diff options
| author | cyfraeviolae <cyfraeviolae> | 2024-04-03 03:10:44 -0400 |
|---|---|---|
| committer | cyfraeviolae <cyfraeviolae> | 2024-04-03 03:10:44 -0400 |
| commit | 6d7ba58f880be618ade07f8ea080fe8c4bf8a896 (patch) | |
| tree | b1c931051ffcebd2bd9d61d98d6233ffa289bbce /venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql | |
| parent | 4f884c9abc32990b4061a1bb6997b4b37e58ea0b (diff) | |
venv
Diffstat (limited to 'venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql')
38 files changed, 13369 insertions, 0 deletions
diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__init__.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__init__.py new file mode 100644 index 0000000..325ea88 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__init__.py @@ -0,0 +1,167 @@ +# dialects/postgresql/__init__.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +from types import ModuleType + +from . import array as arraylib # noqa # keep above base and other dialects +from . import asyncpg # noqa +from . import base +from . import pg8000 # noqa +from . import psycopg # noqa +from . import psycopg2 # noqa +from . import psycopg2cffi # noqa +from .array import All +from .array import Any +from .array import ARRAY +from .array import array +from .base import BIGINT +from .base import BOOLEAN +from .base import CHAR +from .base import DATE +from .base import DOMAIN +from .base import DOUBLE_PRECISION +from .base import FLOAT +from .base import INTEGER +from .base import NUMERIC +from .base import REAL +from .base import SMALLINT +from .base import TEXT +from .base import UUID +from .base import VARCHAR +from .dml import Insert +from .dml import insert +from .ext import aggregate_order_by +from .ext import array_agg +from .ext import ExcludeConstraint +from .ext import phraseto_tsquery +from .ext import plainto_tsquery +from .ext import to_tsquery +from .ext import to_tsvector +from .ext import ts_headline +from .ext import websearch_to_tsquery +from .hstore import HSTORE +from .hstore import hstore +from .json import JSON +from .json import JSONB +from .json import JSONPATH +from .named_types import CreateDomainType +from .named_types import CreateEnumType +from .named_types import DropDomainType +from .named_types import DropEnumType +from .named_types import ENUM +from .named_types import NamedType +from .ranges import AbstractMultiRange +from .ranges import AbstractRange +from .ranges import AbstractSingleRange +from .ranges import DATEMULTIRANGE +from .ranges import DATERANGE +from .ranges import INT4MULTIRANGE +from .ranges import INT4RANGE +from .ranges import INT8MULTIRANGE +from .ranges import INT8RANGE +from .ranges import MultiRange +from .ranges import NUMMULTIRANGE +from .ranges import NUMRANGE +from .ranges import Range +from .ranges import TSMULTIRANGE +from .ranges import TSRANGE +from .ranges import TSTZMULTIRANGE +from .ranges import TSTZRANGE +from .types import BIT +from .types import BYTEA +from .types import CIDR +from .types import CITEXT +from .types import INET +from .types import INTERVAL +from .types import MACADDR +from .types import MACADDR8 +from .types import MONEY +from .types import OID +from .types import REGCLASS +from .types import REGCONFIG +from .types import TIME +from .types import TIMESTAMP +from .types import TSQUERY +from .types import TSVECTOR + + +# Alias psycopg also as psycopg_async +psycopg_async = type( + "psycopg_async", (ModuleType,), {"dialect": psycopg.dialect_async} +) + +base.dialect = dialect = psycopg2.dialect + + +__all__ = ( + "INTEGER", + "BIGINT", + "SMALLINT", + "VARCHAR", + "CHAR", + "TEXT", + "NUMERIC", + "FLOAT", + "REAL", + "INET", + "CIDR", + "CITEXT", + "UUID", + "BIT", + "MACADDR", + "MACADDR8", + "MONEY", + "OID", + "REGCLASS", + "REGCONFIG", + "TSQUERY", + "TSVECTOR", + "DOUBLE_PRECISION", + "TIMESTAMP", + "TIME", + "DATE", + "BYTEA", + "BOOLEAN", + "INTERVAL", + "ARRAY", + "ENUM", + "DOMAIN", + "dialect", + "array", + "HSTORE", + "hstore", + "INT4RANGE", + "INT8RANGE", + "NUMRANGE", + "DATERANGE", + "INT4MULTIRANGE", + "INT8MULTIRANGE", + "NUMMULTIRANGE", + "DATEMULTIRANGE", + "TSVECTOR", + "TSRANGE", + "TSTZRANGE", + "TSMULTIRANGE", + "TSTZMULTIRANGE", + "JSON", + "JSONB", + "JSONPATH", + "Any", + "All", + "DropEnumType", + "DropDomainType", + "CreateDomainType", + "NamedType", + "CreateEnumType", + "ExcludeConstraint", + "Range", + "aggregate_order_by", + "array_agg", + "insert", + "Insert", +) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..620cf0d --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/__init__.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..fc1553f --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/_psycopg_common.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..a257440 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/array.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..50a38f0 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/asyncpg.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..b5aaeaa --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/base.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..27eae2b --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/dml.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..7ae58de --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ext.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..c004810 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/hstore.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..c429892 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/json.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..a075cd8 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/named_types.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..93ccdfb --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/operators.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..1bbbf9c --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg8000.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..23fc72e --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/pg_catalog.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..39fb4c6 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/provision.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..aa442c1 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..1308229 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..bd4f6c4 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/psycopg2cffi.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..d0a785d --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/ranges.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-311.pyc b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-311.pyc Binary files differnew file mode 100644 index 0000000..ed561da --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/__pycache__/types.cpython-311.pyc diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py new file mode 100644 index 0000000..46858c9 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/_psycopg_common.py @@ -0,0 +1,187 @@ +# dialects/postgresql/_psycopg_common.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors +from __future__ import annotations + +import decimal + +from .array import ARRAY as PGARRAY +from .base import _DECIMAL_TYPES +from .base import _FLOAT_TYPES +from .base import _INT_TYPES +from .base import PGDialect +from .base import PGExecutionContext +from .hstore import HSTORE +from .pg_catalog import _SpaceVector +from .pg_catalog import INT2VECTOR +from .pg_catalog import OIDVECTOR +from ... import exc +from ... import types as sqltypes +from ... import util +from ...engine import processors + +_server_side_id = util.counter() + + +class _PsycopgNumeric(sqltypes.Numeric): + def bind_processor(self, dialect): + return None + + def result_processor(self, dialect, coltype): + if self.asdecimal: + if coltype in _FLOAT_TYPES: + return processors.to_decimal_processor_factory( + decimal.Decimal, self._effective_decimal_return_scale + ) + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + # psycopg returns Decimal natively for 1700 + return None + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + else: + if coltype in _FLOAT_TYPES: + # psycopg returns float natively for 701 + return None + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + return processors.to_float + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + + +class _PsycopgFloat(_PsycopgNumeric): + __visit_name__ = "float" + + +class _PsycopgHStore(HSTORE): + def bind_processor(self, dialect): + if dialect._has_native_hstore: + return None + else: + return super().bind_processor(dialect) + + def result_processor(self, dialect, coltype): + if dialect._has_native_hstore: + return None + else: + return super().result_processor(dialect, coltype) + + +class _PsycopgARRAY(PGARRAY): + render_bind_cast = True + + +class _PsycopgINT2VECTOR(_SpaceVector, INT2VECTOR): + pass + + +class _PsycopgOIDVECTOR(_SpaceVector, OIDVECTOR): + pass + + +class _PGExecutionContext_common_psycopg(PGExecutionContext): + def create_server_side_cursor(self): + # use server-side cursors: + # psycopg + # https://www.psycopg.org/psycopg3/docs/advanced/cursors.html#server-side-cursors + # psycopg2 + # https://www.psycopg.org/docs/usage.html#server-side-cursors + ident = "c_%s_%s" % (hex(id(self))[2:], hex(_server_side_id())[2:]) + return self._dbapi_connection.cursor(ident) + + +class _PGDialect_common_psycopg(PGDialect): + supports_statement_cache = True + supports_server_side_cursors = True + + default_paramstyle = "pyformat" + + _has_native_hstore = True + + colspecs = util.update_copy( + PGDialect.colspecs, + { + sqltypes.Numeric: _PsycopgNumeric, + sqltypes.Float: _PsycopgFloat, + HSTORE: _PsycopgHStore, + sqltypes.ARRAY: _PsycopgARRAY, + INT2VECTOR: _PsycopgINT2VECTOR, + OIDVECTOR: _PsycopgOIDVECTOR, + }, + ) + + def __init__( + self, + client_encoding=None, + use_native_hstore=True, + **kwargs, + ): + PGDialect.__init__(self, **kwargs) + if not use_native_hstore: + self._has_native_hstore = False + self.use_native_hstore = use_native_hstore + self.client_encoding = client_encoding + + def create_connect_args(self, url): + opts = url.translate_connect_args(username="user", database="dbname") + + multihosts, multiports = self._split_multihost_from_url(url) + + if opts or url.query: + if not opts: + opts = {} + if "port" in opts: + opts["port"] = int(opts["port"]) + opts.update(url.query) + + if multihosts: + opts["host"] = ",".join(multihosts) + comma_ports = ",".join(str(p) if p else "" for p in multiports) + if comma_ports: + opts["port"] = comma_ports + return ([], opts) + else: + # no connection arguments whatsoever; psycopg2.connect() + # requires that "dsn" be present as a blank string. + return ([""], opts) + + def get_isolation_level_values(self, dbapi_connection): + return ( + "AUTOCOMMIT", + "READ COMMITTED", + "READ UNCOMMITTED", + "REPEATABLE READ", + "SERIALIZABLE", + ) + + def set_deferrable(self, connection, value): + connection.deferrable = value + + def get_deferrable(self, connection): + return connection.deferrable + + def _do_autocommit(self, connection, value): + connection.autocommit = value + + def do_ping(self, dbapi_connection): + cursor = None + before_autocommit = dbapi_connection.autocommit + + if not before_autocommit: + dbapi_connection.autocommit = True + cursor = dbapi_connection.cursor() + try: + cursor.execute(self._dialect_specific_select_one) + finally: + cursor.close() + if not before_autocommit and not dbapi_connection.closed: + dbapi_connection.autocommit = before_autocommit + + return True diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/array.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/array.py new file mode 100644 index 0000000..1d63655 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/array.py @@ -0,0 +1,425 @@ +# dialects/postgresql/array.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + + +from __future__ import annotations + +import re +from typing import Any +from typing import Optional +from typing import TypeVar + +from .operators import CONTAINED_BY +from .operators import CONTAINS +from .operators import OVERLAP +from ... import types as sqltypes +from ... import util +from ...sql import expression +from ...sql import operators +from ...sql._typing import _TypeEngineArgument + + +_T = TypeVar("_T", bound=Any) + + +def Any(other, arrexpr, operator=operators.eq): + """A synonym for the ARRAY-level :meth:`.ARRAY.Comparator.any` method. + See that method for details. + + """ + + return arrexpr.any(other, operator) + + +def All(other, arrexpr, operator=operators.eq): + """A synonym for the ARRAY-level :meth:`.ARRAY.Comparator.all` method. + See that method for details. + + """ + + return arrexpr.all(other, operator) + + +class array(expression.ExpressionClauseList[_T]): + """A PostgreSQL ARRAY literal. + + This is used to produce ARRAY literals in SQL expressions, e.g.:: + + from sqlalchemy.dialects.postgresql import array + from sqlalchemy.dialects import postgresql + from sqlalchemy import select, func + + stmt = select(array([1,2]) + array([3,4,5])) + + print(stmt.compile(dialect=postgresql.dialect())) + + Produces the SQL:: + + SELECT ARRAY[%(param_1)s, %(param_2)s] || + ARRAY[%(param_3)s, %(param_4)s, %(param_5)s]) AS anon_1 + + An instance of :class:`.array` will always have the datatype + :class:`_types.ARRAY`. The "inner" type of the array is inferred from + the values present, unless the ``type_`` keyword argument is passed:: + + array(['foo', 'bar'], type_=CHAR) + + Multidimensional arrays are produced by nesting :class:`.array` constructs. + The dimensionality of the final :class:`_types.ARRAY` + type is calculated by + recursively adding the dimensions of the inner :class:`_types.ARRAY` + type:: + + stmt = select( + array([ + array([1, 2]), array([3, 4]), array([column('q'), column('x')]) + ]) + ) + print(stmt.compile(dialect=postgresql.dialect())) + + Produces:: + + SELECT ARRAY[ARRAY[%(param_1)s, %(param_2)s], + ARRAY[%(param_3)s, %(param_4)s], ARRAY[q, x]] AS anon_1 + + .. versionadded:: 1.3.6 added support for multidimensional array literals + + .. seealso:: + + :class:`_postgresql.ARRAY` + + """ + + __visit_name__ = "array" + + stringify_dialect = "postgresql" + inherit_cache = True + + def __init__(self, clauses, **kw): + type_arg = kw.pop("type_", None) + super().__init__(operators.comma_op, *clauses, **kw) + + self._type_tuple = [arg.type for arg in self.clauses] + + main_type = ( + type_arg + if type_arg is not None + else self._type_tuple[0] if self._type_tuple else sqltypes.NULLTYPE + ) + + if isinstance(main_type, ARRAY): + self.type = ARRAY( + main_type.item_type, + dimensions=( + main_type.dimensions + 1 + if main_type.dimensions is not None + else 2 + ), + ) + else: + self.type = ARRAY(main_type) + + @property + def _select_iterable(self): + return (self,) + + def _bind_param(self, operator, obj, _assume_scalar=False, type_=None): + if _assume_scalar or operator is operators.getitem: + return expression.BindParameter( + None, + obj, + _compared_to_operator=operator, + type_=type_, + _compared_to_type=self.type, + unique=True, + ) + + else: + return array( + [ + self._bind_param( + operator, o, _assume_scalar=True, type_=type_ + ) + for o in obj + ] + ) + + def self_group(self, against=None): + if against in (operators.any_op, operators.all_op, operators.getitem): + return expression.Grouping(self) + else: + return self + + +class ARRAY(sqltypes.ARRAY): + """PostgreSQL ARRAY type. + + The :class:`_postgresql.ARRAY` type is constructed in the same way + as the core :class:`_types.ARRAY` type; a member type is required, and a + number of dimensions is recommended if the type is to be used for more + than one dimension:: + + from sqlalchemy.dialects import postgresql + + mytable = Table("mytable", metadata, + Column("data", postgresql.ARRAY(Integer, dimensions=2)) + ) + + The :class:`_postgresql.ARRAY` type provides all operations defined on the + core :class:`_types.ARRAY` type, including support for "dimensions", + indexed access, and simple matching such as + :meth:`.types.ARRAY.Comparator.any` and + :meth:`.types.ARRAY.Comparator.all`. :class:`_postgresql.ARRAY` + class also + provides PostgreSQL-specific methods for containment operations, including + :meth:`.postgresql.ARRAY.Comparator.contains` + :meth:`.postgresql.ARRAY.Comparator.contained_by`, and + :meth:`.postgresql.ARRAY.Comparator.overlap`, e.g.:: + + mytable.c.data.contains([1, 2]) + + Indexed access is one-based by default, to match that of PostgreSQL; + for zero-based indexed access, set + :paramref:`_postgresql.ARRAY.zero_indexes`. + + Additionally, the :class:`_postgresql.ARRAY` + type does not work directly in + conjunction with the :class:`.ENUM` type. For a workaround, see the + special type at :ref:`postgresql_array_of_enum`. + + .. container:: topic + + **Detecting Changes in ARRAY columns when using the ORM** + + The :class:`_postgresql.ARRAY` type, when used with the SQLAlchemy ORM, + does not detect in-place mutations to the array. In order to detect + these, the :mod:`sqlalchemy.ext.mutable` extension must be used, using + the :class:`.MutableList` class:: + + from sqlalchemy.dialects.postgresql import ARRAY + from sqlalchemy.ext.mutable import MutableList + + class SomeOrmClass(Base): + # ... + + data = Column(MutableList.as_mutable(ARRAY(Integer))) + + This extension will allow "in-place" changes such to the array + such as ``.append()`` to produce events which will be detected by the + unit of work. Note that changes to elements **inside** the array, + including subarrays that are mutated in place, are **not** detected. + + Alternatively, assigning a new array value to an ORM element that + replaces the old one will always trigger a change event. + + .. seealso:: + + :class:`_types.ARRAY` - base array type + + :class:`_postgresql.array` - produces a literal array value. + + """ + + def __init__( + self, + item_type: _TypeEngineArgument[Any], + as_tuple: bool = False, + dimensions: Optional[int] = None, + zero_indexes: bool = False, + ): + """Construct an ARRAY. + + E.g.:: + + Column('myarray', ARRAY(Integer)) + + Arguments are: + + :param item_type: The data type of items of this array. Note that + dimensionality is irrelevant here, so multi-dimensional arrays like + ``INTEGER[][]``, are constructed as ``ARRAY(Integer)``, not as + ``ARRAY(ARRAY(Integer))`` or such. + + :param as_tuple=False: Specify whether return results + should be converted to tuples from lists. DBAPIs such + as psycopg2 return lists by default. When tuples are + returned, the results are hashable. + + :param dimensions: if non-None, the ARRAY will assume a fixed + number of dimensions. This will cause the DDL emitted for this + ARRAY to include the exact number of bracket clauses ``[]``, + and will also optimize the performance of the type overall. + Note that PG arrays are always implicitly "non-dimensioned", + meaning they can store any number of dimensions no matter how + they were declared. + + :param zero_indexes=False: when True, index values will be converted + between Python zero-based and PostgreSQL one-based indexes, e.g. + a value of one will be added to all index values before passing + to the database. + + """ + if isinstance(item_type, ARRAY): + raise ValueError( + "Do not nest ARRAY types; ARRAY(basetype) " + "handles multi-dimensional arrays of basetype" + ) + if isinstance(item_type, type): + item_type = item_type() + self.item_type = item_type + self.as_tuple = as_tuple + self.dimensions = dimensions + self.zero_indexes = zero_indexes + + class Comparator(sqltypes.ARRAY.Comparator): + """Define comparison operations for :class:`_types.ARRAY`. + + Note that these operations are in addition to those provided + by the base :class:`.types.ARRAY.Comparator` class, including + :meth:`.types.ARRAY.Comparator.any` and + :meth:`.types.ARRAY.Comparator.all`. + + """ + + def contains(self, other, **kwargs): + """Boolean expression. Test if elements are a superset of the + elements of the argument array expression. + + kwargs may be ignored by this operator but are required for API + conformance. + """ + return self.operate(CONTAINS, other, result_type=sqltypes.Boolean) + + def contained_by(self, other): + """Boolean expression. Test if elements are a proper subset of the + elements of the argument array expression. + """ + return self.operate( + CONTAINED_BY, other, result_type=sqltypes.Boolean + ) + + def overlap(self, other): + """Boolean expression. Test if array has elements in common with + an argument array expression. + """ + return self.operate(OVERLAP, other, result_type=sqltypes.Boolean) + + comparator_factory = Comparator + + @property + def hashable(self): + return self.as_tuple + + @property + def python_type(self): + return list + + def compare_values(self, x, y): + return x == y + + @util.memoized_property + def _against_native_enum(self): + return ( + isinstance(self.item_type, sqltypes.Enum) + and self.item_type.native_enum + ) + + def literal_processor(self, dialect): + item_proc = self.item_type.dialect_impl(dialect).literal_processor( + dialect + ) + if item_proc is None: + return None + + def to_str(elements): + return f"ARRAY[{', '.join(elements)}]" + + def process(value): + inner = self._apply_item_processor( + value, item_proc, self.dimensions, to_str + ) + return inner + + return process + + def bind_processor(self, dialect): + item_proc = self.item_type.dialect_impl(dialect).bind_processor( + dialect + ) + + def process(value): + if value is None: + return value + else: + return self._apply_item_processor( + value, item_proc, self.dimensions, list + ) + + return process + + def result_processor(self, dialect, coltype): + item_proc = self.item_type.dialect_impl(dialect).result_processor( + dialect, coltype + ) + + def process(value): + if value is None: + return value + else: + return self._apply_item_processor( + value, + item_proc, + self.dimensions, + tuple if self.as_tuple else list, + ) + + if self._against_native_enum: + super_rp = process + pattern = re.compile(r"^{(.*)}$") + + def handle_raw_string(value): + inner = pattern.match(value).group(1) + return _split_enum_values(inner) + + def process(value): + if value is None: + return value + # isinstance(value, str) is required to handle + # the case where a TypeDecorator for and Array of Enum is + # used like was required in sa < 1.3.17 + return super_rp( + handle_raw_string(value) + if isinstance(value, str) + else value + ) + + return process + + +def _split_enum_values(array_string): + if '"' not in array_string: + # no escape char is present so it can just split on the comma + return array_string.split(",") if array_string else [] + + # handles quoted strings from: + # r'abc,"quoted","also\\\\quoted", "quoted, comma", "esc \" quot", qpr' + # returns + # ['abc', 'quoted', 'also\\quoted', 'quoted, comma', 'esc " quot', 'qpr'] + text = array_string.replace(r"\"", "_$ESC_QUOTE$_") + text = text.replace(r"\\", "\\") + result = [] + on_quotes = re.split(r'(")', text) + in_quotes = False + for tok in on_quotes: + if tok == '"': + in_quotes = not in_quotes + elif in_quotes: + result.append(tok.replace("_$ESC_QUOTE$_", '"')) + else: + result.extend(re.findall(r"([^\s,]+),?", tok)) + return result diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py new file mode 100644 index 0000000..df2656d --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/asyncpg.py @@ -0,0 +1,1262 @@ +# dialects/postgresql/asyncpg.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors <see AUTHORS +# file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql+asyncpg + :name: asyncpg + :dbapi: asyncpg + :connectstring: postgresql+asyncpg://user:password@host:port/dbname[?key=value&key=value...] + :url: https://magicstack.github.io/asyncpg/ + +The asyncpg dialect is SQLAlchemy's first Python asyncio dialect. + +Using a special asyncio mediation layer, the asyncpg dialect is usable +as the backend for the :ref:`SQLAlchemy asyncio <asyncio_toplevel>` +extension package. + +This dialect should normally be used only with the +:func:`_asyncio.create_async_engine` engine creation function:: + + from sqlalchemy.ext.asyncio import create_async_engine + engine = create_async_engine("postgresql+asyncpg://user:pass@hostname/dbname") + +.. versionadded:: 1.4 + +.. note:: + + By default asyncpg does not decode the ``json`` and ``jsonb`` types and + returns them as strings. SQLAlchemy sets default type decoder for ``json`` + and ``jsonb`` types using the python builtin ``json.loads`` function. + The json implementation used can be changed by setting the attribute + ``json_deserializer`` when creating the engine with + :func:`create_engine` or :func:`create_async_engine`. + +.. _asyncpg_multihost: + +Multihost Connections +-------------------------- + +The asyncpg dialect features support for multiple fallback hosts in the +same way as that of the psycopg2 and psycopg dialects. The +syntax is the same, +using ``host=<host>:<port>`` combinations as additional query string arguments; +however, there is no default port, so all hosts must have a complete port number +present, otherwise an exception is raised:: + + engine = create_async_engine( + "postgresql+asyncpg://user:password@/dbname?host=HostA:5432&host=HostB:5432&host=HostC:5432" + ) + +For complete background on this syntax, see :ref:`psycopg2_multi_host`. + +.. versionadded:: 2.0.18 + +.. seealso:: + + :ref:`psycopg2_multi_host` + +.. _asyncpg_prepared_statement_cache: + +Prepared Statement Cache +-------------------------- + +The asyncpg SQLAlchemy dialect makes use of ``asyncpg.connection.prepare()`` +for all statements. The prepared statement objects are cached after +construction which appears to grant a 10% or more performance improvement for +statement invocation. The cache is on a per-DBAPI connection basis, which +means that the primary storage for prepared statements is within DBAPI +connections pooled within the connection pool. The size of this cache +defaults to 100 statements per DBAPI connection and may be adjusted using the +``prepared_statement_cache_size`` DBAPI argument (note that while this argument +is implemented by SQLAlchemy, it is part of the DBAPI emulation portion of the +asyncpg dialect, therefore is handled as a DBAPI argument, not a dialect +argument):: + + + engine = create_async_engine("postgresql+asyncpg://user:pass@hostname/dbname?prepared_statement_cache_size=500") + +To disable the prepared statement cache, use a value of zero:: + + engine = create_async_engine("postgresql+asyncpg://user:pass@hostname/dbname?prepared_statement_cache_size=0") + +.. versionadded:: 1.4.0b2 Added ``prepared_statement_cache_size`` for asyncpg. + + +.. warning:: The ``asyncpg`` database driver necessarily uses caches for + PostgreSQL type OIDs, which become stale when custom PostgreSQL datatypes + such as ``ENUM`` objects are changed via DDL operations. Additionally, + prepared statements themselves which are optionally cached by SQLAlchemy's + driver as described above may also become "stale" when DDL has been emitted + to the PostgreSQL database which modifies the tables or other objects + involved in a particular prepared statement. + + The SQLAlchemy asyncpg dialect will invalidate these caches within its local + process when statements that represent DDL are emitted on a local + connection, but this is only controllable within a single Python process / + database engine. If DDL changes are made from other database engines + and/or processes, a running application may encounter asyncpg exceptions + ``InvalidCachedStatementError`` and/or ``InternalServerError("cache lookup + failed for type <oid>")`` if it refers to pooled database connections which + operated upon the previous structures. The SQLAlchemy asyncpg dialect will + recover from these error cases when the driver raises these exceptions by + clearing its internal caches as well as those of the asyncpg driver in + response to them, but cannot prevent them from being raised in the first + place if the cached prepared statement or asyncpg type caches have gone + stale, nor can it retry the statement as the PostgreSQL transaction is + invalidated when these errors occur. + +.. _asyncpg_prepared_statement_name: + +Prepared Statement Name with PGBouncer +-------------------------------------- + +By default, asyncpg enumerates prepared statements in numeric order, which +can lead to errors if a name has already been taken for another prepared +statement. This issue can arise if your application uses database proxies +such as PgBouncer to handle connections. One possible workaround is to +use dynamic prepared statement names, which asyncpg now supports through +an optional ``name`` value for the statement name. This allows you to +generate your own unique names that won't conflict with existing ones. +To achieve this, you can provide a function that will be called every time +a prepared statement is prepared:: + + from uuid import uuid4 + + engine = create_async_engine( + "postgresql+asyncpg://user:pass@somepgbouncer/dbname", + poolclass=NullPool, + connect_args={ + 'prepared_statement_name_func': lambda: f'__asyncpg_{uuid4()}__', + }, + ) + +.. seealso:: + + https://github.com/MagicStack/asyncpg/issues/837 + + https://github.com/sqlalchemy/sqlalchemy/issues/6467 + +.. warning:: When using PGBouncer, to prevent a buildup of useless prepared statements in + your application, it's important to use the :class:`.NullPool` pool + class, and to configure PgBouncer to use `DISCARD <https://www.postgresql.org/docs/current/sql-discard.html>`_ + when returning connections. The DISCARD command is used to release resources held by the db connection, + including prepared statements. Without proper setup, prepared statements can + accumulate quickly and cause performance issues. + +Disabling the PostgreSQL JIT to improve ENUM datatype handling +--------------------------------------------------------------- + +Asyncpg has an `issue <https://github.com/MagicStack/asyncpg/issues/727>`_ when +using PostgreSQL ENUM datatypes, where upon the creation of new database +connections, an expensive query may be emitted in order to retrieve metadata +regarding custom types which has been shown to negatively affect performance. +To mitigate this issue, the PostgreSQL "jit" setting may be disabled from the +client using this setting passed to :func:`_asyncio.create_async_engine`:: + + engine = create_async_engine( + "postgresql+asyncpg://user:password@localhost/tmp", + connect_args={"server_settings": {"jit": "off"}}, + ) + +.. seealso:: + + https://github.com/MagicStack/asyncpg/issues/727 + +""" # noqa + +from __future__ import annotations + +import collections +import decimal +import json as _py_json +import re +import time + +from . import json +from . import ranges +from .array import ARRAY as PGARRAY +from .base import _DECIMAL_TYPES +from .base import _FLOAT_TYPES +from .base import _INT_TYPES +from .base import ENUM +from .base import INTERVAL +from .base import OID +from .base import PGCompiler +from .base import PGDialect +from .base import PGExecutionContext +from .base import PGIdentifierPreparer +from .base import REGCLASS +from .base import REGCONFIG +from .types import BIT +from .types import BYTEA +from .types import CITEXT +from ... import exc +from ... import pool +from ... import util +from ...engine import AdaptedConnection +from ...engine import processors +from ...sql import sqltypes +from ...util.concurrency import asyncio +from ...util.concurrency import await_fallback +from ...util.concurrency import await_only + + +class AsyncpgARRAY(PGARRAY): + render_bind_cast = True + + +class AsyncpgString(sqltypes.String): + render_bind_cast = True + + +class AsyncpgREGCONFIG(REGCONFIG): + render_bind_cast = True + + +class AsyncpgTime(sqltypes.Time): + render_bind_cast = True + + +class AsyncpgBit(BIT): + render_bind_cast = True + + +class AsyncpgByteA(BYTEA): + render_bind_cast = True + + +class AsyncpgDate(sqltypes.Date): + render_bind_cast = True + + +class AsyncpgDateTime(sqltypes.DateTime): + render_bind_cast = True + + +class AsyncpgBoolean(sqltypes.Boolean): + render_bind_cast = True + + +class AsyncPgInterval(INTERVAL): + render_bind_cast = True + + @classmethod + def adapt_emulated_to_native(cls, interval, **kw): + return AsyncPgInterval(precision=interval.second_precision) + + +class AsyncPgEnum(ENUM): + render_bind_cast = True + + +class AsyncpgInteger(sqltypes.Integer): + render_bind_cast = True + + +class AsyncpgBigInteger(sqltypes.BigInteger): + render_bind_cast = True + + +class AsyncpgJSON(json.JSON): + render_bind_cast = True + + def result_processor(self, dialect, coltype): + return None + + +class AsyncpgJSONB(json.JSONB): + render_bind_cast = True + + def result_processor(self, dialect, coltype): + return None + + +class AsyncpgJSONIndexType(sqltypes.JSON.JSONIndexType): + pass + + +class AsyncpgJSONIntIndexType(sqltypes.JSON.JSONIntIndexType): + __visit_name__ = "json_int_index" + + render_bind_cast = True + + +class AsyncpgJSONStrIndexType(sqltypes.JSON.JSONStrIndexType): + __visit_name__ = "json_str_index" + + render_bind_cast = True + + +class AsyncpgJSONPathType(json.JSONPathType): + def bind_processor(self, dialect): + def process(value): + if isinstance(value, str): + # If it's already a string assume that it's in json path + # format. This allows using cast with json paths literals + return value + elif value: + tokens = [str(elem) for elem in value] + return tokens + else: + return [] + + return process + + +class AsyncpgNumeric(sqltypes.Numeric): + render_bind_cast = True + + def bind_processor(self, dialect): + return None + + def result_processor(self, dialect, coltype): + if self.asdecimal: + if coltype in _FLOAT_TYPES: + return processors.to_decimal_processor_factory( + decimal.Decimal, self._effective_decimal_return_scale + ) + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + # pg8000 returns Decimal natively for 1700 + return None + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + else: + if coltype in _FLOAT_TYPES: + # pg8000 returns float natively for 701 + return None + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + return processors.to_float + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + + +class AsyncpgFloat(AsyncpgNumeric, sqltypes.Float): + __visit_name__ = "float" + render_bind_cast = True + + +class AsyncpgREGCLASS(REGCLASS): + render_bind_cast = True + + +class AsyncpgOID(OID): + render_bind_cast = True + + +class AsyncpgCHAR(sqltypes.CHAR): + render_bind_cast = True + + +class _AsyncpgRange(ranges.AbstractSingleRangeImpl): + def bind_processor(self, dialect): + asyncpg_Range = dialect.dbapi.asyncpg.Range + + def to_range(value): + if isinstance(value, ranges.Range): + value = asyncpg_Range( + value.lower, + value.upper, + lower_inc=value.bounds[0] == "[", + upper_inc=value.bounds[1] == "]", + empty=value.empty, + ) + return value + + return to_range + + def result_processor(self, dialect, coltype): + def to_range(value): + if value is not None: + empty = value.isempty + value = ranges.Range( + value.lower, + value.upper, + bounds=f"{'[' if empty or value.lower_inc else '('}" # type: ignore # noqa: E501 + f"{']' if not empty and value.upper_inc else ')'}", + empty=empty, + ) + return value + + return to_range + + +class _AsyncpgMultiRange(ranges.AbstractMultiRangeImpl): + def bind_processor(self, dialect): + asyncpg_Range = dialect.dbapi.asyncpg.Range + + NoneType = type(None) + + def to_range(value): + if isinstance(value, (str, NoneType)): + return value + + def to_range(value): + if isinstance(value, ranges.Range): + value = asyncpg_Range( + value.lower, + value.upper, + lower_inc=value.bounds[0] == "[", + upper_inc=value.bounds[1] == "]", + empty=value.empty, + ) + return value + + return [to_range(element) for element in value] + + return to_range + + def result_processor(self, dialect, coltype): + def to_range_array(value): + def to_range(rvalue): + if rvalue is not None: + empty = rvalue.isempty + rvalue = ranges.Range( + rvalue.lower, + rvalue.upper, + bounds=f"{'[' if empty or rvalue.lower_inc else '('}" # type: ignore # noqa: E501 + f"{']' if not empty and rvalue.upper_inc else ')'}", + empty=empty, + ) + return rvalue + + if value is not None: + value = ranges.MultiRange(to_range(elem) for elem in value) + + return value + + return to_range_array + + +class PGExecutionContext_asyncpg(PGExecutionContext): + def handle_dbapi_exception(self, e): + if isinstance( + e, + ( + self.dialect.dbapi.InvalidCachedStatementError, + self.dialect.dbapi.InternalServerError, + ), + ): + self.dialect._invalidate_schema_cache() + + def pre_exec(self): + if self.isddl: + self.dialect._invalidate_schema_cache() + + self.cursor._invalidate_schema_cache_asof = ( + self.dialect._invalidate_schema_cache_asof + ) + + if not self.compiled: + return + + def create_server_side_cursor(self): + return self._dbapi_connection.cursor(server_side=True) + + +class PGCompiler_asyncpg(PGCompiler): + pass + + +class PGIdentifierPreparer_asyncpg(PGIdentifierPreparer): + pass + + +class AsyncAdapt_asyncpg_cursor: + __slots__ = ( + "_adapt_connection", + "_connection", + "_rows", + "description", + "arraysize", + "rowcount", + "_cursor", + "_invalidate_schema_cache_asof", + ) + + server_side = False + + def __init__(self, adapt_connection): + self._adapt_connection = adapt_connection + self._connection = adapt_connection._connection + self._rows = [] + self._cursor = None + self.description = None + self.arraysize = 1 + self.rowcount = -1 + self._invalidate_schema_cache_asof = 0 + + def close(self): + self._rows[:] = [] + + def _handle_exception(self, error): + self._adapt_connection._handle_exception(error) + + async def _prepare_and_execute(self, operation, parameters): + adapt_connection = self._adapt_connection + + async with adapt_connection._execute_mutex: + if not adapt_connection._started: + await adapt_connection._start_transaction() + + if parameters is None: + parameters = () + + try: + prepared_stmt, attributes = await adapt_connection._prepare( + operation, self._invalidate_schema_cache_asof + ) + + if attributes: + self.description = [ + ( + attr.name, + attr.type.oid, + None, + None, + None, + None, + None, + ) + for attr in attributes + ] + else: + self.description = None + + if self.server_side: + self._cursor = await prepared_stmt.cursor(*parameters) + self.rowcount = -1 + else: + self._rows = await prepared_stmt.fetch(*parameters) + status = prepared_stmt.get_statusmsg() + + reg = re.match( + r"(?:SELECT|UPDATE|DELETE|INSERT \d+) (\d+)", status + ) + if reg: + self.rowcount = int(reg.group(1)) + else: + self.rowcount = -1 + + except Exception as error: + self._handle_exception(error) + + async def _executemany(self, operation, seq_of_parameters): + adapt_connection = self._adapt_connection + + self.description = None + async with adapt_connection._execute_mutex: + await adapt_connection._check_type_cache_invalidation( + self._invalidate_schema_cache_asof + ) + + if not adapt_connection._started: + await adapt_connection._start_transaction() + + try: + return await self._connection.executemany( + operation, seq_of_parameters + ) + except Exception as error: + self._handle_exception(error) + + def execute(self, operation, parameters=None): + self._adapt_connection.await_( + self._prepare_and_execute(operation, parameters) + ) + + def executemany(self, operation, seq_of_parameters): + return self._adapt_connection.await_( + self._executemany(operation, seq_of_parameters) + ) + + def setinputsizes(self, *inputsizes): + raise NotImplementedError() + + def __iter__(self): + while self._rows: + yield self._rows.pop(0) + + def fetchone(self): + if self._rows: + return self._rows.pop(0) + else: + return None + + def fetchmany(self, size=None): + if size is None: + size = self.arraysize + + retval = self._rows[0:size] + self._rows[:] = self._rows[size:] + return retval + + def fetchall(self): + retval = self._rows[:] + self._rows[:] = [] + return retval + + +class AsyncAdapt_asyncpg_ss_cursor(AsyncAdapt_asyncpg_cursor): + server_side = True + __slots__ = ("_rowbuffer",) + + def __init__(self, adapt_connection): + super().__init__(adapt_connection) + self._rowbuffer = None + + def close(self): + self._cursor = None + self._rowbuffer = None + + def _buffer_rows(self): + new_rows = self._adapt_connection.await_(self._cursor.fetch(50)) + self._rowbuffer = collections.deque(new_rows) + + def __aiter__(self): + return self + + async def __anext__(self): + if not self._rowbuffer: + self._buffer_rows() + + while True: + while self._rowbuffer: + yield self._rowbuffer.popleft() + + self._buffer_rows() + if not self._rowbuffer: + break + + def fetchone(self): + if not self._rowbuffer: + self._buffer_rows() + if not self._rowbuffer: + return None + return self._rowbuffer.popleft() + + def fetchmany(self, size=None): + if size is None: + return self.fetchall() + + if not self._rowbuffer: + self._buffer_rows() + + buf = list(self._rowbuffer) + lb = len(buf) + if size > lb: + buf.extend( + self._adapt_connection.await_(self._cursor.fetch(size - lb)) + ) + + result = buf[0:size] + self._rowbuffer = collections.deque(buf[size:]) + return result + + def fetchall(self): + ret = list(self._rowbuffer) + list( + self._adapt_connection.await_(self._all()) + ) + self._rowbuffer.clear() + return ret + + async def _all(self): + rows = [] + + # TODO: looks like we have to hand-roll some kind of batching here. + # hardcoding for the moment but this should be improved. + while True: + batch = await self._cursor.fetch(1000) + if batch: + rows.extend(batch) + continue + else: + break + return rows + + def executemany(self, operation, seq_of_parameters): + raise NotImplementedError( + "server side cursor doesn't support executemany yet" + ) + + +class AsyncAdapt_asyncpg_connection(AdaptedConnection): + __slots__ = ( + "dbapi", + "isolation_level", + "_isolation_setting", + "readonly", + "deferrable", + "_transaction", + "_started", + "_prepared_statement_cache", + "_prepared_statement_name_func", + "_invalidate_schema_cache_asof", + "_execute_mutex", + ) + + await_ = staticmethod(await_only) + + def __init__( + self, + dbapi, + connection, + prepared_statement_cache_size=100, + prepared_statement_name_func=None, + ): + self.dbapi = dbapi + self._connection = connection + self.isolation_level = self._isolation_setting = "read_committed" + self.readonly = False + self.deferrable = False + self._transaction = None + self._started = False + self._invalidate_schema_cache_asof = time.time() + self._execute_mutex = asyncio.Lock() + + if prepared_statement_cache_size: + self._prepared_statement_cache = util.LRUCache( + prepared_statement_cache_size + ) + else: + self._prepared_statement_cache = None + + if prepared_statement_name_func: + self._prepared_statement_name_func = prepared_statement_name_func + else: + self._prepared_statement_name_func = self._default_name_func + + async def _check_type_cache_invalidation(self, invalidate_timestamp): + if invalidate_timestamp > self._invalidate_schema_cache_asof: + await self._connection.reload_schema_state() + self._invalidate_schema_cache_asof = invalidate_timestamp + + async def _prepare(self, operation, invalidate_timestamp): + await self._check_type_cache_invalidation(invalidate_timestamp) + + cache = self._prepared_statement_cache + if cache is None: + prepared_stmt = await self._connection.prepare( + operation, name=self._prepared_statement_name_func() + ) + attributes = prepared_stmt.get_attributes() + return prepared_stmt, attributes + + # asyncpg uses a type cache for the "attributes" which seems to go + # stale independently of the PreparedStatement itself, so place that + # collection in the cache as well. + if operation in cache: + prepared_stmt, attributes, cached_timestamp = cache[operation] + + # preparedstatements themselves also go stale for certain DDL + # changes such as size of a VARCHAR changing, so there is also + # a cross-connection invalidation timestamp + if cached_timestamp > invalidate_timestamp: + return prepared_stmt, attributes + + prepared_stmt = await self._connection.prepare( + operation, name=self._prepared_statement_name_func() + ) + attributes = prepared_stmt.get_attributes() + cache[operation] = (prepared_stmt, attributes, time.time()) + + return prepared_stmt, attributes + + def _handle_exception(self, error): + if self._connection.is_closed(): + self._transaction = None + self._started = False + + if not isinstance(error, AsyncAdapt_asyncpg_dbapi.Error): + exception_mapping = self.dbapi._asyncpg_error_translate + + for super_ in type(error).__mro__: + if super_ in exception_mapping: + translated_error = exception_mapping[super_]( + "%s: %s" % (type(error), error) + ) + translated_error.pgcode = translated_error.sqlstate = ( + getattr(error, "sqlstate", None) + ) + raise translated_error from error + else: + raise error + else: + raise error + + @property + def autocommit(self): + return self.isolation_level == "autocommit" + + @autocommit.setter + def autocommit(self, value): + if value: + self.isolation_level = "autocommit" + else: + self.isolation_level = self._isolation_setting + + def ping(self): + try: + _ = self.await_(self._async_ping()) + except Exception as error: + self._handle_exception(error) + + async def _async_ping(self): + if self._transaction is None and self.isolation_level != "autocommit": + # create a tranasction explicitly to support pgbouncer + # transaction mode. See #10226 + tr = self._connection.transaction() + await tr.start() + try: + await self._connection.fetchrow(";") + finally: + await tr.rollback() + else: + await self._connection.fetchrow(";") + + def set_isolation_level(self, level): + if self._started: + self.rollback() + self.isolation_level = self._isolation_setting = level + + async def _start_transaction(self): + if self.isolation_level == "autocommit": + return + + try: + self._transaction = self._connection.transaction( + isolation=self.isolation_level, + readonly=self.readonly, + deferrable=self.deferrable, + ) + await self._transaction.start() + except Exception as error: + self._handle_exception(error) + else: + self._started = True + + def cursor(self, server_side=False): + if server_side: + return AsyncAdapt_asyncpg_ss_cursor(self) + else: + return AsyncAdapt_asyncpg_cursor(self) + + def rollback(self): + if self._started: + try: + self.await_(self._transaction.rollback()) + except Exception as error: + self._handle_exception(error) + finally: + self._transaction = None + self._started = False + + def commit(self): + if self._started: + try: + self.await_(self._transaction.commit()) + except Exception as error: + self._handle_exception(error) + finally: + self._transaction = None + self._started = False + + def close(self): + self.rollback() + + self.await_(self._connection.close()) + + def terminate(self): + if util.concurrency.in_greenlet(): + # in a greenlet; this is the connection was invalidated + # case. + try: + # try to gracefully close; see #10717 + # timeout added in asyncpg 0.14.0 December 2017 + self.await_(self._connection.close(timeout=2)) + except ( + asyncio.TimeoutError, + OSError, + self.dbapi.asyncpg.PostgresError, + ): + # in the case where we are recycling an old connection + # that may have already been disconnected, close() will + # fail with the above timeout. in this case, terminate + # the connection without any further waiting. + # see issue #8419 + self._connection.terminate() + else: + # not in a greenlet; this is the gc cleanup case + self._connection.terminate() + self._started = False + + @staticmethod + def _default_name_func(): + return None + + +class AsyncAdaptFallback_asyncpg_connection(AsyncAdapt_asyncpg_connection): + __slots__ = () + + await_ = staticmethod(await_fallback) + + +class AsyncAdapt_asyncpg_dbapi: + def __init__(self, asyncpg): + self.asyncpg = asyncpg + self.paramstyle = "numeric_dollar" + + def connect(self, *arg, **kw): + async_fallback = kw.pop("async_fallback", False) + creator_fn = kw.pop("async_creator_fn", self.asyncpg.connect) + prepared_statement_cache_size = kw.pop( + "prepared_statement_cache_size", 100 + ) + prepared_statement_name_func = kw.pop( + "prepared_statement_name_func", None + ) + + if util.asbool(async_fallback): + return AsyncAdaptFallback_asyncpg_connection( + self, + await_fallback(creator_fn(*arg, **kw)), + prepared_statement_cache_size=prepared_statement_cache_size, + prepared_statement_name_func=prepared_statement_name_func, + ) + else: + return AsyncAdapt_asyncpg_connection( + self, + await_only(creator_fn(*arg, **kw)), + prepared_statement_cache_size=prepared_statement_cache_size, + prepared_statement_name_func=prepared_statement_name_func, + ) + + class Error(Exception): + pass + + class Warning(Exception): # noqa + pass + + class InterfaceError(Error): + pass + + class DatabaseError(Error): + pass + + class InternalError(DatabaseError): + pass + + class OperationalError(DatabaseError): + pass + + class ProgrammingError(DatabaseError): + pass + + class IntegrityError(DatabaseError): + pass + + class DataError(DatabaseError): + pass + + class NotSupportedError(DatabaseError): + pass + + class InternalServerError(InternalError): + pass + + class InvalidCachedStatementError(NotSupportedError): + def __init__(self, message): + super().__init__( + message + " (SQLAlchemy asyncpg dialect will now invalidate " + "all prepared caches in response to this exception)", + ) + + # pep-249 datatype placeholders. As of SQLAlchemy 2.0 these aren't + # used, however the test suite looks for these in a few cases. + STRING = util.symbol("STRING") + NUMBER = util.symbol("NUMBER") + DATETIME = util.symbol("DATETIME") + + @util.memoized_property + def _asyncpg_error_translate(self): + import asyncpg + + return { + asyncpg.exceptions.IntegrityConstraintViolationError: self.IntegrityError, # noqa: E501 + asyncpg.exceptions.PostgresError: self.Error, + asyncpg.exceptions.SyntaxOrAccessError: self.ProgrammingError, + asyncpg.exceptions.InterfaceError: self.InterfaceError, + asyncpg.exceptions.InvalidCachedStatementError: self.InvalidCachedStatementError, # noqa: E501 + asyncpg.exceptions.InternalServerError: self.InternalServerError, + } + + def Binary(self, value): + return value + + +class PGDialect_asyncpg(PGDialect): + driver = "asyncpg" + supports_statement_cache = True + + supports_server_side_cursors = True + + render_bind_cast = True + has_terminate = True + + default_paramstyle = "numeric_dollar" + supports_sane_multi_rowcount = False + execution_ctx_cls = PGExecutionContext_asyncpg + statement_compiler = PGCompiler_asyncpg + preparer = PGIdentifierPreparer_asyncpg + + colspecs = util.update_copy( + PGDialect.colspecs, + { + sqltypes.String: AsyncpgString, + sqltypes.ARRAY: AsyncpgARRAY, + BIT: AsyncpgBit, + CITEXT: CITEXT, + REGCONFIG: AsyncpgREGCONFIG, + sqltypes.Time: AsyncpgTime, + sqltypes.Date: AsyncpgDate, + sqltypes.DateTime: AsyncpgDateTime, + sqltypes.Interval: AsyncPgInterval, + INTERVAL: AsyncPgInterval, + sqltypes.Boolean: AsyncpgBoolean, + sqltypes.Integer: AsyncpgInteger, + sqltypes.BigInteger: AsyncpgBigInteger, + sqltypes.Numeric: AsyncpgNumeric, + sqltypes.Float: AsyncpgFloat, + sqltypes.JSON: AsyncpgJSON, + sqltypes.LargeBinary: AsyncpgByteA, + json.JSONB: AsyncpgJSONB, + sqltypes.JSON.JSONPathType: AsyncpgJSONPathType, + sqltypes.JSON.JSONIndexType: AsyncpgJSONIndexType, + sqltypes.JSON.JSONIntIndexType: AsyncpgJSONIntIndexType, + sqltypes.JSON.JSONStrIndexType: AsyncpgJSONStrIndexType, + sqltypes.Enum: AsyncPgEnum, + OID: AsyncpgOID, + REGCLASS: AsyncpgREGCLASS, + sqltypes.CHAR: AsyncpgCHAR, + ranges.AbstractSingleRange: _AsyncpgRange, + ranges.AbstractMultiRange: _AsyncpgMultiRange, + }, + ) + is_async = True + _invalidate_schema_cache_asof = 0 + + def _invalidate_schema_cache(self): + self._invalidate_schema_cache_asof = time.time() + + @util.memoized_property + def _dbapi_version(self): + if self.dbapi and hasattr(self.dbapi, "__version__"): + return tuple( + [ + int(x) + for x in re.findall( + r"(\d+)(?:[-\.]?|$)", self.dbapi.__version__ + ) + ] + ) + else: + return (99, 99, 99) + + @classmethod + def import_dbapi(cls): + return AsyncAdapt_asyncpg_dbapi(__import__("asyncpg")) + + @util.memoized_property + def _isolation_lookup(self): + return { + "AUTOCOMMIT": "autocommit", + "READ COMMITTED": "read_committed", + "REPEATABLE READ": "repeatable_read", + "SERIALIZABLE": "serializable", + } + + def get_isolation_level_values(self, dbapi_connection): + return list(self._isolation_lookup) + + def set_isolation_level(self, dbapi_connection, level): + dbapi_connection.set_isolation_level(self._isolation_lookup[level]) + + def set_readonly(self, connection, value): + connection.readonly = value + + def get_readonly(self, connection): + return connection.readonly + + def set_deferrable(self, connection, value): + connection.deferrable = value + + def get_deferrable(self, connection): + return connection.deferrable + + def do_terminate(self, dbapi_connection) -> None: + dbapi_connection.terminate() + + def create_connect_args(self, url): + opts = url.translate_connect_args(username="user") + multihosts, multiports = self._split_multihost_from_url(url) + + opts.update(url.query) + + if multihosts: + assert multiports + if len(multihosts) == 1: + opts["host"] = multihosts[0] + if multiports[0] is not None: + opts["port"] = multiports[0] + elif not all(multihosts): + raise exc.ArgumentError( + "All hosts are required to be present" + " for asyncpg multiple host URL" + ) + elif not all(multiports): + raise exc.ArgumentError( + "All ports are required to be present" + " for asyncpg multiple host URL" + ) + else: + opts["host"] = list(multihosts) + opts["port"] = list(multiports) + else: + util.coerce_kw_type(opts, "port", int) + util.coerce_kw_type(opts, "prepared_statement_cache_size", int) + return ([], opts) + + def do_ping(self, dbapi_connection): + dbapi_connection.ping() + return True + + @classmethod + def get_pool_class(cls, url): + async_fallback = url.query.get("async_fallback", False) + + if util.asbool(async_fallback): + return pool.FallbackAsyncAdaptedQueuePool + else: + return pool.AsyncAdaptedQueuePool + + def is_disconnect(self, e, connection, cursor): + if connection: + return connection._connection.is_closed() + else: + return isinstance( + e, self.dbapi.InterfaceError + ) and "connection is closed" in str(e) + + async def setup_asyncpg_json_codec(self, conn): + """set up JSON codec for asyncpg. + + This occurs for all new connections and + can be overridden by third party dialects. + + .. versionadded:: 1.4.27 + + """ + + asyncpg_connection = conn._connection + deserializer = self._json_deserializer or _py_json.loads + + def _json_decoder(bin_value): + return deserializer(bin_value.decode()) + + await asyncpg_connection.set_type_codec( + "json", + encoder=str.encode, + decoder=_json_decoder, + schema="pg_catalog", + format="binary", + ) + + async def setup_asyncpg_jsonb_codec(self, conn): + """set up JSONB codec for asyncpg. + + This occurs for all new connections and + can be overridden by third party dialects. + + .. versionadded:: 1.4.27 + + """ + + asyncpg_connection = conn._connection + deserializer = self._json_deserializer or _py_json.loads + + def _jsonb_encoder(str_value): + # \x01 is the prefix for jsonb used by PostgreSQL. + # asyncpg requires it when format='binary' + return b"\x01" + str_value.encode() + + deserializer = self._json_deserializer or _py_json.loads + + def _jsonb_decoder(bin_value): + # the byte is the \x01 prefix for jsonb used by PostgreSQL. + # asyncpg returns it when format='binary' + return deserializer(bin_value[1:].decode()) + + await asyncpg_connection.set_type_codec( + "jsonb", + encoder=_jsonb_encoder, + decoder=_jsonb_decoder, + schema="pg_catalog", + format="binary", + ) + + async def _disable_asyncpg_inet_codecs(self, conn): + asyncpg_connection = conn._connection + + await asyncpg_connection.set_type_codec( + "inet", + encoder=lambda s: s, + decoder=lambda s: s, + schema="pg_catalog", + format="text", + ) + + await asyncpg_connection.set_type_codec( + "cidr", + encoder=lambda s: s, + decoder=lambda s: s, + schema="pg_catalog", + format="text", + ) + + def on_connect(self): + """on_connect for asyncpg + + A major component of this for asyncpg is to set up type decoders at the + asyncpg level. + + See https://github.com/MagicStack/asyncpg/issues/623 for + notes on JSON/JSONB implementation. + + """ + + super_connect = super().on_connect() + + def connect(conn): + conn.await_(self.setup_asyncpg_json_codec(conn)) + conn.await_(self.setup_asyncpg_jsonb_codec(conn)) + + if self._native_inet_types is False: + conn.await_(self._disable_asyncpg_inet_codecs(conn)) + if super_connect is not None: + super_connect(conn) + + return connect + + def get_driver_connection(self, connection): + return connection._connection + + +dialect = PGDialect_asyncpg diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/base.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/base.py new file mode 100644 index 0000000..4ab3ca2 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/base.py @@ -0,0 +1,5007 @@ +# dialects/postgresql/base.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql + :name: PostgreSQL + :full_support: 12, 13, 14, 15 + :normal_support: 9.6+ + :best_effort: 9+ + +.. _postgresql_sequences: + +Sequences/SERIAL/IDENTITY +------------------------- + +PostgreSQL supports sequences, and SQLAlchemy uses these as the default means +of creating new primary key values for integer-based primary key columns. When +creating tables, SQLAlchemy will issue the ``SERIAL`` datatype for +integer-based primary key columns, which generates a sequence and server side +default corresponding to the column. + +To specify a specific named sequence to be used for primary key generation, +use the :func:`~sqlalchemy.schema.Sequence` construct:: + + Table( + "sometable", + metadata, + Column( + "id", Integer, Sequence("some_id_seq", start=1), primary_key=True + ) + ) + +When SQLAlchemy issues a single INSERT statement, to fulfill the contract of +having the "last insert identifier" available, a RETURNING clause is added to +the INSERT statement which specifies the primary key columns should be +returned after the statement completes. The RETURNING functionality only takes +place if PostgreSQL 8.2 or later is in use. As a fallback approach, the +sequence, whether specified explicitly or implicitly via ``SERIAL``, is +executed independently beforehand, the returned value to be used in the +subsequent insert. Note that when an +:func:`~sqlalchemy.sql.expression.insert()` construct is executed using +"executemany" semantics, the "last inserted identifier" functionality does not +apply; no RETURNING clause is emitted nor is the sequence pre-executed in this +case. + + +PostgreSQL 10 and above IDENTITY columns +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +PostgreSQL 10 and above have a new IDENTITY feature that supersedes the use +of SERIAL. The :class:`_schema.Identity` construct in a +:class:`_schema.Column` can be used to control its behavior:: + + from sqlalchemy import Table, Column, MetaData, Integer, Computed + + metadata = MetaData() + + data = Table( + "data", + metadata, + Column( + 'id', Integer, Identity(start=42, cycle=True), primary_key=True + ), + Column('data', String) + ) + +The CREATE TABLE for the above :class:`_schema.Table` object would be: + +.. sourcecode:: sql + + CREATE TABLE data ( + id INTEGER GENERATED BY DEFAULT AS IDENTITY (START WITH 42 CYCLE), + data VARCHAR, + PRIMARY KEY (id) + ) + +.. versionchanged:: 1.4 Added :class:`_schema.Identity` construct + in a :class:`_schema.Column` to specify the option of an autoincrementing + column. + +.. note:: + + Previous versions of SQLAlchemy did not have built-in support for rendering + of IDENTITY, and could use the following compilation hook to replace + occurrences of SERIAL with IDENTITY:: + + from sqlalchemy.schema import CreateColumn + from sqlalchemy.ext.compiler import compiles + + + @compiles(CreateColumn, 'postgresql') + def use_identity(element, compiler, **kw): + text = compiler.visit_create_column(element, **kw) + text = text.replace( + "SERIAL", "INT GENERATED BY DEFAULT AS IDENTITY" + ) + return text + + Using the above, a table such as:: + + t = Table( + 't', m, + Column('id', Integer, primary_key=True), + Column('data', String) + ) + + Will generate on the backing database as:: + + CREATE TABLE t ( + id INT GENERATED BY DEFAULT AS IDENTITY, + data VARCHAR, + PRIMARY KEY (id) + ) + +.. _postgresql_ss_cursors: + +Server Side Cursors +------------------- + +Server-side cursor support is available for the psycopg2, asyncpg +dialects and may also be available in others. + +Server side cursors are enabled on a per-statement basis by using the +:paramref:`.Connection.execution_options.stream_results` connection execution +option:: + + with engine.connect() as conn: + result = conn.execution_options(stream_results=True).execute(text("select * from table")) + +Note that some kinds of SQL statements may not be supported with +server side cursors; generally, only SQL statements that return rows should be +used with this option. + +.. deprecated:: 1.4 The dialect-level server_side_cursors flag is deprecated + and will be removed in a future release. Please use the + :paramref:`_engine.Connection.stream_results` execution option for + unbuffered cursor support. + +.. seealso:: + + :ref:`engine_stream_results` + +.. _postgresql_isolation_level: + +Transaction Isolation Level +--------------------------- + +Most SQLAlchemy dialects support setting of transaction isolation level +using the :paramref:`_sa.create_engine.isolation_level` parameter +at the :func:`_sa.create_engine` level, and at the :class:`_engine.Connection` +level via the :paramref:`.Connection.execution_options.isolation_level` +parameter. + +For PostgreSQL dialects, this feature works either by making use of the +DBAPI-specific features, such as psycopg2's isolation level flags which will +embed the isolation level setting inline with the ``"BEGIN"`` statement, or for +DBAPIs with no direct support by emitting ``SET SESSION CHARACTERISTICS AS +TRANSACTION ISOLATION LEVEL <level>`` ahead of the ``"BEGIN"`` statement +emitted by the DBAPI. For the special AUTOCOMMIT isolation level, +DBAPI-specific techniques are used which is typically an ``.autocommit`` +flag on the DBAPI connection object. + +To set isolation level using :func:`_sa.create_engine`:: + + engine = create_engine( + "postgresql+pg8000://scott:tiger@localhost/test", + isolation_level = "REPEATABLE READ" + ) + +To set using per-connection execution options:: + + with engine.connect() as conn: + conn = conn.execution_options( + isolation_level="REPEATABLE READ" + ) + with conn.begin(): + # ... work with transaction + +There are also more options for isolation level configurations, such as +"sub-engine" objects linked to a main :class:`_engine.Engine` which each apply +different isolation level settings. See the discussion at +:ref:`dbapi_autocommit` for background. + +Valid values for ``isolation_level`` on most PostgreSQL dialects include: + +* ``READ COMMITTED`` +* ``READ UNCOMMITTED`` +* ``REPEATABLE READ`` +* ``SERIALIZABLE`` +* ``AUTOCOMMIT`` + +.. seealso:: + + :ref:`dbapi_autocommit` + + :ref:`postgresql_readonly_deferrable` + + :ref:`psycopg2_isolation_level` + + :ref:`pg8000_isolation_level` + +.. _postgresql_readonly_deferrable: + +Setting READ ONLY / DEFERRABLE +------------------------------ + +Most PostgreSQL dialects support setting the "READ ONLY" and "DEFERRABLE" +characteristics of the transaction, which is in addition to the isolation level +setting. These two attributes can be established either in conjunction with or +independently of the isolation level by passing the ``postgresql_readonly`` and +``postgresql_deferrable`` flags with +:meth:`_engine.Connection.execution_options`. The example below illustrates +passing the ``"SERIALIZABLE"`` isolation level at the same time as setting +"READ ONLY" and "DEFERRABLE":: + + with engine.connect() as conn: + conn = conn.execution_options( + isolation_level="SERIALIZABLE", + postgresql_readonly=True, + postgresql_deferrable=True + ) + with conn.begin(): + # ... work with transaction + +Note that some DBAPIs such as asyncpg only support "readonly" with +SERIALIZABLE isolation. + +.. versionadded:: 1.4 added support for the ``postgresql_readonly`` + and ``postgresql_deferrable`` execution options. + +.. _postgresql_reset_on_return: + +Temporary Table / Resource Reset for Connection Pooling +------------------------------------------------------- + +The :class:`.QueuePool` connection pool implementation used +by the SQLAlchemy :class:`.Engine` object includes +:ref:`reset on return <pool_reset_on_return>` behavior that will invoke +the DBAPI ``.rollback()`` method when connections are returned to the pool. +While this rollback will clear out the immediate state used by the previous +transaction, it does not cover a wider range of session-level state, including +temporary tables as well as other server state such as prepared statement +handles and statement caches. The PostgreSQL database includes a variety +of commands which may be used to reset this state, including +``DISCARD``, ``RESET``, ``DEALLOCATE``, and ``UNLISTEN``. + + +To install +one or more of these commands as the means of performing reset-on-return, +the :meth:`.PoolEvents.reset` event hook may be used, as demonstrated +in the example below. The implementation +will end transactions in progress as well as discard temporary tables +using the ``CLOSE``, ``RESET`` and ``DISCARD`` commands; see the PostgreSQL +documentation for background on what each of these statements do. + +The :paramref:`_sa.create_engine.pool_reset_on_return` parameter +is set to ``None`` so that the custom scheme can replace the default behavior +completely. The custom hook implementation calls ``.rollback()`` in any case, +as it's usually important that the DBAPI's own tracking of commit/rollback +will remain consistent with the state of the transaction:: + + + from sqlalchemy import create_engine + from sqlalchemy import event + + postgresql_engine = create_engine( + "postgresql+pyscopg2://scott:tiger@hostname/dbname", + + # disable default reset-on-return scheme + pool_reset_on_return=None, + ) + + + @event.listens_for(postgresql_engine, "reset") + def _reset_postgresql(dbapi_connection, connection_record, reset_state): + if not reset_state.terminate_only: + dbapi_connection.execute("CLOSE ALL") + dbapi_connection.execute("RESET ALL") + dbapi_connection.execute("DISCARD TEMP") + + # so that the DBAPI itself knows that the connection has been + # reset + dbapi_connection.rollback() + +.. versionchanged:: 2.0.0b3 Added additional state arguments to + the :meth:`.PoolEvents.reset` event and additionally ensured the event + is invoked for all "reset" occurrences, so that it's appropriate + as a place for custom "reset" handlers. Previous schemes which + use the :meth:`.PoolEvents.checkin` handler remain usable as well. + +.. seealso:: + + :ref:`pool_reset_on_return` - in the :ref:`pooling_toplevel` documentation + +.. _postgresql_alternate_search_path: + +Setting Alternate Search Paths on Connect +------------------------------------------ + +The PostgreSQL ``search_path`` variable refers to the list of schema names +that will be implicitly referenced when a particular table or other +object is referenced in a SQL statement. As detailed in the next section +:ref:`postgresql_schema_reflection`, SQLAlchemy is generally organized around +the concept of keeping this variable at its default value of ``public``, +however, in order to have it set to any arbitrary name or names when connections +are used automatically, the "SET SESSION search_path" command may be invoked +for all connections in a pool using the following event handler, as discussed +at :ref:`schema_set_default_connections`:: + + from sqlalchemy import event + from sqlalchemy import create_engine + + engine = create_engine("postgresql+psycopg2://scott:tiger@host/dbname") + + @event.listens_for(engine, "connect", insert=True) + def set_search_path(dbapi_connection, connection_record): + existing_autocommit = dbapi_connection.autocommit + dbapi_connection.autocommit = True + cursor = dbapi_connection.cursor() + cursor.execute("SET SESSION search_path='%s'" % schema_name) + cursor.close() + dbapi_connection.autocommit = existing_autocommit + +The reason the recipe is complicated by use of the ``.autocommit`` DBAPI +attribute is so that when the ``SET SESSION search_path`` directive is invoked, +it is invoked outside of the scope of any transaction and therefore will not +be reverted when the DBAPI connection has a rollback. + +.. seealso:: + + :ref:`schema_set_default_connections` - in the :ref:`metadata_toplevel` documentation + + + + +.. _postgresql_schema_reflection: + +Remote-Schema Table Introspection and PostgreSQL search_path +------------------------------------------------------------ + +.. admonition:: Section Best Practices Summarized + + keep the ``search_path`` variable set to its default of ``public``, without + any other schema names. Ensure the username used to connect **does not** + match remote schemas, or ensure the ``"$user"`` token is **removed** from + ``search_path``. For other schema names, name these explicitly + within :class:`_schema.Table` definitions. Alternatively, the + ``postgresql_ignore_search_path`` option will cause all reflected + :class:`_schema.Table` objects to have a :attr:`_schema.Table.schema` + attribute set up. + +The PostgreSQL dialect can reflect tables from any schema, as outlined in +:ref:`metadata_reflection_schemas`. + +In all cases, the first thing SQLAlchemy does when reflecting tables is +to **determine the default schema for the current database connection**. +It does this using the PostgreSQL ``current_schema()`` +function, illustated below using a PostgreSQL client session (i.e. using +the ``psql`` tool):: + + test=> select current_schema(); + current_schema + ---------------- + public + (1 row) + +Above we see that on a plain install of PostgreSQL, the default schema name +is the name ``public``. + +However, if your database username **matches the name of a schema**, PostgreSQL's +default is to then **use that name as the default schema**. Below, we log in +using the username ``scott``. When we create a schema named ``scott``, **it +implicitly changes the default schema**:: + + test=> select current_schema(); + current_schema + ---------------- + public + (1 row) + + test=> create schema scott; + CREATE SCHEMA + test=> select current_schema(); + current_schema + ---------------- + scott + (1 row) + +The behavior of ``current_schema()`` is derived from the +`PostgreSQL search path +<https://www.postgresql.org/docs/current/static/ddl-schemas.html#DDL-SCHEMAS-PATH>`_ +variable ``search_path``, which in modern PostgreSQL versions defaults to this:: + + test=> show search_path; + search_path + ----------------- + "$user", public + (1 row) + +Where above, the ``"$user"`` variable will inject the current username as the +default schema, if one exists. Otherwise, ``public`` is used. + +When a :class:`_schema.Table` object is reflected, if it is present in the +schema indicated by the ``current_schema()`` function, **the schema name assigned +to the ".schema" attribute of the Table is the Python "None" value**. Otherwise, the +".schema" attribute will be assigned the string name of that schema. + +With regards to tables which these :class:`_schema.Table` +objects refer to via foreign key constraint, a decision must be made as to how +the ``.schema`` is represented in those remote tables, in the case where that +remote schema name is also a member of the current ``search_path``. + +By default, the PostgreSQL dialect mimics the behavior encouraged by +PostgreSQL's own ``pg_get_constraintdef()`` builtin procedure. This function +returns a sample definition for a particular foreign key constraint, +omitting the referenced schema name from that definition when the name is +also in the PostgreSQL schema search path. The interaction below +illustrates this behavior:: + + test=> CREATE TABLE test_schema.referred(id INTEGER PRIMARY KEY); + CREATE TABLE + test=> CREATE TABLE referring( + test(> id INTEGER PRIMARY KEY, + test(> referred_id INTEGER REFERENCES test_schema.referred(id)); + CREATE TABLE + test=> SET search_path TO public, test_schema; + test=> SELECT pg_catalog.pg_get_constraintdef(r.oid, true) FROM + test-> pg_catalog.pg_class c JOIN pg_catalog.pg_namespace n + test-> ON n.oid = c.relnamespace + test-> JOIN pg_catalog.pg_constraint r ON c.oid = r.conrelid + test-> WHERE c.relname='referring' AND r.contype = 'f' + test-> ; + pg_get_constraintdef + --------------------------------------------------- + FOREIGN KEY (referred_id) REFERENCES referred(id) + (1 row) + +Above, we created a table ``referred`` as a member of the remote schema +``test_schema``, however when we added ``test_schema`` to the +PG ``search_path`` and then asked ``pg_get_constraintdef()`` for the +``FOREIGN KEY`` syntax, ``test_schema`` was not included in the output of +the function. + +On the other hand, if we set the search path back to the typical default +of ``public``:: + + test=> SET search_path TO public; + SET + +The same query against ``pg_get_constraintdef()`` now returns the fully +schema-qualified name for us:: + + test=> SELECT pg_catalog.pg_get_constraintdef(r.oid, true) FROM + test-> pg_catalog.pg_class c JOIN pg_catalog.pg_namespace n + test-> ON n.oid = c.relnamespace + test-> JOIN pg_catalog.pg_constraint r ON c.oid = r.conrelid + test-> WHERE c.relname='referring' AND r.contype = 'f'; + pg_get_constraintdef + --------------------------------------------------------------- + FOREIGN KEY (referred_id) REFERENCES test_schema.referred(id) + (1 row) + +SQLAlchemy will by default use the return value of ``pg_get_constraintdef()`` +in order to determine the remote schema name. That is, if our ``search_path`` +were set to include ``test_schema``, and we invoked a table +reflection process as follows:: + + >>> from sqlalchemy import Table, MetaData, create_engine, text + >>> engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test") + >>> with engine.connect() as conn: + ... conn.execute(text("SET search_path TO test_schema, public")) + ... metadata_obj = MetaData() + ... referring = Table('referring', metadata_obj, + ... autoload_with=conn) + ... + <sqlalchemy.engine.result.CursorResult object at 0x101612ed0> + +The above process would deliver to the :attr:`_schema.MetaData.tables` +collection +``referred`` table named **without** the schema:: + + >>> metadata_obj.tables['referred'].schema is None + True + +To alter the behavior of reflection such that the referred schema is +maintained regardless of the ``search_path`` setting, use the +``postgresql_ignore_search_path`` option, which can be specified as a +dialect-specific argument to both :class:`_schema.Table` as well as +:meth:`_schema.MetaData.reflect`:: + + >>> with engine.connect() as conn: + ... conn.execute(text("SET search_path TO test_schema, public")) + ... metadata_obj = MetaData() + ... referring = Table('referring', metadata_obj, + ... autoload_with=conn, + ... postgresql_ignore_search_path=True) + ... + <sqlalchemy.engine.result.CursorResult object at 0x1016126d0> + +We will now have ``test_schema.referred`` stored as schema-qualified:: + + >>> metadata_obj.tables['test_schema.referred'].schema + 'test_schema' + +.. sidebar:: Best Practices for PostgreSQL Schema reflection + + The description of PostgreSQL schema reflection behavior is complex, and + is the product of many years of dealing with widely varied use cases and + user preferences. But in fact, there's no need to understand any of it if + you just stick to the simplest use pattern: leave the ``search_path`` set + to its default of ``public`` only, never refer to the name ``public`` as + an explicit schema name otherwise, and refer to all other schema names + explicitly when building up a :class:`_schema.Table` object. The options + described here are only for those users who can't, or prefer not to, stay + within these guidelines. + +.. seealso:: + + :ref:`reflection_schema_qualified_interaction` - discussion of the issue + from a backend-agnostic perspective + + `The Schema Search Path + <https://www.postgresql.org/docs/current/static/ddl-schemas.html#DDL-SCHEMAS-PATH>`_ + - on the PostgreSQL website. + +INSERT/UPDATE...RETURNING +------------------------- + +The dialect supports PG 8.2's ``INSERT..RETURNING``, ``UPDATE..RETURNING`` and +``DELETE..RETURNING`` syntaxes. ``INSERT..RETURNING`` is used by default +for single-row INSERT statements in order to fetch newly generated +primary key identifiers. To specify an explicit ``RETURNING`` clause, +use the :meth:`._UpdateBase.returning` method on a per-statement basis:: + + # INSERT..RETURNING + result = table.insert().returning(table.c.col1, table.c.col2).\ + values(name='foo') + print(result.fetchall()) + + # UPDATE..RETURNING + result = table.update().returning(table.c.col1, table.c.col2).\ + where(table.c.name=='foo').values(name='bar') + print(result.fetchall()) + + # DELETE..RETURNING + result = table.delete().returning(table.c.col1, table.c.col2).\ + where(table.c.name=='foo') + print(result.fetchall()) + +.. _postgresql_insert_on_conflict: + +INSERT...ON CONFLICT (Upsert) +------------------------------ + +Starting with version 9.5, PostgreSQL allows "upserts" (update or insert) of +rows into a table via the ``ON CONFLICT`` clause of the ``INSERT`` statement. A +candidate row will only be inserted if that row does not violate any unique +constraints. In the case of a unique constraint violation, a secondary action +can occur which can be either "DO UPDATE", indicating that the data in the +target row should be updated, or "DO NOTHING", which indicates to silently skip +this row. + +Conflicts are determined using existing unique constraints and indexes. These +constraints may be identified either using their name as stated in DDL, +or they may be inferred by stating the columns and conditions that comprise +the indexes. + +SQLAlchemy provides ``ON CONFLICT`` support via the PostgreSQL-specific +:func:`_postgresql.insert()` function, which provides +the generative methods :meth:`_postgresql.Insert.on_conflict_do_update` +and :meth:`~.postgresql.Insert.on_conflict_do_nothing`: + +.. sourcecode:: pycon+sql + + >>> from sqlalchemy.dialects.postgresql import insert + >>> insert_stmt = insert(my_table).values( + ... id='some_existing_id', + ... data='inserted value') + >>> do_nothing_stmt = insert_stmt.on_conflict_do_nothing( + ... index_elements=['id'] + ... ) + >>> print(do_nothing_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO NOTHING + {stop} + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... constraint='pk_my_table', + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT ON CONSTRAINT pk_my_table DO UPDATE SET data = %(param_1)s + +.. seealso:: + + `INSERT .. ON CONFLICT + <https://www.postgresql.org/docs/current/static/sql-insert.html#SQL-ON-CONFLICT>`_ + - in the PostgreSQL documentation. + +Specifying the Target +^^^^^^^^^^^^^^^^^^^^^ + +Both methods supply the "target" of the conflict using either the +named constraint or by column inference: + +* The :paramref:`_postgresql.Insert.on_conflict_do_update.index_elements` argument + specifies a sequence containing string column names, :class:`_schema.Column` + objects, and/or SQL expression elements, which would identify a unique + index: + + .. sourcecode:: pycon+sql + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... index_elements=['id'], + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s + {stop} + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... index_elements=[my_table.c.id], + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s + +* When using :paramref:`_postgresql.Insert.on_conflict_do_update.index_elements` to + infer an index, a partial index can be inferred by also specifying the + use the :paramref:`_postgresql.Insert.on_conflict_do_update.index_where` parameter: + + .. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values(user_email='a@b.com', data='inserted data') + >>> stmt = stmt.on_conflict_do_update( + ... index_elements=[my_table.c.user_email], + ... index_where=my_table.c.user_email.like('%@gmail.com'), + ... set_=dict(data=stmt.excluded.data) + ... ) + >>> print(stmt) + {printsql}INSERT INTO my_table (data, user_email) + VALUES (%(data)s, %(user_email)s) ON CONFLICT (user_email) + WHERE user_email LIKE %(user_email_1)s DO UPDATE SET data = excluded.data + +* The :paramref:`_postgresql.Insert.on_conflict_do_update.constraint` argument is + used to specify an index directly rather than inferring it. This can be + the name of a UNIQUE constraint, a PRIMARY KEY constraint, or an INDEX: + + .. sourcecode:: pycon+sql + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... constraint='my_table_idx_1', + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT ON CONSTRAINT my_table_idx_1 DO UPDATE SET data = %(param_1)s + {stop} + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... constraint='my_table_pk', + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT ON CONSTRAINT my_table_pk DO UPDATE SET data = %(param_1)s + {stop} + +* The :paramref:`_postgresql.Insert.on_conflict_do_update.constraint` argument may + also refer to a SQLAlchemy construct representing a constraint, + e.g. :class:`.UniqueConstraint`, :class:`.PrimaryKeyConstraint`, + :class:`.Index`, or :class:`.ExcludeConstraint`. In this use, + if the constraint has a name, it is used directly. Otherwise, if the + constraint is unnamed, then inference will be used, where the expressions + and optional WHERE clause of the constraint will be spelled out in the + construct. This use is especially convenient + to refer to the named or unnamed primary key of a :class:`_schema.Table` + using the + :attr:`_schema.Table.primary_key` attribute: + + .. sourcecode:: pycon+sql + + >>> do_update_stmt = insert_stmt.on_conflict_do_update( + ... constraint=my_table.primary_key, + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s + +The SET Clause +^^^^^^^^^^^^^^^ + +``ON CONFLICT...DO UPDATE`` is used to perform an update of the already +existing row, using any combination of new values as well as values +from the proposed insertion. These values are specified using the +:paramref:`_postgresql.Insert.on_conflict_do_update.set_` parameter. This +parameter accepts a dictionary which consists of direct values +for UPDATE: + +.. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values(id='some_id', data='inserted value') + >>> do_update_stmt = stmt.on_conflict_do_update( + ... index_elements=['id'], + ... set_=dict(data='updated value') + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s + +.. warning:: + + The :meth:`_expression.Insert.on_conflict_do_update` + method does **not** take into + account Python-side default UPDATE values or generation functions, e.g. + those specified using :paramref:`_schema.Column.onupdate`. + These values will not be exercised for an ON CONFLICT style of UPDATE, + unless they are manually specified in the + :paramref:`_postgresql.Insert.on_conflict_do_update.set_` dictionary. + +Updating using the Excluded INSERT Values +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +In order to refer to the proposed insertion row, the special alias +:attr:`~.postgresql.Insert.excluded` is available as an attribute on +the :class:`_postgresql.Insert` object; this object is a +:class:`_expression.ColumnCollection` +which alias contains all columns of the target +table: + +.. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values( + ... id='some_id', + ... data='inserted value', + ... author='jlh' + ... ) + >>> do_update_stmt = stmt.on_conflict_do_update( + ... index_elements=['id'], + ... set_=dict(data='updated value', author=stmt.excluded.author) + ... ) + >>> print(do_update_stmt) + {printsql}INSERT INTO my_table (id, data, author) + VALUES (%(id)s, %(data)s, %(author)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s, author = excluded.author + +Additional WHERE Criteria +^^^^^^^^^^^^^^^^^^^^^^^^^ + +The :meth:`_expression.Insert.on_conflict_do_update` method also accepts +a WHERE clause using the :paramref:`_postgresql.Insert.on_conflict_do_update.where` +parameter, which will limit those rows which receive an UPDATE: + +.. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values( + ... id='some_id', + ... data='inserted value', + ... author='jlh' + ... ) + >>> on_update_stmt = stmt.on_conflict_do_update( + ... index_elements=['id'], + ... set_=dict(data='updated value', author=stmt.excluded.author), + ... where=(my_table.c.status == 2) + ... ) + >>> print(on_update_stmt) + {printsql}INSERT INTO my_table (id, data, author) + VALUES (%(id)s, %(data)s, %(author)s) + ON CONFLICT (id) DO UPDATE SET data = %(param_1)s, author = excluded.author + WHERE my_table.status = %(status_1)s + +Skipping Rows with DO NOTHING +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +``ON CONFLICT`` may be used to skip inserting a row entirely +if any conflict with a unique or exclusion constraint occurs; below +this is illustrated using the +:meth:`~.postgresql.Insert.on_conflict_do_nothing` method: + +.. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values(id='some_id', data='inserted value') + >>> stmt = stmt.on_conflict_do_nothing(index_elements=['id']) + >>> print(stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT (id) DO NOTHING + +If ``DO NOTHING`` is used without specifying any columns or constraint, +it has the effect of skipping the INSERT for any unique or exclusion +constraint violation which occurs: + +.. sourcecode:: pycon+sql + + >>> stmt = insert(my_table).values(id='some_id', data='inserted value') + >>> stmt = stmt.on_conflict_do_nothing() + >>> print(stmt) + {printsql}INSERT INTO my_table (id, data) VALUES (%(id)s, %(data)s) + ON CONFLICT DO NOTHING + +.. _postgresql_match: + +Full Text Search +---------------- + +PostgreSQL's full text search system is available through the use of the +:data:`.func` namespace, combined with the use of custom operators +via the :meth:`.Operators.bool_op` method. For simple cases with some +degree of cross-backend compatibility, the :meth:`.Operators.match` operator +may also be used. + +.. _postgresql_simple_match: + +Simple plain text matching with ``match()`` +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +The :meth:`.Operators.match` operator provides for cross-compatible simple +text matching. For the PostgreSQL backend, it's hardcoded to generate +an expression using the ``@@`` operator in conjunction with the +``plainto_tsquery()`` PostgreSQL function. + +On the PostgreSQL dialect, an expression like the following:: + + select(sometable.c.text.match("search string")) + +would emit to the database:: + + SELECT text @@ plainto_tsquery('search string') FROM table + +Above, passing a plain string to :meth:`.Operators.match` will automatically +make use of ``plainto_tsquery()`` to specify the type of tsquery. This +establishes basic database cross-compatibility for :meth:`.Operators.match` +with other backends. + +.. versionchanged:: 2.0 The default tsquery generation function used by the + PostgreSQL dialect with :meth:`.Operators.match` is ``plainto_tsquery()``. + + To render exactly what was rendered in 1.4, use the following form:: + + from sqlalchemy import func + + select( + sometable.c.text.bool_op("@@")(func.to_tsquery("search string")) + ) + + Which would emit:: + + SELECT text @@ to_tsquery('search string') FROM table + +Using PostgreSQL full text functions and operators directly +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +Text search operations beyond the simple use of :meth:`.Operators.match` +may make use of the :data:`.func` namespace to generate PostgreSQL full-text +functions, in combination with :meth:`.Operators.bool_op` to generate +any boolean operator. + +For example, the query:: + + select( + func.to_tsquery('cat').bool_op("@>")(func.to_tsquery('cat & rat')) + ) + +would generate: + +.. sourcecode:: sql + + SELECT to_tsquery('cat') @> to_tsquery('cat & rat') + + +The :class:`_postgresql.TSVECTOR` type can provide for explicit CAST:: + + from sqlalchemy.dialects.postgresql import TSVECTOR + from sqlalchemy import select, cast + select(cast("some text", TSVECTOR)) + +produces a statement equivalent to:: + + SELECT CAST('some text' AS TSVECTOR) AS anon_1 + +The ``func`` namespace is augmented by the PostgreSQL dialect to set up +correct argument and return types for most full text search functions. +These functions are used automatically by the :attr:`_sql.func` namespace +assuming the ``sqlalchemy.dialects.postgresql`` package has been imported, +or :func:`_sa.create_engine` has been invoked using a ``postgresql`` +dialect. These functions are documented at: + +* :class:`_postgresql.to_tsvector` +* :class:`_postgresql.to_tsquery` +* :class:`_postgresql.plainto_tsquery` +* :class:`_postgresql.phraseto_tsquery` +* :class:`_postgresql.websearch_to_tsquery` +* :class:`_postgresql.ts_headline` + +Specifying the "regconfig" with ``match()`` or custom operators +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +PostgreSQL's ``plainto_tsquery()`` function accepts an optional +"regconfig" argument that is used to instruct PostgreSQL to use a +particular pre-computed GIN or GiST index in order to perform the search. +When using :meth:`.Operators.match`, this additional parameter may be +specified using the ``postgresql_regconfig`` parameter, such as:: + + select(mytable.c.id).where( + mytable.c.title.match('somestring', postgresql_regconfig='english') + ) + +Which would emit:: + + SELECT mytable.id FROM mytable + WHERE mytable.title @@ plainto_tsquery('english', 'somestring') + +When using other PostgreSQL search functions with :data:`.func`, the +"regconfig" parameter may be passed directly as the initial argument:: + + select(mytable.c.id).where( + func.to_tsvector("english", mytable.c.title).bool_op("@@")( + func.to_tsquery("english", "somestring") + ) + ) + +produces a statement equivalent to:: + + SELECT mytable.id FROM mytable + WHERE to_tsvector('english', mytable.title) @@ + to_tsquery('english', 'somestring') + +It is recommended that you use the ``EXPLAIN ANALYZE...`` tool from +PostgreSQL to ensure that you are generating queries with SQLAlchemy that +take full advantage of any indexes you may have created for full text search. + +.. seealso:: + + `Full Text Search <https://www.postgresql.org/docs/current/textsearch-controls.html>`_ - in the PostgreSQL documentation + + +FROM ONLY ... +------------- + +The dialect supports PostgreSQL's ONLY keyword for targeting only a particular +table in an inheritance hierarchy. This can be used to produce the +``SELECT ... FROM ONLY``, ``UPDATE ONLY ...``, and ``DELETE FROM ONLY ...`` +syntaxes. It uses SQLAlchemy's hints mechanism:: + + # SELECT ... FROM ONLY ... + result = table.select().with_hint(table, 'ONLY', 'postgresql') + print(result.fetchall()) + + # UPDATE ONLY ... + table.update(values=dict(foo='bar')).with_hint('ONLY', + dialect_name='postgresql') + + # DELETE FROM ONLY ... + table.delete().with_hint('ONLY', dialect_name='postgresql') + + +.. _postgresql_indexes: + +PostgreSQL-Specific Index Options +--------------------------------- + +Several extensions to the :class:`.Index` construct are available, specific +to the PostgreSQL dialect. + +Covering Indexes +^^^^^^^^^^^^^^^^ + +The ``postgresql_include`` option renders INCLUDE(colname) for the given +string names:: + + Index("my_index", table.c.x, postgresql_include=['y']) + +would render the index as ``CREATE INDEX my_index ON table (x) INCLUDE (y)`` + +Note that this feature requires PostgreSQL 11 or later. + +.. versionadded:: 1.4 + +.. _postgresql_partial_indexes: + +Partial Indexes +^^^^^^^^^^^^^^^ + +Partial indexes add criterion to the index definition so that the index is +applied to a subset of rows. These can be specified on :class:`.Index` +using the ``postgresql_where`` keyword argument:: + + Index('my_index', my_table.c.id, postgresql_where=my_table.c.value > 10) + +.. _postgresql_operator_classes: + +Operator Classes +^^^^^^^^^^^^^^^^ + +PostgreSQL allows the specification of an *operator class* for each column of +an index (see +https://www.postgresql.org/docs/current/interactive/indexes-opclass.html). +The :class:`.Index` construct allows these to be specified via the +``postgresql_ops`` keyword argument:: + + Index( + 'my_index', my_table.c.id, my_table.c.data, + postgresql_ops={ + 'data': 'text_pattern_ops', + 'id': 'int4_ops' + }) + +Note that the keys in the ``postgresql_ops`` dictionaries are the +"key" name of the :class:`_schema.Column`, i.e. the name used to access it from +the ``.c`` collection of :class:`_schema.Table`, which can be configured to be +different than the actual name of the column as expressed in the database. + +If ``postgresql_ops`` is to be used against a complex SQL expression such +as a function call, then to apply to the column it must be given a label +that is identified in the dictionary by name, e.g.:: + + Index( + 'my_index', my_table.c.id, + func.lower(my_table.c.data).label('data_lower'), + postgresql_ops={ + 'data_lower': 'text_pattern_ops', + 'id': 'int4_ops' + }) + +Operator classes are also supported by the +:class:`_postgresql.ExcludeConstraint` construct using the +:paramref:`_postgresql.ExcludeConstraint.ops` parameter. See that parameter for +details. + +.. versionadded:: 1.3.21 added support for operator classes with + :class:`_postgresql.ExcludeConstraint`. + + +Index Types +^^^^^^^^^^^ + +PostgreSQL provides several index types: B-Tree, Hash, GiST, and GIN, as well +as the ability for users to create their own (see +https://www.postgresql.org/docs/current/static/indexes-types.html). These can be +specified on :class:`.Index` using the ``postgresql_using`` keyword argument:: + + Index('my_index', my_table.c.data, postgresql_using='gin') + +The value passed to the keyword argument will be simply passed through to the +underlying CREATE INDEX command, so it *must* be a valid index type for your +version of PostgreSQL. + +.. _postgresql_index_storage: + +Index Storage Parameters +^^^^^^^^^^^^^^^^^^^^^^^^ + +PostgreSQL allows storage parameters to be set on indexes. The storage +parameters available depend on the index method used by the index. Storage +parameters can be specified on :class:`.Index` using the ``postgresql_with`` +keyword argument:: + + Index('my_index', my_table.c.data, postgresql_with={"fillfactor": 50}) + +PostgreSQL allows to define the tablespace in which to create the index. +The tablespace can be specified on :class:`.Index` using the +``postgresql_tablespace`` keyword argument:: + + Index('my_index', my_table.c.data, postgresql_tablespace='my_tablespace') + +Note that the same option is available on :class:`_schema.Table` as well. + +.. _postgresql_index_concurrently: + +Indexes with CONCURRENTLY +^^^^^^^^^^^^^^^^^^^^^^^^^ + +The PostgreSQL index option CONCURRENTLY is supported by passing the +flag ``postgresql_concurrently`` to the :class:`.Index` construct:: + + tbl = Table('testtbl', m, Column('data', Integer)) + + idx1 = Index('test_idx1', tbl.c.data, postgresql_concurrently=True) + +The above index construct will render DDL for CREATE INDEX, assuming +PostgreSQL 8.2 or higher is detected or for a connection-less dialect, as:: + + CREATE INDEX CONCURRENTLY test_idx1 ON testtbl (data) + +For DROP INDEX, assuming PostgreSQL 9.2 or higher is detected or for +a connection-less dialect, it will emit:: + + DROP INDEX CONCURRENTLY test_idx1 + +When using CONCURRENTLY, the PostgreSQL database requires that the statement +be invoked outside of a transaction block. The Python DBAPI enforces that +even for a single statement, a transaction is present, so to use this +construct, the DBAPI's "autocommit" mode must be used:: + + metadata = MetaData() + table = Table( + "foo", metadata, + Column("id", String)) + index = Index( + "foo_idx", table.c.id, postgresql_concurrently=True) + + with engine.connect() as conn: + with conn.execution_options(isolation_level='AUTOCOMMIT'): + table.create(conn) + +.. seealso:: + + :ref:`postgresql_isolation_level` + +.. _postgresql_index_reflection: + +PostgreSQL Index Reflection +--------------------------- + +The PostgreSQL database creates a UNIQUE INDEX implicitly whenever the +UNIQUE CONSTRAINT construct is used. When inspecting a table using +:class:`_reflection.Inspector`, the :meth:`_reflection.Inspector.get_indexes` +and the :meth:`_reflection.Inspector.get_unique_constraints` +will report on these +two constructs distinctly; in the case of the index, the key +``duplicates_constraint`` will be present in the index entry if it is +detected as mirroring a constraint. When performing reflection using +``Table(..., autoload_with=engine)``, the UNIQUE INDEX is **not** returned +in :attr:`_schema.Table.indexes` when it is detected as mirroring a +:class:`.UniqueConstraint` in the :attr:`_schema.Table.constraints` collection +. + +Special Reflection Options +-------------------------- + +The :class:`_reflection.Inspector` +used for the PostgreSQL backend is an instance +of :class:`.PGInspector`, which offers additional methods:: + + from sqlalchemy import create_engine, inspect + + engine = create_engine("postgresql+psycopg2://localhost/test") + insp = inspect(engine) # will be a PGInspector + + print(insp.get_enums()) + +.. autoclass:: PGInspector + :members: + +.. _postgresql_table_options: + +PostgreSQL Table Options +------------------------ + +Several options for CREATE TABLE are supported directly by the PostgreSQL +dialect in conjunction with the :class:`_schema.Table` construct: + +* ``INHERITS``:: + + Table("some_table", metadata, ..., postgresql_inherits="some_supertable") + + Table("some_table", metadata, ..., postgresql_inherits=("t1", "t2", ...)) + +* ``ON COMMIT``:: + + Table("some_table", metadata, ..., postgresql_on_commit='PRESERVE ROWS') + +* ``PARTITION BY``:: + + Table("some_table", metadata, ..., + postgresql_partition_by='LIST (part_column)') + + .. versionadded:: 1.2.6 + +* ``TABLESPACE``:: + + Table("some_table", metadata, ..., postgresql_tablespace='some_tablespace') + + The above option is also available on the :class:`.Index` construct. + +* ``USING``:: + + Table("some_table", metadata, ..., postgresql_using='heap') + + .. versionadded:: 2.0.26 + +* ``WITH OIDS``:: + + Table("some_table", metadata, ..., postgresql_with_oids=True) + +* ``WITHOUT OIDS``:: + + Table("some_table", metadata, ..., postgresql_with_oids=False) + +.. seealso:: + + `PostgreSQL CREATE TABLE options + <https://www.postgresql.org/docs/current/static/sql-createtable.html>`_ - + in the PostgreSQL documentation. + +.. _postgresql_constraint_options: + +PostgreSQL Constraint Options +----------------------------- + +The following option(s) are supported by the PostgreSQL dialect in conjunction +with selected constraint constructs: + +* ``NOT VALID``: This option applies towards CHECK and FOREIGN KEY constraints + when the constraint is being added to an existing table via ALTER TABLE, + and has the effect that existing rows are not scanned during the ALTER + operation against the constraint being added. + + When using a SQL migration tool such as `Alembic <https://alembic.sqlalchemy.org>`_ + that renders ALTER TABLE constructs, the ``postgresql_not_valid`` argument + may be specified as an additional keyword argument within the operation + that creates the constraint, as in the following Alembic example:: + + def update(): + op.create_foreign_key( + "fk_user_address", + "address", + "user", + ["user_id"], + ["id"], + postgresql_not_valid=True + ) + + The keyword is ultimately accepted directly by the + :class:`_schema.CheckConstraint`, :class:`_schema.ForeignKeyConstraint` + and :class:`_schema.ForeignKey` constructs; when using a tool like + Alembic, dialect-specific keyword arguments are passed through to + these constructs from the migration operation directives:: + + CheckConstraint("some_field IS NOT NULL", postgresql_not_valid=True) + + ForeignKeyConstraint(["some_id"], ["some_table.some_id"], postgresql_not_valid=True) + + .. versionadded:: 1.4.32 + + .. seealso:: + + `PostgreSQL ALTER TABLE options + <https://www.postgresql.org/docs/current/static/sql-altertable.html>`_ - + in the PostgreSQL documentation. + +.. _postgresql_table_valued_overview: + +Table values, Table and Column valued functions, Row and Tuple objects +----------------------------------------------------------------------- + +PostgreSQL makes great use of modern SQL forms such as table-valued functions, +tables and rows as values. These constructs are commonly used as part +of PostgreSQL's support for complex datatypes such as JSON, ARRAY, and other +datatypes. SQLAlchemy's SQL expression language has native support for +most table-valued and row-valued forms. + +.. _postgresql_table_valued: + +Table-Valued Functions +^^^^^^^^^^^^^^^^^^^^^^^ + +Many PostgreSQL built-in functions are intended to be used in the FROM clause +of a SELECT statement, and are capable of returning table rows or sets of table +rows. A large portion of PostgreSQL's JSON functions for example such as +``json_array_elements()``, ``json_object_keys()``, ``json_each_text()``, +``json_each()``, ``json_to_record()``, ``json_populate_recordset()`` use such +forms. These classes of SQL function calling forms in SQLAlchemy are available +using the :meth:`_functions.FunctionElement.table_valued` method in conjunction +with :class:`_functions.Function` objects generated from the :data:`_sql.func` +namespace. + +Examples from PostgreSQL's reference documentation follow below: + +* ``json_each()``: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import select, func + >>> stmt = select(func.json_each('{"a":"foo", "b":"bar"}').table_valued("key", "value")) + >>> print(stmt) + {printsql}SELECT anon_1.key, anon_1.value + FROM json_each(:json_each_1) AS anon_1 + +* ``json_populate_record()``: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import select, func, literal_column + >>> stmt = select( + ... func.json_populate_record( + ... literal_column("null::myrowtype"), + ... '{"a":1,"b":2}' + ... ).table_valued("a", "b", name="x") + ... ) + >>> print(stmt) + {printsql}SELECT x.a, x.b + FROM json_populate_record(null::myrowtype, :json_populate_record_1) AS x + +* ``json_to_record()`` - this form uses a PostgreSQL specific form of derived + columns in the alias, where we may make use of :func:`_sql.column` elements with + types to produce them. The :meth:`_functions.FunctionElement.table_valued` + method produces a :class:`_sql.TableValuedAlias` construct, and the method + :meth:`_sql.TableValuedAlias.render_derived` method sets up the derived + columns specification: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import select, func, column, Integer, Text + >>> stmt = select( + ... func.json_to_record('{"a":1,"b":[1,2,3],"c":"bar"}').table_valued( + ... column("a", Integer), column("b", Text), column("d", Text), + ... ).render_derived(name="x", with_types=True) + ... ) + >>> print(stmt) + {printsql}SELECT x.a, x.b, x.d + FROM json_to_record(:json_to_record_1) AS x(a INTEGER, b TEXT, d TEXT) + +* ``WITH ORDINALITY`` - part of the SQL standard, ``WITH ORDINALITY`` adds an + ordinal counter to the output of a function and is accepted by a limited set + of PostgreSQL functions including ``unnest()`` and ``generate_series()``. The + :meth:`_functions.FunctionElement.table_valued` method accepts a keyword + parameter ``with_ordinality`` for this purpose, which accepts the string name + that will be applied to the "ordinality" column: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import select, func + >>> stmt = select( + ... func.generate_series(4, 1, -1). + ... table_valued("value", with_ordinality="ordinality"). + ... render_derived() + ... ) + >>> print(stmt) + {printsql}SELECT anon_1.value, anon_1.ordinality + FROM generate_series(:generate_series_1, :generate_series_2, :generate_series_3) + WITH ORDINALITY AS anon_1(value, ordinality) + +.. versionadded:: 1.4.0b2 + +.. seealso:: + + :ref:`tutorial_functions_table_valued` - in the :ref:`unified_tutorial` + +.. _postgresql_column_valued: + +Column Valued Functions +^^^^^^^^^^^^^^^^^^^^^^^ + +Similar to the table valued function, a column valued function is present +in the FROM clause, but delivers itself to the columns clause as a single +scalar value. PostgreSQL functions such as ``json_array_elements()``, +``unnest()`` and ``generate_series()`` may use this form. Column valued functions are available using the +:meth:`_functions.FunctionElement.column_valued` method of :class:`_functions.FunctionElement`: + +* ``json_array_elements()``: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import select, func + >>> stmt = select(func.json_array_elements('["one", "two"]').column_valued("x")) + >>> print(stmt) + {printsql}SELECT x + FROM json_array_elements(:json_array_elements_1) AS x + +* ``unnest()`` - in order to generate a PostgreSQL ARRAY literal, the + :func:`_postgresql.array` construct may be used: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy.dialects.postgresql import array + >>> from sqlalchemy import select, func + >>> stmt = select(func.unnest(array([1, 2])).column_valued()) + >>> print(stmt) + {printsql}SELECT anon_1 + FROM unnest(ARRAY[%(param_1)s, %(param_2)s]) AS anon_1 + + The function can of course be used against an existing table-bound column + that's of type :class:`_types.ARRAY`: + + .. sourcecode:: pycon+sql + + >>> from sqlalchemy import table, column, ARRAY, Integer + >>> from sqlalchemy import select, func + >>> t = table("t", column('value', ARRAY(Integer))) + >>> stmt = select(func.unnest(t.c.value).column_valued("unnested_value")) + >>> print(stmt) + {printsql}SELECT unnested_value + FROM unnest(t.value) AS unnested_value + +.. seealso:: + + :ref:`tutorial_functions_column_valued` - in the :ref:`unified_tutorial` + + +Row Types +^^^^^^^^^ + +Built-in support for rendering a ``ROW`` may be approximated using +``func.ROW`` with the :attr:`_sa.func` namespace, or by using the +:func:`_sql.tuple_` construct: + +.. sourcecode:: pycon+sql + + >>> from sqlalchemy import table, column, func, tuple_ + >>> t = table("t", column("id"), column("fk")) + >>> stmt = t.select().where( + ... tuple_(t.c.id, t.c.fk) > (1,2) + ... ).where( + ... func.ROW(t.c.id, t.c.fk) < func.ROW(3, 7) + ... ) + >>> print(stmt) + {printsql}SELECT t.id, t.fk + FROM t + WHERE (t.id, t.fk) > (:param_1, :param_2) AND ROW(t.id, t.fk) < ROW(:ROW_1, :ROW_2) + +.. seealso:: + + `PostgreSQL Row Constructors + <https://www.postgresql.org/docs/current/sql-expressions.html#SQL-SYNTAX-ROW-CONSTRUCTORS>`_ + + `PostgreSQL Row Constructor Comparison + <https://www.postgresql.org/docs/current/functions-comparisons.html#ROW-WISE-COMPARISON>`_ + +Table Types passed to Functions +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +PostgreSQL supports passing a table as an argument to a function, which is +known as a "record" type. SQLAlchemy :class:`_sql.FromClause` objects +such as :class:`_schema.Table` support this special form using the +:meth:`_sql.FromClause.table_valued` method, which is comparable to the +:meth:`_functions.FunctionElement.table_valued` method except that the collection +of columns is already established by that of the :class:`_sql.FromClause` +itself: + +.. sourcecode:: pycon+sql + + >>> from sqlalchemy import table, column, func, select + >>> a = table( "a", column("id"), column("x"), column("y")) + >>> stmt = select(func.row_to_json(a.table_valued())) + >>> print(stmt) + {printsql}SELECT row_to_json(a) AS row_to_json_1 + FROM a + +.. versionadded:: 1.4.0b2 + + + +""" # noqa: E501 + +from __future__ import annotations + +from collections import defaultdict +from functools import lru_cache +import re +from typing import Any +from typing import cast +from typing import List +from typing import Optional +from typing import Tuple +from typing import TYPE_CHECKING +from typing import Union + +from . import arraylib as _array +from . import json as _json +from . import pg_catalog +from . import ranges as _ranges +from .ext import _regconfig_fn +from .ext import aggregate_order_by +from .hstore import HSTORE +from .named_types import CreateDomainType as CreateDomainType # noqa: F401 +from .named_types import CreateEnumType as CreateEnumType # noqa: F401 +from .named_types import DOMAIN as DOMAIN # noqa: F401 +from .named_types import DropDomainType as DropDomainType # noqa: F401 +from .named_types import DropEnumType as DropEnumType # noqa: F401 +from .named_types import ENUM as ENUM # noqa: F401 +from .named_types import NamedType as NamedType # noqa: F401 +from .types import _DECIMAL_TYPES # noqa: F401 +from .types import _FLOAT_TYPES # noqa: F401 +from .types import _INT_TYPES # noqa: F401 +from .types import BIT as BIT +from .types import BYTEA as BYTEA +from .types import CIDR as CIDR +from .types import CITEXT as CITEXT +from .types import INET as INET +from .types import INTERVAL as INTERVAL +from .types import MACADDR as MACADDR +from .types import MACADDR8 as MACADDR8 +from .types import MONEY as MONEY +from .types import OID as OID +from .types import PGBit as PGBit # noqa: F401 +from .types import PGCidr as PGCidr # noqa: F401 +from .types import PGInet as PGInet # noqa: F401 +from .types import PGInterval as PGInterval # noqa: F401 +from .types import PGMacAddr as PGMacAddr # noqa: F401 +from .types import PGMacAddr8 as PGMacAddr8 # noqa: F401 +from .types import PGUuid as PGUuid +from .types import REGCLASS as REGCLASS +from .types import REGCONFIG as REGCONFIG # noqa: F401 +from .types import TIME as TIME +from .types import TIMESTAMP as TIMESTAMP +from .types import TSVECTOR as TSVECTOR +from ... import exc +from ... import schema +from ... import select +from ... import sql +from ... import util +from ...engine import characteristics +from ...engine import default +from ...engine import interfaces +from ...engine import ObjectKind +from ...engine import ObjectScope +from ...engine import reflection +from ...engine import URL +from ...engine.reflection import ReflectionDefaults +from ...sql import bindparam +from ...sql import coercions +from ...sql import compiler +from ...sql import elements +from ...sql import expression +from ...sql import roles +from ...sql import sqltypes +from ...sql import util as sql_util +from ...sql.compiler import InsertmanyvaluesSentinelOpts +from ...sql.visitors import InternalTraversal +from ...types import BIGINT +from ...types import BOOLEAN +from ...types import CHAR +from ...types import DATE +from ...types import DOUBLE_PRECISION +from ...types import FLOAT +from ...types import INTEGER +from ...types import NUMERIC +from ...types import REAL +from ...types import SMALLINT +from ...types import TEXT +from ...types import UUID as UUID +from ...types import VARCHAR +from ...util.typing import TypedDict + +IDX_USING = re.compile(r"^(?:btree|hash|gist|gin|[\w_]+)$", re.I) + +RESERVED_WORDS = { + "all", + "analyse", + "analyze", + "and", + "any", + "array", + "as", + "asc", + "asymmetric", + "both", + "case", + "cast", + "check", + "collate", + "column", + "constraint", + "create", + "current_catalog", + "current_date", + "current_role", + "current_time", + "current_timestamp", + "current_user", + "default", + "deferrable", + "desc", + "distinct", + "do", + "else", + "end", + "except", + "false", + "fetch", + "for", + "foreign", + "from", + "grant", + "group", + "having", + "in", + "initially", + "intersect", + "into", + "leading", + "limit", + "localtime", + "localtimestamp", + "new", + "not", + "null", + "of", + "off", + "offset", + "old", + "on", + "only", + "or", + "order", + "placing", + "primary", + "references", + "returning", + "select", + "session_user", + "some", + "symmetric", + "table", + "then", + "to", + "trailing", + "true", + "union", + "unique", + "user", + "using", + "variadic", + "when", + "where", + "window", + "with", + "authorization", + "between", + "binary", + "cross", + "current_schema", + "freeze", + "full", + "ilike", + "inner", + "is", + "isnull", + "join", + "left", + "like", + "natural", + "notnull", + "outer", + "over", + "overlaps", + "right", + "similar", + "verbose", +} + +colspecs = { + sqltypes.ARRAY: _array.ARRAY, + sqltypes.Interval: INTERVAL, + sqltypes.Enum: ENUM, + sqltypes.JSON.JSONPathType: _json.JSONPATH, + sqltypes.JSON: _json.JSON, + sqltypes.Uuid: PGUuid, +} + + +ischema_names = { + "_array": _array.ARRAY, + "hstore": HSTORE, + "json": _json.JSON, + "jsonb": _json.JSONB, + "int4range": _ranges.INT4RANGE, + "int8range": _ranges.INT8RANGE, + "numrange": _ranges.NUMRANGE, + "daterange": _ranges.DATERANGE, + "tsrange": _ranges.TSRANGE, + "tstzrange": _ranges.TSTZRANGE, + "int4multirange": _ranges.INT4MULTIRANGE, + "int8multirange": _ranges.INT8MULTIRANGE, + "nummultirange": _ranges.NUMMULTIRANGE, + "datemultirange": _ranges.DATEMULTIRANGE, + "tsmultirange": _ranges.TSMULTIRANGE, + "tstzmultirange": _ranges.TSTZMULTIRANGE, + "integer": INTEGER, + "bigint": BIGINT, + "smallint": SMALLINT, + "character varying": VARCHAR, + "character": CHAR, + '"char"': sqltypes.String, + "name": sqltypes.String, + "text": TEXT, + "numeric": NUMERIC, + "float": FLOAT, + "real": REAL, + "inet": INET, + "cidr": CIDR, + "citext": CITEXT, + "uuid": UUID, + "bit": BIT, + "bit varying": BIT, + "macaddr": MACADDR, + "macaddr8": MACADDR8, + "money": MONEY, + "oid": OID, + "regclass": REGCLASS, + "double precision": DOUBLE_PRECISION, + "timestamp": TIMESTAMP, + "timestamp with time zone": TIMESTAMP, + "timestamp without time zone": TIMESTAMP, + "time with time zone": TIME, + "time without time zone": TIME, + "date": DATE, + "time": TIME, + "bytea": BYTEA, + "boolean": BOOLEAN, + "interval": INTERVAL, + "tsvector": TSVECTOR, +} + + +class PGCompiler(compiler.SQLCompiler): + def visit_to_tsvector_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def visit_to_tsquery_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def visit_plainto_tsquery_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def visit_phraseto_tsquery_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def visit_websearch_to_tsquery_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def visit_ts_headline_func(self, element, **kw): + return self._assert_pg_ts_ext(element, **kw) + + def _assert_pg_ts_ext(self, element, **kw): + if not isinstance(element, _regconfig_fn): + # other options here include trying to rewrite the function + # with the correct types. however, that means we have to + # "un-SQL-ize" the first argument, which can't work in a + # generalized way. Also, parent compiler class has already added + # the incorrect return type to the result map. So let's just + # make sure the function we want is used up front. + + raise exc.CompileError( + f'Can\'t compile "{element.name}()" full text search ' + f"function construct that does not originate from the " + f'"sqlalchemy.dialects.postgresql" package. ' + f'Please ensure "import sqlalchemy.dialects.postgresql" is ' + f"called before constructing " + f'"sqlalchemy.func.{element.name}()" to ensure registration ' + f"of the correct argument and return types." + ) + + return f"{element.name}{self.function_argspec(element, **kw)}" + + def render_bind_cast(self, type_, dbapi_type, sqltext): + if dbapi_type._type_affinity is sqltypes.String and dbapi_type.length: + # use VARCHAR with no length for VARCHAR cast. + # see #9511 + dbapi_type = sqltypes.STRINGTYPE + return f"""{sqltext}::{ + self.dialect.type_compiler_instance.process( + dbapi_type, identifier_preparer=self.preparer + ) + }""" + + def visit_array(self, element, **kw): + return "ARRAY[%s]" % self.visit_clauselist(element, **kw) + + def visit_slice(self, element, **kw): + return "%s:%s" % ( + self.process(element.start, **kw), + self.process(element.stop, **kw), + ) + + def visit_bitwise_xor_op_binary(self, binary, operator, **kw): + return self._generate_generic_binary(binary, " # ", **kw) + + def visit_json_getitem_op_binary( + self, binary, operator, _cast_applied=False, **kw + ): + if ( + not _cast_applied + and binary.type._type_affinity is not sqltypes.JSON + ): + kw["_cast_applied"] = True + return self.process(sql.cast(binary, binary.type), **kw) + + kw["eager_grouping"] = True + + return self._generate_generic_binary( + binary, " -> " if not _cast_applied else " ->> ", **kw + ) + + def visit_json_path_getitem_op_binary( + self, binary, operator, _cast_applied=False, **kw + ): + if ( + not _cast_applied + and binary.type._type_affinity is not sqltypes.JSON + ): + kw["_cast_applied"] = True + return self.process(sql.cast(binary, binary.type), **kw) + + kw["eager_grouping"] = True + return self._generate_generic_binary( + binary, " #> " if not _cast_applied else " #>> ", **kw + ) + + def visit_getitem_binary(self, binary, operator, **kw): + return "%s[%s]" % ( + self.process(binary.left, **kw), + self.process(binary.right, **kw), + ) + + def visit_aggregate_order_by(self, element, **kw): + return "%s ORDER BY %s" % ( + self.process(element.target, **kw), + self.process(element.order_by, **kw), + ) + + def visit_match_op_binary(self, binary, operator, **kw): + if "postgresql_regconfig" in binary.modifiers: + regconfig = self.render_literal_value( + binary.modifiers["postgresql_regconfig"], sqltypes.STRINGTYPE + ) + if regconfig: + return "%s @@ plainto_tsquery(%s, %s)" % ( + self.process(binary.left, **kw), + regconfig, + self.process(binary.right, **kw), + ) + return "%s @@ plainto_tsquery(%s)" % ( + self.process(binary.left, **kw), + self.process(binary.right, **kw), + ) + + def visit_ilike_case_insensitive_operand(self, element, **kw): + return element.element._compiler_dispatch(self, **kw) + + def visit_ilike_op_binary(self, binary, operator, **kw): + escape = binary.modifiers.get("escape", None) + + return "%s ILIKE %s" % ( + self.process(binary.left, **kw), + self.process(binary.right, **kw), + ) + ( + " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE) + if escape is not None + else "" + ) + + def visit_not_ilike_op_binary(self, binary, operator, **kw): + escape = binary.modifiers.get("escape", None) + return "%s NOT ILIKE %s" % ( + self.process(binary.left, **kw), + self.process(binary.right, **kw), + ) + ( + " ESCAPE " + self.render_literal_value(escape, sqltypes.STRINGTYPE) + if escape is not None + else "" + ) + + def _regexp_match(self, base_op, binary, operator, kw): + flags = binary.modifiers["flags"] + if flags is None: + return self._generate_generic_binary( + binary, " %s " % base_op, **kw + ) + if flags == "i": + return self._generate_generic_binary( + binary, " %s* " % base_op, **kw + ) + return "%s %s CONCAT('(?', %s, ')', %s)" % ( + self.process(binary.left, **kw), + base_op, + self.render_literal_value(flags, sqltypes.STRINGTYPE), + self.process(binary.right, **kw), + ) + + def visit_regexp_match_op_binary(self, binary, operator, **kw): + return self._regexp_match("~", binary, operator, kw) + + def visit_not_regexp_match_op_binary(self, binary, operator, **kw): + return self._regexp_match("!~", binary, operator, kw) + + def visit_regexp_replace_op_binary(self, binary, operator, **kw): + string = self.process(binary.left, **kw) + pattern_replace = self.process(binary.right, **kw) + flags = binary.modifiers["flags"] + if flags is None: + return "REGEXP_REPLACE(%s, %s)" % ( + string, + pattern_replace, + ) + else: + return "REGEXP_REPLACE(%s, %s, %s)" % ( + string, + pattern_replace, + self.render_literal_value(flags, sqltypes.STRINGTYPE), + ) + + def visit_empty_set_expr(self, element_types, **kw): + # cast the empty set to the type we are comparing against. if + # we are comparing against the null type, pick an arbitrary + # datatype for the empty set + return "SELECT %s WHERE 1!=1" % ( + ", ".join( + "CAST(NULL AS %s)" + % self.dialect.type_compiler_instance.process( + INTEGER() if type_._isnull else type_ + ) + for type_ in element_types or [INTEGER()] + ), + ) + + def render_literal_value(self, value, type_): + value = super().render_literal_value(value, type_) + + if self.dialect._backslash_escapes: + value = value.replace("\\", "\\\\") + return value + + def visit_aggregate_strings_func(self, fn, **kw): + return "string_agg%s" % self.function_argspec(fn) + + def visit_sequence(self, seq, **kw): + return "nextval('%s')" % self.preparer.format_sequence(seq) + + def limit_clause(self, select, **kw): + text = "" + if select._limit_clause is not None: + text += " \n LIMIT " + self.process(select._limit_clause, **kw) + if select._offset_clause is not None: + if select._limit_clause is None: + text += "\n LIMIT ALL" + text += " OFFSET " + self.process(select._offset_clause, **kw) + return text + + def format_from_hint_text(self, sqltext, table, hint, iscrud): + if hint.upper() != "ONLY": + raise exc.CompileError("Unrecognized hint: %r" % hint) + return "ONLY " + sqltext + + def get_select_precolumns(self, select, **kw): + # Do not call super().get_select_precolumns because + # it will warn/raise when distinct on is present + if select._distinct or select._distinct_on: + if select._distinct_on: + return ( + "DISTINCT ON (" + + ", ".join( + [ + self.process(col, **kw) + for col in select._distinct_on + ] + ) + + ") " + ) + else: + return "DISTINCT " + else: + return "" + + def for_update_clause(self, select, **kw): + if select._for_update_arg.read: + if select._for_update_arg.key_share: + tmp = " FOR KEY SHARE" + else: + tmp = " FOR SHARE" + elif select._for_update_arg.key_share: + tmp = " FOR NO KEY UPDATE" + else: + tmp = " FOR UPDATE" + + if select._for_update_arg.of: + tables = util.OrderedSet() + for c in select._for_update_arg.of: + tables.update(sql_util.surface_selectables_only(c)) + + tmp += " OF " + ", ".join( + self.process(table, ashint=True, use_schema=False, **kw) + for table in tables + ) + + if select._for_update_arg.nowait: + tmp += " NOWAIT" + if select._for_update_arg.skip_locked: + tmp += " SKIP LOCKED" + + return tmp + + def visit_substring_func(self, func, **kw): + s = self.process(func.clauses.clauses[0], **kw) + start = self.process(func.clauses.clauses[1], **kw) + if len(func.clauses.clauses) > 2: + length = self.process(func.clauses.clauses[2], **kw) + return "SUBSTRING(%s FROM %s FOR %s)" % (s, start, length) + else: + return "SUBSTRING(%s FROM %s)" % (s, start) + + def _on_conflict_target(self, clause, **kw): + if clause.constraint_target is not None: + # target may be a name of an Index, UniqueConstraint or + # ExcludeConstraint. While there is a separate + # "max_identifier_length" for indexes, PostgreSQL uses the same + # length for all objects so we can use + # truncate_and_render_constraint_name + target_text = ( + "ON CONSTRAINT %s" + % self.preparer.truncate_and_render_constraint_name( + clause.constraint_target + ) + ) + elif clause.inferred_target_elements is not None: + target_text = "(%s)" % ", ".join( + ( + self.preparer.quote(c) + if isinstance(c, str) + else self.process(c, include_table=False, use_schema=False) + ) + for c in clause.inferred_target_elements + ) + if clause.inferred_target_whereclause is not None: + target_text += " WHERE %s" % self.process( + clause.inferred_target_whereclause, + include_table=False, + use_schema=False, + ) + else: + target_text = "" + + return target_text + + def visit_on_conflict_do_nothing(self, on_conflict, **kw): + target_text = self._on_conflict_target(on_conflict, **kw) + + if target_text: + return "ON CONFLICT %s DO NOTHING" % target_text + else: + return "ON CONFLICT DO NOTHING" + + def visit_on_conflict_do_update(self, on_conflict, **kw): + clause = on_conflict + + target_text = self._on_conflict_target(on_conflict, **kw) + + action_set_ops = [] + + set_parameters = dict(clause.update_values_to_set) + # create a list of column assignment clauses as tuples + + insert_statement = self.stack[-1]["selectable"] + cols = insert_statement.table.c + for c in cols: + col_key = c.key + + if col_key in set_parameters: + value = set_parameters.pop(col_key) + elif c in set_parameters: + value = set_parameters.pop(c) + else: + continue + + if coercions._is_literal(value): + value = elements.BindParameter(None, value, type_=c.type) + + else: + if ( + isinstance(value, elements.BindParameter) + and value.type._isnull + ): + value = value._clone() + value.type = c.type + value_text = self.process(value.self_group(), use_schema=False) + + key_text = self.preparer.quote(c.name) + action_set_ops.append("%s = %s" % (key_text, value_text)) + + # check for names that don't match columns + if set_parameters: + util.warn( + "Additional column names not matching " + "any column keys in table '%s': %s" + % ( + self.current_executable.table.name, + (", ".join("'%s'" % c for c in set_parameters)), + ) + ) + for k, v in set_parameters.items(): + key_text = ( + self.preparer.quote(k) + if isinstance(k, str) + else self.process(k, use_schema=False) + ) + value_text = self.process( + coercions.expect(roles.ExpressionElementRole, v), + use_schema=False, + ) + action_set_ops.append("%s = %s" % (key_text, value_text)) + + action_text = ", ".join(action_set_ops) + if clause.update_whereclause is not None: + action_text += " WHERE %s" % self.process( + clause.update_whereclause, include_table=True, use_schema=False + ) + + return "ON CONFLICT %s DO UPDATE SET %s" % (target_text, action_text) + + def update_from_clause( + self, update_stmt, from_table, extra_froms, from_hints, **kw + ): + kw["asfrom"] = True + return "FROM " + ", ".join( + t._compiler_dispatch(self, fromhints=from_hints, **kw) + for t in extra_froms + ) + + def delete_extra_from_clause( + self, delete_stmt, from_table, extra_froms, from_hints, **kw + ): + """Render the DELETE .. USING clause specific to PostgreSQL.""" + kw["asfrom"] = True + return "USING " + ", ".join( + t._compiler_dispatch(self, fromhints=from_hints, **kw) + for t in extra_froms + ) + + def fetch_clause(self, select, **kw): + # pg requires parens for non literal clauses. It's also required for + # bind parameters if a ::type casts is used by the driver (asyncpg), + # so it's easiest to just always add it + text = "" + if select._offset_clause is not None: + text += "\n OFFSET (%s) ROWS" % self.process( + select._offset_clause, **kw + ) + if select._fetch_clause is not None: + text += "\n FETCH FIRST (%s)%s ROWS %s" % ( + self.process(select._fetch_clause, **kw), + " PERCENT" if select._fetch_clause_options["percent"] else "", + ( + "WITH TIES" + if select._fetch_clause_options["with_ties"] + else "ONLY" + ), + ) + return text + + +class PGDDLCompiler(compiler.DDLCompiler): + def get_column_specification(self, column, **kwargs): + colspec = self.preparer.format_column(column) + impl_type = column.type.dialect_impl(self.dialect) + if isinstance(impl_type, sqltypes.TypeDecorator): + impl_type = impl_type.impl + + has_identity = ( + column.identity is not None + and self.dialect.supports_identity_columns + ) + + if ( + column.primary_key + and column is column.table._autoincrement_column + and ( + self.dialect.supports_smallserial + or not isinstance(impl_type, sqltypes.SmallInteger) + ) + and not has_identity + and ( + column.default is None + or ( + isinstance(column.default, schema.Sequence) + and column.default.optional + ) + ) + ): + if isinstance(impl_type, sqltypes.BigInteger): + colspec += " BIGSERIAL" + elif isinstance(impl_type, sqltypes.SmallInteger): + colspec += " SMALLSERIAL" + else: + colspec += " SERIAL" + else: + colspec += " " + self.dialect.type_compiler_instance.process( + column.type, + type_expression=column, + identifier_preparer=self.preparer, + ) + default = self.get_column_default_string(column) + if default is not None: + colspec += " DEFAULT " + default + + if column.computed is not None: + colspec += " " + self.process(column.computed) + if has_identity: + colspec += " " + self.process(column.identity) + + if not column.nullable and not has_identity: + colspec += " NOT NULL" + elif column.nullable and has_identity: + colspec += " NULL" + return colspec + + def _define_constraint_validity(self, constraint): + not_valid = constraint.dialect_options["postgresql"]["not_valid"] + return " NOT VALID" if not_valid else "" + + def visit_check_constraint(self, constraint, **kw): + if constraint._type_bound: + typ = list(constraint.columns)[0].type + if ( + isinstance(typ, sqltypes.ARRAY) + and isinstance(typ.item_type, sqltypes.Enum) + and not typ.item_type.native_enum + ): + raise exc.CompileError( + "PostgreSQL dialect cannot produce the CHECK constraint " + "for ARRAY of non-native ENUM; please specify " + "create_constraint=False on this Enum datatype." + ) + + text = super().visit_check_constraint(constraint) + text += self._define_constraint_validity(constraint) + return text + + def visit_foreign_key_constraint(self, constraint, **kw): + text = super().visit_foreign_key_constraint(constraint) + text += self._define_constraint_validity(constraint) + return text + + def visit_create_enum_type(self, create, **kw): + type_ = create.element + + return "CREATE TYPE %s AS ENUM (%s)" % ( + self.preparer.format_type(type_), + ", ".join( + self.sql_compiler.process(sql.literal(e), literal_binds=True) + for e in type_.enums + ), + ) + + def visit_drop_enum_type(self, drop, **kw): + type_ = drop.element + + return "DROP TYPE %s" % (self.preparer.format_type(type_)) + + def visit_create_domain_type(self, create, **kw): + domain: DOMAIN = create.element + + options = [] + if domain.collation is not None: + options.append(f"COLLATE {self.preparer.quote(domain.collation)}") + if domain.default is not None: + default = self.render_default_string(domain.default) + options.append(f"DEFAULT {default}") + if domain.constraint_name is not None: + name = self.preparer.truncate_and_render_constraint_name( + domain.constraint_name + ) + options.append(f"CONSTRAINT {name}") + if domain.not_null: + options.append("NOT NULL") + if domain.check is not None: + check = self.sql_compiler.process( + domain.check, include_table=False, literal_binds=True + ) + options.append(f"CHECK ({check})") + + return ( + f"CREATE DOMAIN {self.preparer.format_type(domain)} AS " + f"{self.type_compiler.process(domain.data_type)} " + f"{' '.join(options)}" + ) + + def visit_drop_domain_type(self, drop, **kw): + domain = drop.element + return f"DROP DOMAIN {self.preparer.format_type(domain)}" + + def visit_create_index(self, create, **kw): + preparer = self.preparer + index = create.element + self._verify_index_table(index) + text = "CREATE " + if index.unique: + text += "UNIQUE " + + text += "INDEX " + + if self.dialect._supports_create_index_concurrently: + concurrently = index.dialect_options["postgresql"]["concurrently"] + if concurrently: + text += "CONCURRENTLY " + + if create.if_not_exists: + text += "IF NOT EXISTS " + + text += "%s ON %s " % ( + self._prepared_index_name(index, include_schema=False), + preparer.format_table(index.table), + ) + + using = index.dialect_options["postgresql"]["using"] + if using: + text += ( + "USING %s " + % self.preparer.validate_sql_phrase(using, IDX_USING).lower() + ) + + ops = index.dialect_options["postgresql"]["ops"] + text += "(%s)" % ( + ", ".join( + [ + self.sql_compiler.process( + ( + expr.self_group() + if not isinstance(expr, expression.ColumnClause) + else expr + ), + include_table=False, + literal_binds=True, + ) + + ( + (" " + ops[expr.key]) + if hasattr(expr, "key") and expr.key in ops + else "" + ) + for expr in index.expressions + ] + ) + ) + + includeclause = index.dialect_options["postgresql"]["include"] + if includeclause: + inclusions = [ + index.table.c[col] if isinstance(col, str) else col + for col in includeclause + ] + text += " INCLUDE (%s)" % ", ".join( + [preparer.quote(c.name) for c in inclusions] + ) + + nulls_not_distinct = index.dialect_options["postgresql"][ + "nulls_not_distinct" + ] + if nulls_not_distinct is True: + text += " NULLS NOT DISTINCT" + elif nulls_not_distinct is False: + text += " NULLS DISTINCT" + + withclause = index.dialect_options["postgresql"]["with"] + if withclause: + text += " WITH (%s)" % ( + ", ".join( + [ + "%s = %s" % storage_parameter + for storage_parameter in withclause.items() + ] + ) + ) + + tablespace_name = index.dialect_options["postgresql"]["tablespace"] + if tablespace_name: + text += " TABLESPACE %s" % preparer.quote(tablespace_name) + + whereclause = index.dialect_options["postgresql"]["where"] + if whereclause is not None: + whereclause = coercions.expect( + roles.DDLExpressionRole, whereclause + ) + + where_compiled = self.sql_compiler.process( + whereclause, include_table=False, literal_binds=True + ) + text += " WHERE " + where_compiled + + return text + + def define_unique_constraint_distinct(self, constraint, **kw): + nulls_not_distinct = constraint.dialect_options["postgresql"][ + "nulls_not_distinct" + ] + if nulls_not_distinct is True: + nulls_not_distinct_param = "NULLS NOT DISTINCT " + elif nulls_not_distinct is False: + nulls_not_distinct_param = "NULLS DISTINCT " + else: + nulls_not_distinct_param = "" + return nulls_not_distinct_param + + def visit_drop_index(self, drop, **kw): + index = drop.element + + text = "\nDROP INDEX " + + if self.dialect._supports_drop_index_concurrently: + concurrently = index.dialect_options["postgresql"]["concurrently"] + if concurrently: + text += "CONCURRENTLY " + + if drop.if_exists: + text += "IF EXISTS " + + text += self._prepared_index_name(index, include_schema=True) + return text + + def visit_exclude_constraint(self, constraint, **kw): + text = "" + if constraint.name is not None: + text += "CONSTRAINT %s " % self.preparer.format_constraint( + constraint + ) + elements = [] + kw["include_table"] = False + kw["literal_binds"] = True + for expr, name, op in constraint._render_exprs: + exclude_element = self.sql_compiler.process(expr, **kw) + ( + (" " + constraint.ops[expr.key]) + if hasattr(expr, "key") and expr.key in constraint.ops + else "" + ) + + elements.append("%s WITH %s" % (exclude_element, op)) + text += "EXCLUDE USING %s (%s)" % ( + self.preparer.validate_sql_phrase( + constraint.using, IDX_USING + ).lower(), + ", ".join(elements), + ) + if constraint.where is not None: + text += " WHERE (%s)" % self.sql_compiler.process( + constraint.where, literal_binds=True + ) + text += self.define_constraint_deferrability(constraint) + return text + + def post_create_table(self, table): + table_opts = [] + pg_opts = table.dialect_options["postgresql"] + + inherits = pg_opts.get("inherits") + if inherits is not None: + if not isinstance(inherits, (list, tuple)): + inherits = (inherits,) + table_opts.append( + "\n INHERITS ( " + + ", ".join(self.preparer.quote(name) for name in inherits) + + " )" + ) + + if pg_opts["partition_by"]: + table_opts.append("\n PARTITION BY %s" % pg_opts["partition_by"]) + + if pg_opts["using"]: + table_opts.append("\n USING %s" % pg_opts["using"]) + + if pg_opts["with_oids"] is True: + table_opts.append("\n WITH OIDS") + elif pg_opts["with_oids"] is False: + table_opts.append("\n WITHOUT OIDS") + + if pg_opts["on_commit"]: + on_commit_options = pg_opts["on_commit"].replace("_", " ").upper() + table_opts.append("\n ON COMMIT %s" % on_commit_options) + + if pg_opts["tablespace"]: + tablespace_name = pg_opts["tablespace"] + table_opts.append( + "\n TABLESPACE %s" % self.preparer.quote(tablespace_name) + ) + + return "".join(table_opts) + + def visit_computed_column(self, generated, **kw): + if generated.persisted is False: + raise exc.CompileError( + "PostrgreSQL computed columns do not support 'virtual' " + "persistence; set the 'persisted' flag to None or True for " + "PostgreSQL support." + ) + + return "GENERATED ALWAYS AS (%s) STORED" % self.sql_compiler.process( + generated.sqltext, include_table=False, literal_binds=True + ) + + def visit_create_sequence(self, create, **kw): + prefix = None + if create.element.data_type is not None: + prefix = " AS %s" % self.type_compiler.process( + create.element.data_type + ) + + return super().visit_create_sequence(create, prefix=prefix, **kw) + + def _can_comment_on_constraint(self, ddl_instance): + constraint = ddl_instance.element + if constraint.name is None: + raise exc.CompileError( + f"Can't emit COMMENT ON for constraint {constraint!r}: " + "it has no name" + ) + if constraint.table is None: + raise exc.CompileError( + f"Can't emit COMMENT ON for constraint {constraint!r}: " + "it has no associated table" + ) + + def visit_set_constraint_comment(self, create, **kw): + self._can_comment_on_constraint(create) + return "COMMENT ON CONSTRAINT %s ON %s IS %s" % ( + self.preparer.format_constraint(create.element), + self.preparer.format_table(create.element.table), + self.sql_compiler.render_literal_value( + create.element.comment, sqltypes.String() + ), + ) + + def visit_drop_constraint_comment(self, drop, **kw): + self._can_comment_on_constraint(drop) + return "COMMENT ON CONSTRAINT %s ON %s IS NULL" % ( + self.preparer.format_constraint(drop.element), + self.preparer.format_table(drop.element.table), + ) + + +class PGTypeCompiler(compiler.GenericTypeCompiler): + def visit_TSVECTOR(self, type_, **kw): + return "TSVECTOR" + + def visit_TSQUERY(self, type_, **kw): + return "TSQUERY" + + def visit_INET(self, type_, **kw): + return "INET" + + def visit_CIDR(self, type_, **kw): + return "CIDR" + + def visit_CITEXT(self, type_, **kw): + return "CITEXT" + + def visit_MACADDR(self, type_, **kw): + return "MACADDR" + + def visit_MACADDR8(self, type_, **kw): + return "MACADDR8" + + def visit_MONEY(self, type_, **kw): + return "MONEY" + + def visit_OID(self, type_, **kw): + return "OID" + + def visit_REGCONFIG(self, type_, **kw): + return "REGCONFIG" + + def visit_REGCLASS(self, type_, **kw): + return "REGCLASS" + + def visit_FLOAT(self, type_, **kw): + if not type_.precision: + return "FLOAT" + else: + return "FLOAT(%(precision)s)" % {"precision": type_.precision} + + def visit_double(self, type_, **kw): + return self.visit_DOUBLE_PRECISION(type, **kw) + + def visit_BIGINT(self, type_, **kw): + return "BIGINT" + + def visit_HSTORE(self, type_, **kw): + return "HSTORE" + + def visit_JSON(self, type_, **kw): + return "JSON" + + def visit_JSONB(self, type_, **kw): + return "JSONB" + + def visit_INT4MULTIRANGE(self, type_, **kw): + return "INT4MULTIRANGE" + + def visit_INT8MULTIRANGE(self, type_, **kw): + return "INT8MULTIRANGE" + + def visit_NUMMULTIRANGE(self, type_, **kw): + return "NUMMULTIRANGE" + + def visit_DATEMULTIRANGE(self, type_, **kw): + return "DATEMULTIRANGE" + + def visit_TSMULTIRANGE(self, type_, **kw): + return "TSMULTIRANGE" + + def visit_TSTZMULTIRANGE(self, type_, **kw): + return "TSTZMULTIRANGE" + + def visit_INT4RANGE(self, type_, **kw): + return "INT4RANGE" + + def visit_INT8RANGE(self, type_, **kw): + return "INT8RANGE" + + def visit_NUMRANGE(self, type_, **kw): + return "NUMRANGE" + + def visit_DATERANGE(self, type_, **kw): + return "DATERANGE" + + def visit_TSRANGE(self, type_, **kw): + return "TSRANGE" + + def visit_TSTZRANGE(self, type_, **kw): + return "TSTZRANGE" + + def visit_json_int_index(self, type_, **kw): + return "INT" + + def visit_json_str_index(self, type_, **kw): + return "TEXT" + + def visit_datetime(self, type_, **kw): + return self.visit_TIMESTAMP(type_, **kw) + + def visit_enum(self, type_, **kw): + if not type_.native_enum or not self.dialect.supports_native_enum: + return super().visit_enum(type_, **kw) + else: + return self.visit_ENUM(type_, **kw) + + def visit_ENUM(self, type_, identifier_preparer=None, **kw): + if identifier_preparer is None: + identifier_preparer = self.dialect.identifier_preparer + return identifier_preparer.format_type(type_) + + def visit_DOMAIN(self, type_, identifier_preparer=None, **kw): + if identifier_preparer is None: + identifier_preparer = self.dialect.identifier_preparer + return identifier_preparer.format_type(type_) + + def visit_TIMESTAMP(self, type_, **kw): + return "TIMESTAMP%s %s" % ( + ( + "(%d)" % type_.precision + if getattr(type_, "precision", None) is not None + else "" + ), + (type_.timezone and "WITH" or "WITHOUT") + " TIME ZONE", + ) + + def visit_TIME(self, type_, **kw): + return "TIME%s %s" % ( + ( + "(%d)" % type_.precision + if getattr(type_, "precision", None) is not None + else "" + ), + (type_.timezone and "WITH" or "WITHOUT") + " TIME ZONE", + ) + + def visit_INTERVAL(self, type_, **kw): + text = "INTERVAL" + if type_.fields is not None: + text += " " + type_.fields + if type_.precision is not None: + text += " (%d)" % type_.precision + return text + + def visit_BIT(self, type_, **kw): + if type_.varying: + compiled = "BIT VARYING" + if type_.length is not None: + compiled += "(%d)" % type_.length + else: + compiled = "BIT(%d)" % type_.length + return compiled + + def visit_uuid(self, type_, **kw): + if type_.native_uuid: + return self.visit_UUID(type_, **kw) + else: + return super().visit_uuid(type_, **kw) + + def visit_UUID(self, type_, **kw): + return "UUID" + + def visit_large_binary(self, type_, **kw): + return self.visit_BYTEA(type_, **kw) + + def visit_BYTEA(self, type_, **kw): + return "BYTEA" + + def visit_ARRAY(self, type_, **kw): + inner = self.process(type_.item_type, **kw) + return re.sub( + r"((?: COLLATE.*)?)$", + ( + r"%s\1" + % ( + "[]" + * (type_.dimensions if type_.dimensions is not None else 1) + ) + ), + inner, + count=1, + ) + + def visit_json_path(self, type_, **kw): + return self.visit_JSONPATH(type_, **kw) + + def visit_JSONPATH(self, type_, **kw): + return "JSONPATH" + + +class PGIdentifierPreparer(compiler.IdentifierPreparer): + reserved_words = RESERVED_WORDS + + def _unquote_identifier(self, value): + if value[0] == self.initial_quote: + value = value[1:-1].replace( + self.escape_to_quote, self.escape_quote + ) + return value + + def format_type(self, type_, use_schema=True): + if not type_.name: + raise exc.CompileError( + f"PostgreSQL {type_.__class__.__name__} type requires a name." + ) + + name = self.quote(type_.name) + effective_schema = self.schema_for_object(type_) + + if ( + not self.omit_schema + and use_schema + and effective_schema is not None + ): + name = f"{self.quote_schema(effective_schema)}.{name}" + return name + + +class ReflectedNamedType(TypedDict): + """Represents a reflected named type.""" + + name: str + """Name of the type.""" + schema: str + """The schema of the type.""" + visible: bool + """Indicates if this type is in the current search path.""" + + +class ReflectedDomainConstraint(TypedDict): + """Represents a reflect check constraint of a domain.""" + + name: str + """Name of the constraint.""" + check: str + """The check constraint text.""" + + +class ReflectedDomain(ReflectedNamedType): + """Represents a reflected enum.""" + + type: str + """The string name of the underlying data type of the domain.""" + nullable: bool + """Indicates if the domain allows null or not.""" + default: Optional[str] + """The string representation of the default value of this domain + or ``None`` if none present. + """ + constraints: List[ReflectedDomainConstraint] + """The constraints defined in the domain, if any. + The constraint are in order of evaluation by postgresql. + """ + collation: Optional[str] + """The collation for the domain.""" + + +class ReflectedEnum(ReflectedNamedType): + """Represents a reflected enum.""" + + labels: List[str] + """The labels that compose the enum.""" + + +class PGInspector(reflection.Inspector): + dialect: PGDialect + + def get_table_oid( + self, table_name: str, schema: Optional[str] = None + ) -> int: + """Return the OID for the given table name. + + :param table_name: string name of the table. For special quoting, + use :class:`.quoted_name`. + + :param schema: string schema name; if omitted, uses the default schema + of the database connection. For special quoting, + use :class:`.quoted_name`. + + """ + + with self._operation_context() as conn: + return self.dialect.get_table_oid( + conn, table_name, schema, info_cache=self.info_cache + ) + + def get_domains( + self, schema: Optional[str] = None + ) -> List[ReflectedDomain]: + """Return a list of DOMAIN objects. + + Each member is a dictionary containing these fields: + + * name - name of the domain + * schema - the schema name for the domain. + * visible - boolean, whether or not this domain is visible + in the default search path. + * type - the type defined by this domain. + * nullable - Indicates if this domain can be ``NULL``. + * default - The default value of the domain or ``None`` if the + domain has no default. + * constraints - A list of dict wit the constraint defined by this + domain. Each element constaints two keys: ``name`` of the + constraint and ``check`` with the constraint text. + + :param schema: schema name. If None, the default schema + (typically 'public') is used. May also be set to ``'*'`` to + indicate load domains for all schemas. + + .. versionadded:: 2.0 + + """ + with self._operation_context() as conn: + return self.dialect._load_domains( + conn, schema, info_cache=self.info_cache + ) + + def get_enums(self, schema: Optional[str] = None) -> List[ReflectedEnum]: + """Return a list of ENUM objects. + + Each member is a dictionary containing these fields: + + * name - name of the enum + * schema - the schema name for the enum. + * visible - boolean, whether or not this enum is visible + in the default search path. + * labels - a list of string labels that apply to the enum. + + :param schema: schema name. If None, the default schema + (typically 'public') is used. May also be set to ``'*'`` to + indicate load enums for all schemas. + + """ + with self._operation_context() as conn: + return self.dialect._load_enums( + conn, schema, info_cache=self.info_cache + ) + + def get_foreign_table_names( + self, schema: Optional[str] = None + ) -> List[str]: + """Return a list of FOREIGN TABLE names. + + Behavior is similar to that of + :meth:`_reflection.Inspector.get_table_names`, + except that the list is limited to those tables that report a + ``relkind`` value of ``f``. + + """ + with self._operation_context() as conn: + return self.dialect._get_foreign_table_names( + conn, schema, info_cache=self.info_cache + ) + + def has_type( + self, type_name: str, schema: Optional[str] = None, **kw: Any + ) -> bool: + """Return if the database has the specified type in the provided + schema. + + :param type_name: the type to check. + :param schema: schema name. If None, the default schema + (typically 'public') is used. May also be set to ``'*'`` to + check in all schemas. + + .. versionadded:: 2.0 + + """ + with self._operation_context() as conn: + return self.dialect.has_type( + conn, type_name, schema, info_cache=self.info_cache + ) + + +class PGExecutionContext(default.DefaultExecutionContext): + def fire_sequence(self, seq, type_): + return self._execute_scalar( + ( + "select nextval('%s')" + % self.identifier_preparer.format_sequence(seq) + ), + type_, + ) + + def get_insert_default(self, column): + if column.primary_key and column is column.table._autoincrement_column: + if column.server_default and column.server_default.has_argument: + # pre-execute passive defaults on primary key columns + return self._execute_scalar( + "select %s" % column.server_default.arg, column.type + ) + + elif column.default is None or ( + column.default.is_sequence and column.default.optional + ): + # execute the sequence associated with a SERIAL primary + # key column. for non-primary-key SERIAL, the ID just + # generates server side. + + try: + seq_name = column._postgresql_seq_name + except AttributeError: + tab = column.table.name + col = column.name + tab = tab[0 : 29 + max(0, (29 - len(col)))] + col = col[0 : 29 + max(0, (29 - len(tab)))] + name = "%s_%s_seq" % (tab, col) + column._postgresql_seq_name = seq_name = name + + if column.table is not None: + effective_schema = self.connection.schema_for_object( + column.table + ) + else: + effective_schema = None + + if effective_schema is not None: + exc = 'select nextval(\'"%s"."%s"\')' % ( + effective_schema, + seq_name, + ) + else: + exc = "select nextval('\"%s\"')" % (seq_name,) + + return self._execute_scalar(exc, column.type) + + return super().get_insert_default(column) + + +class PGReadOnlyConnectionCharacteristic( + characteristics.ConnectionCharacteristic +): + transactional = True + + def reset_characteristic(self, dialect, dbapi_conn): + dialect.set_readonly(dbapi_conn, False) + + def set_characteristic(self, dialect, dbapi_conn, value): + dialect.set_readonly(dbapi_conn, value) + + def get_characteristic(self, dialect, dbapi_conn): + return dialect.get_readonly(dbapi_conn) + + +class PGDeferrableConnectionCharacteristic( + characteristics.ConnectionCharacteristic +): + transactional = True + + def reset_characteristic(self, dialect, dbapi_conn): + dialect.set_deferrable(dbapi_conn, False) + + def set_characteristic(self, dialect, dbapi_conn, value): + dialect.set_deferrable(dbapi_conn, value) + + def get_characteristic(self, dialect, dbapi_conn): + return dialect.get_deferrable(dbapi_conn) + + +class PGDialect(default.DefaultDialect): + name = "postgresql" + supports_statement_cache = True + supports_alter = True + max_identifier_length = 63 + supports_sane_rowcount = True + + bind_typing = interfaces.BindTyping.RENDER_CASTS + + supports_native_enum = True + supports_native_boolean = True + supports_native_uuid = True + supports_smallserial = True + + supports_sequences = True + sequences_optional = True + preexecute_autoincrement_sequences = True + postfetch_lastrowid = False + use_insertmanyvalues = True + + returns_native_bytes = True + + insertmanyvalues_implicit_sentinel = ( + InsertmanyvaluesSentinelOpts.ANY_AUTOINCREMENT + | InsertmanyvaluesSentinelOpts.USE_INSERT_FROM_SELECT + | InsertmanyvaluesSentinelOpts.RENDER_SELECT_COL_CASTS + ) + + supports_comments = True + supports_constraint_comments = True + supports_default_values = True + + supports_default_metavalue = True + + supports_empty_insert = False + supports_multivalues_insert = True + + supports_identity_columns = True + + default_paramstyle = "pyformat" + ischema_names = ischema_names + colspecs = colspecs + + statement_compiler = PGCompiler + ddl_compiler = PGDDLCompiler + type_compiler_cls = PGTypeCompiler + preparer = PGIdentifierPreparer + execution_ctx_cls = PGExecutionContext + inspector = PGInspector + + update_returning = True + delete_returning = True + insert_returning = True + update_returning_multifrom = True + delete_returning_multifrom = True + + connection_characteristics = ( + default.DefaultDialect.connection_characteristics + ) + connection_characteristics = connection_characteristics.union( + { + "postgresql_readonly": PGReadOnlyConnectionCharacteristic(), + "postgresql_deferrable": PGDeferrableConnectionCharacteristic(), + } + ) + + construct_arguments = [ + ( + schema.Index, + { + "using": False, + "include": None, + "where": None, + "ops": {}, + "concurrently": False, + "with": {}, + "tablespace": None, + "nulls_not_distinct": None, + }, + ), + ( + schema.Table, + { + "ignore_search_path": False, + "tablespace": None, + "partition_by": None, + "with_oids": None, + "on_commit": None, + "inherits": None, + "using": None, + }, + ), + ( + schema.CheckConstraint, + { + "not_valid": False, + }, + ), + ( + schema.ForeignKeyConstraint, + { + "not_valid": False, + }, + ), + ( + schema.UniqueConstraint, + {"nulls_not_distinct": None}, + ), + ] + + reflection_options = ("postgresql_ignore_search_path",) + + _backslash_escapes = True + _supports_create_index_concurrently = True + _supports_drop_index_concurrently = True + + def __init__( + self, + native_inet_types=None, + json_serializer=None, + json_deserializer=None, + **kwargs, + ): + default.DefaultDialect.__init__(self, **kwargs) + + self._native_inet_types = native_inet_types + self._json_deserializer = json_deserializer + self._json_serializer = json_serializer + + def initialize(self, connection): + super().initialize(connection) + + # https://www.postgresql.org/docs/9.3/static/release-9-2.html#AEN116689 + self.supports_smallserial = self.server_version_info >= (9, 2) + + self._set_backslash_escapes(connection) + + self._supports_drop_index_concurrently = self.server_version_info >= ( + 9, + 2, + ) + self.supports_identity_columns = self.server_version_info >= (10,) + + def get_isolation_level_values(self, dbapi_conn): + # note the generic dialect doesn't have AUTOCOMMIT, however + # all postgresql dialects should include AUTOCOMMIT. + return ( + "SERIALIZABLE", + "READ UNCOMMITTED", + "READ COMMITTED", + "REPEATABLE READ", + ) + + def set_isolation_level(self, dbapi_connection, level): + cursor = dbapi_connection.cursor() + cursor.execute( + "SET SESSION CHARACTERISTICS AS TRANSACTION " + f"ISOLATION LEVEL {level}" + ) + cursor.execute("COMMIT") + cursor.close() + + def get_isolation_level(self, dbapi_connection): + cursor = dbapi_connection.cursor() + cursor.execute("show transaction isolation level") + val = cursor.fetchone()[0] + cursor.close() + return val.upper() + + def set_readonly(self, connection, value): + raise NotImplementedError() + + def get_readonly(self, connection): + raise NotImplementedError() + + def set_deferrable(self, connection, value): + raise NotImplementedError() + + def get_deferrable(self, connection): + raise NotImplementedError() + + def _split_multihost_from_url(self, url: URL) -> Union[ + Tuple[None, None], + Tuple[Tuple[Optional[str], ...], Tuple[Optional[int], ...]], + ]: + hosts: Optional[Tuple[Optional[str], ...]] = None + ports_str: Union[str, Tuple[Optional[str], ...], None] = None + + integrated_multihost = False + + if "host" in url.query: + if isinstance(url.query["host"], (list, tuple)): + integrated_multihost = True + hosts, ports_str = zip( + *[ + token.split(":") if ":" in token else (token, None) + for token in url.query["host"] + ] + ) + + elif isinstance(url.query["host"], str): + hosts = tuple(url.query["host"].split(",")) + + if ( + "port" not in url.query + and len(hosts) == 1 + and ":" in hosts[0] + ): + # internet host is alphanumeric plus dots or hyphens. + # this is essentially rfc1123, which refers to rfc952. + # https://stackoverflow.com/questions/3523028/ + # valid-characters-of-a-hostname + host_port_match = re.match( + r"^([a-zA-Z0-9\-\.]*)(?:\:(\d*))?$", hosts[0] + ) + if host_port_match: + integrated_multihost = True + h, p = host_port_match.group(1, 2) + if TYPE_CHECKING: + assert isinstance(h, str) + assert isinstance(p, str) + hosts = (h,) + ports_str = cast( + "Tuple[Optional[str], ...]", (p,) if p else (None,) + ) + + if "port" in url.query: + if integrated_multihost: + raise exc.ArgumentError( + "Can't mix 'multihost' formats together; use " + '"host=h1,h2,h3&port=p1,p2,p3" or ' + '"host=h1:p1&host=h2:p2&host=h3:p3" separately' + ) + if isinstance(url.query["port"], (list, tuple)): + ports_str = url.query["port"] + elif isinstance(url.query["port"], str): + ports_str = tuple(url.query["port"].split(",")) + + ports: Optional[Tuple[Optional[int], ...]] = None + + if ports_str: + try: + ports = tuple(int(x) if x else None for x in ports_str) + except ValueError: + raise exc.ArgumentError( + f"Received non-integer port arguments: {ports_str}" + ) from None + + if ports and ( + (not hosts and len(ports) > 1) + or ( + hosts + and ports + and len(hosts) != len(ports) + and (len(hosts) > 1 or len(ports) > 1) + ) + ): + raise exc.ArgumentError("number of hosts and ports don't match") + + if hosts is not None: + if ports is None: + ports = tuple(None for _ in hosts) + + return hosts, ports # type: ignore + + def do_begin_twophase(self, connection, xid): + self.do_begin(connection.connection) + + def do_prepare_twophase(self, connection, xid): + connection.exec_driver_sql("PREPARE TRANSACTION '%s'" % xid) + + def do_rollback_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + if is_prepared: + if recover: + # FIXME: ugly hack to get out of transaction + # context when committing recoverable transactions + # Must find out a way how to make the dbapi not + # open a transaction. + connection.exec_driver_sql("ROLLBACK") + connection.exec_driver_sql("ROLLBACK PREPARED '%s'" % xid) + connection.exec_driver_sql("BEGIN") + self.do_rollback(connection.connection) + else: + self.do_rollback(connection.connection) + + def do_commit_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + if is_prepared: + if recover: + connection.exec_driver_sql("ROLLBACK") + connection.exec_driver_sql("COMMIT PREPARED '%s'" % xid) + connection.exec_driver_sql("BEGIN") + self.do_rollback(connection.connection) + else: + self.do_commit(connection.connection) + + def do_recover_twophase(self, connection): + return connection.scalars( + sql.text("SELECT gid FROM pg_prepared_xacts") + ).all() + + def _get_default_schema_name(self, connection): + return connection.exec_driver_sql("select current_schema()").scalar() + + @reflection.cache + def has_schema(self, connection, schema, **kw): + query = select(pg_catalog.pg_namespace.c.nspname).where( + pg_catalog.pg_namespace.c.nspname == schema + ) + return bool(connection.scalar(query)) + + def _pg_class_filter_scope_schema( + self, query, schema, scope, pg_class_table=None + ): + if pg_class_table is None: + pg_class_table = pg_catalog.pg_class + query = query.join( + pg_catalog.pg_namespace, + pg_catalog.pg_namespace.c.oid == pg_class_table.c.relnamespace, + ) + + if scope is ObjectScope.DEFAULT: + query = query.where(pg_class_table.c.relpersistence != "t") + elif scope is ObjectScope.TEMPORARY: + query = query.where(pg_class_table.c.relpersistence == "t") + + if schema is None: + query = query.where( + pg_catalog.pg_table_is_visible(pg_class_table.c.oid), + # ignore pg_catalog schema + pg_catalog.pg_namespace.c.nspname != "pg_catalog", + ) + else: + query = query.where(pg_catalog.pg_namespace.c.nspname == schema) + return query + + def _pg_class_relkind_condition(self, relkinds, pg_class_table=None): + if pg_class_table is None: + pg_class_table = pg_catalog.pg_class + # uses the any form instead of in otherwise postgresql complaings + # that 'IN could not convert type character to "char"' + return pg_class_table.c.relkind == sql.any_(_array.array(relkinds)) + + @lru_cache() + def _has_table_query(self, schema): + query = select(pg_catalog.pg_class.c.relname).where( + pg_catalog.pg_class.c.relname == bindparam("table_name"), + self._pg_class_relkind_condition( + pg_catalog.RELKINDS_ALL_TABLE_LIKE + ), + ) + return self._pg_class_filter_scope_schema( + query, schema, scope=ObjectScope.ANY + ) + + @reflection.cache + def has_table(self, connection, table_name, schema=None, **kw): + self._ensure_has_table_connection(connection) + query = self._has_table_query(schema) + return bool(connection.scalar(query, {"table_name": table_name})) + + @reflection.cache + def has_sequence(self, connection, sequence_name, schema=None, **kw): + query = select(pg_catalog.pg_class.c.relname).where( + pg_catalog.pg_class.c.relkind == "S", + pg_catalog.pg_class.c.relname == sequence_name, + ) + query = self._pg_class_filter_scope_schema( + query, schema, scope=ObjectScope.ANY + ) + return bool(connection.scalar(query)) + + @reflection.cache + def has_type(self, connection, type_name, schema=None, **kw): + query = ( + select(pg_catalog.pg_type.c.typname) + .join( + pg_catalog.pg_namespace, + pg_catalog.pg_namespace.c.oid + == pg_catalog.pg_type.c.typnamespace, + ) + .where(pg_catalog.pg_type.c.typname == type_name) + ) + if schema is None: + query = query.where( + pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid), + # ignore pg_catalog schema + pg_catalog.pg_namespace.c.nspname != "pg_catalog", + ) + elif schema != "*": + query = query.where(pg_catalog.pg_namespace.c.nspname == schema) + + return bool(connection.scalar(query)) + + def _get_server_version_info(self, connection): + v = connection.exec_driver_sql("select pg_catalog.version()").scalar() + m = re.match( + r".*(?:PostgreSQL|EnterpriseDB) " + r"(\d+)\.?(\d+)?(?:\.(\d+))?(?:\.\d+)?(?:devel|beta)?", + v, + ) + if not m: + raise AssertionError( + "Could not determine version from string '%s'" % v + ) + return tuple([int(x) for x in m.group(1, 2, 3) if x is not None]) + + @reflection.cache + def get_table_oid(self, connection, table_name, schema=None, **kw): + """Fetch the oid for schema.table_name.""" + query = select(pg_catalog.pg_class.c.oid).where( + pg_catalog.pg_class.c.relname == table_name, + self._pg_class_relkind_condition( + pg_catalog.RELKINDS_ALL_TABLE_LIKE + ), + ) + query = self._pg_class_filter_scope_schema( + query, schema, scope=ObjectScope.ANY + ) + table_oid = connection.scalar(query) + if table_oid is None: + raise exc.NoSuchTableError( + f"{schema}.{table_name}" if schema else table_name + ) + return table_oid + + @reflection.cache + def get_schema_names(self, connection, **kw): + query = ( + select(pg_catalog.pg_namespace.c.nspname) + .where(pg_catalog.pg_namespace.c.nspname.not_like("pg_%")) + .order_by(pg_catalog.pg_namespace.c.nspname) + ) + return connection.scalars(query).all() + + def _get_relnames_for_relkinds(self, connection, schema, relkinds, scope): + query = select(pg_catalog.pg_class.c.relname).where( + self._pg_class_relkind_condition(relkinds) + ) + query = self._pg_class_filter_scope_schema(query, schema, scope=scope) + return connection.scalars(query).all() + + @reflection.cache + def get_table_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, + schema, + pg_catalog.RELKINDS_TABLE_NO_FOREIGN, + scope=ObjectScope.DEFAULT, + ) + + @reflection.cache + def get_temp_table_names(self, connection, **kw): + return self._get_relnames_for_relkinds( + connection, + schema=None, + relkinds=pg_catalog.RELKINDS_TABLE_NO_FOREIGN, + scope=ObjectScope.TEMPORARY, + ) + + @reflection.cache + def _get_foreign_table_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, schema, relkinds=("f",), scope=ObjectScope.ANY + ) + + @reflection.cache + def get_view_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, + schema, + pg_catalog.RELKINDS_VIEW, + scope=ObjectScope.DEFAULT, + ) + + @reflection.cache + def get_materialized_view_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, + schema, + pg_catalog.RELKINDS_MAT_VIEW, + scope=ObjectScope.DEFAULT, + ) + + @reflection.cache + def get_temp_view_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, + schema, + # NOTE: do not include temp materialzied views (that do not + # seem to be a thing at least up to version 14) + pg_catalog.RELKINDS_VIEW, + scope=ObjectScope.TEMPORARY, + ) + + @reflection.cache + def get_sequence_names(self, connection, schema=None, **kw): + return self._get_relnames_for_relkinds( + connection, schema, relkinds=("S",), scope=ObjectScope.ANY + ) + + @reflection.cache + def get_view_definition(self, connection, view_name, schema=None, **kw): + query = ( + select(pg_catalog.pg_get_viewdef(pg_catalog.pg_class.c.oid)) + .select_from(pg_catalog.pg_class) + .where( + pg_catalog.pg_class.c.relname == view_name, + self._pg_class_relkind_condition( + pg_catalog.RELKINDS_VIEW + pg_catalog.RELKINDS_MAT_VIEW + ), + ) + ) + query = self._pg_class_filter_scope_schema( + query, schema, scope=ObjectScope.ANY + ) + res = connection.scalar(query) + if res is None: + raise exc.NoSuchTableError( + f"{schema}.{view_name}" if schema else view_name + ) + else: + return res + + def _value_or_raise(self, data, table, schema): + try: + return dict(data)[(schema, table)] + except KeyError: + raise exc.NoSuchTableError( + f"{schema}.{table}" if schema else table + ) from None + + def _prepare_filter_names(self, filter_names): + if filter_names: + return True, {"filter_names": filter_names} + else: + return False, {} + + def _kind_to_relkinds(self, kind: ObjectKind) -> Tuple[str, ...]: + if kind is ObjectKind.ANY: + return pg_catalog.RELKINDS_ALL_TABLE_LIKE + relkinds = () + if ObjectKind.TABLE in kind: + relkinds += pg_catalog.RELKINDS_TABLE + if ObjectKind.VIEW in kind: + relkinds += pg_catalog.RELKINDS_VIEW + if ObjectKind.MATERIALIZED_VIEW in kind: + relkinds += pg_catalog.RELKINDS_MAT_VIEW + return relkinds + + @reflection.cache + def get_columns(self, connection, table_name, schema=None, **kw): + data = self.get_multi_columns( + connection, + schema=schema, + filter_names=[table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + @lru_cache() + def _columns_query(self, schema, has_filter_names, scope, kind): + # NOTE: the query with the default and identity options scalar + # subquery is faster than trying to use outer joins for them + generated = ( + pg_catalog.pg_attribute.c.attgenerated.label("generated") + if self.server_version_info >= (12,) + else sql.null().label("generated") + ) + if self.server_version_info >= (10,): + # join lateral performs worse (~2x slower) than a scalar_subquery + identity = ( + select( + sql.func.json_build_object( + "always", + pg_catalog.pg_attribute.c.attidentity == "a", + "start", + pg_catalog.pg_sequence.c.seqstart, + "increment", + pg_catalog.pg_sequence.c.seqincrement, + "minvalue", + pg_catalog.pg_sequence.c.seqmin, + "maxvalue", + pg_catalog.pg_sequence.c.seqmax, + "cache", + pg_catalog.pg_sequence.c.seqcache, + "cycle", + pg_catalog.pg_sequence.c.seqcycle, + ) + ) + .select_from(pg_catalog.pg_sequence) + .where( + # attidentity != '' is required or it will reflect also + # serial columns as identity. + pg_catalog.pg_attribute.c.attidentity != "", + pg_catalog.pg_sequence.c.seqrelid + == sql.cast( + sql.cast( + pg_catalog.pg_get_serial_sequence( + sql.cast( + sql.cast( + pg_catalog.pg_attribute.c.attrelid, + REGCLASS, + ), + TEXT, + ), + pg_catalog.pg_attribute.c.attname, + ), + REGCLASS, + ), + OID, + ), + ) + .correlate(pg_catalog.pg_attribute) + .scalar_subquery() + .label("identity_options") + ) + else: + identity = sql.null().label("identity_options") + + # join lateral performs the same as scalar_subquery here + default = ( + select( + pg_catalog.pg_get_expr( + pg_catalog.pg_attrdef.c.adbin, + pg_catalog.pg_attrdef.c.adrelid, + ) + ) + .select_from(pg_catalog.pg_attrdef) + .where( + pg_catalog.pg_attrdef.c.adrelid + == pg_catalog.pg_attribute.c.attrelid, + pg_catalog.pg_attrdef.c.adnum + == pg_catalog.pg_attribute.c.attnum, + pg_catalog.pg_attribute.c.atthasdef, + ) + .correlate(pg_catalog.pg_attribute) + .scalar_subquery() + .label("default") + ) + relkinds = self._kind_to_relkinds(kind) + query = ( + select( + pg_catalog.pg_attribute.c.attname.label("name"), + pg_catalog.format_type( + pg_catalog.pg_attribute.c.atttypid, + pg_catalog.pg_attribute.c.atttypmod, + ).label("format_type"), + default, + pg_catalog.pg_attribute.c.attnotnull.label("not_null"), + pg_catalog.pg_class.c.relname.label("table_name"), + pg_catalog.pg_description.c.description.label("comment"), + generated, + identity, + ) + .select_from(pg_catalog.pg_class) + # NOTE: postgresql support table with no user column, meaning + # there is no row with pg_attribute.attnum > 0. use a left outer + # join to avoid filtering these tables. + .outerjoin( + pg_catalog.pg_attribute, + sql.and_( + pg_catalog.pg_class.c.oid + == pg_catalog.pg_attribute.c.attrelid, + pg_catalog.pg_attribute.c.attnum > 0, + ~pg_catalog.pg_attribute.c.attisdropped, + ), + ) + .outerjoin( + pg_catalog.pg_description, + sql.and_( + pg_catalog.pg_description.c.objoid + == pg_catalog.pg_attribute.c.attrelid, + pg_catalog.pg_description.c.objsubid + == pg_catalog.pg_attribute.c.attnum, + ), + ) + .where(self._pg_class_relkind_condition(relkinds)) + .order_by( + pg_catalog.pg_class.c.relname, pg_catalog.pg_attribute.c.attnum + ) + ) + query = self._pg_class_filter_scope_schema(query, schema, scope=scope) + if has_filter_names: + query = query.where( + pg_catalog.pg_class.c.relname.in_(bindparam("filter_names")) + ) + return query + + def get_multi_columns( + self, connection, schema, filter_names, scope, kind, **kw + ): + has_filter_names, params = self._prepare_filter_names(filter_names) + query = self._columns_query(schema, has_filter_names, scope, kind) + rows = connection.execute(query, params).mappings() + + # dictionary with (name, ) if default search path or (schema, name) + # as keys + domains = { + ((d["schema"], d["name"]) if not d["visible"] else (d["name"],)): d + for d in self._load_domains( + connection, schema="*", info_cache=kw.get("info_cache") + ) + } + + # dictionary with (name, ) if default search path or (schema, name) + # as keys + enums = dict( + ( + ((rec["name"],), rec) + if rec["visible"] + else ((rec["schema"], rec["name"]), rec) + ) + for rec in self._load_enums( + connection, schema="*", info_cache=kw.get("info_cache") + ) + ) + + columns = self._get_columns_info(rows, domains, enums, schema) + + return columns.items() + + _format_type_args_pattern = re.compile(r"\((.*)\)") + _format_type_args_delim = re.compile(r"\s*,\s*") + _format_array_spec_pattern = re.compile(r"((?:\[\])*)$") + + def _reflect_type( + self, + format_type: Optional[str], + domains: dict[str, ReflectedDomain], + enums: dict[str, ReflectedEnum], + type_description: str, + ) -> sqltypes.TypeEngine[Any]: + """ + Attempts to reconstruct a column type defined in ischema_names based + on the information available in the format_type. + + If the `format_type` cannot be associated with a known `ischema_names`, + it is treated as a reference to a known PostgreSQL named `ENUM` or + `DOMAIN` type. + """ + type_description = type_description or "unknown type" + if format_type is None: + util.warn( + "PostgreSQL format_type() returned NULL for %s" + % type_description + ) + return sqltypes.NULLTYPE + + attype_args_match = self._format_type_args_pattern.search(format_type) + if attype_args_match and attype_args_match.group(1): + attype_args = self._format_type_args_delim.split( + attype_args_match.group(1) + ) + else: + attype_args = () + + match_array_dim = self._format_array_spec_pattern.search(format_type) + # Each "[]" in array specs corresponds to an array dimension + array_dim = len(match_array_dim.group(1) or "") // 2 + + # Remove all parameters and array specs from format_type to obtain an + # ischema_name candidate + attype = self._format_type_args_pattern.sub("", format_type) + attype = self._format_array_spec_pattern.sub("", attype) + + schema_type = self.ischema_names.get(attype.lower(), None) + args, kwargs = (), {} + + if attype == "numeric": + if len(attype_args) == 2: + precision, scale = map(int, attype_args) + args = (precision, scale) + + elif attype == "double precision": + args = (53,) + + elif attype == "integer": + args = () + + elif attype in ("timestamp with time zone", "time with time zone"): + kwargs["timezone"] = True + if len(attype_args) == 1: + kwargs["precision"] = int(attype_args[0]) + + elif attype in ( + "timestamp without time zone", + "time without time zone", + "time", + ): + kwargs["timezone"] = False + if len(attype_args) == 1: + kwargs["precision"] = int(attype_args[0]) + + elif attype == "bit varying": + kwargs["varying"] = True + if len(attype_args) == 1: + charlen = int(attype_args[0]) + args = (charlen,) + + elif attype.startswith("interval"): + schema_type = INTERVAL + + field_match = re.match(r"interval (.+)", attype) + if field_match: + kwargs["fields"] = field_match.group(1) + + if len(attype_args) == 1: + kwargs["precision"] = int(attype_args[0]) + + else: + enum_or_domain_key = tuple(util.quoted_token_parser(attype)) + + if enum_or_domain_key in enums: + schema_type = ENUM + enum = enums[enum_or_domain_key] + + args = tuple(enum["labels"]) + kwargs["name"] = enum["name"] + + if not enum["visible"]: + kwargs["schema"] = enum["schema"] + args = tuple(enum["labels"]) + elif enum_or_domain_key in domains: + schema_type = DOMAIN + domain = domains[enum_or_domain_key] + + data_type = self._reflect_type( + domain["type"], + domains, + enums, + type_description="DOMAIN '%s'" % domain["name"], + ) + args = (domain["name"], data_type) + + kwargs["collation"] = domain["collation"] + kwargs["default"] = domain["default"] + kwargs["not_null"] = not domain["nullable"] + kwargs["create_type"] = False + + if domain["constraints"]: + # We only support a single constraint + check_constraint = domain["constraints"][0] + + kwargs["constraint_name"] = check_constraint["name"] + kwargs["check"] = check_constraint["check"] + + if not domain["visible"]: + kwargs["schema"] = domain["schema"] + + else: + try: + charlen = int(attype_args[0]) + args = (charlen, *attype_args[1:]) + except (ValueError, IndexError): + args = attype_args + + if not schema_type: + util.warn( + "Did not recognize type '%s' of %s" + % (attype, type_description) + ) + return sqltypes.NULLTYPE + + data_type = schema_type(*args, **kwargs) + if array_dim >= 1: + # postgres does not preserve dimensionality or size of array types. + data_type = _array.ARRAY(data_type) + + return data_type + + def _get_columns_info(self, rows, domains, enums, schema): + columns = defaultdict(list) + for row_dict in rows: + # ensure that each table has an entry, even if it has no columns + if row_dict["name"] is None: + columns[(schema, row_dict["table_name"])] = ( + ReflectionDefaults.columns() + ) + continue + table_cols = columns[(schema, row_dict["table_name"])] + + coltype = self._reflect_type( + row_dict["format_type"], + domains, + enums, + type_description="column '%s'" % row_dict["name"], + ) + + default = row_dict["default"] + name = row_dict["name"] + generated = row_dict["generated"] + nullable = not row_dict["not_null"] + + if isinstance(coltype, DOMAIN): + if not default: + # domain can override the default value but + # cant set it to None + if coltype.default is not None: + default = coltype.default + + nullable = nullable and not coltype.not_null + + identity = row_dict["identity_options"] + + # If a zero byte or blank string depending on driver (is also + # absent for older PG versions), then not a generated column. + # Otherwise, s = stored. (Other values might be added in the + # future.) + if generated not in (None, "", b"\x00"): + computed = dict( + sqltext=default, persisted=generated in ("s", b"s") + ) + default = None + else: + computed = None + + # adjust the default value + autoincrement = False + if default is not None: + match = re.search(r"""(nextval\(')([^']+)('.*$)""", default) + if match is not None: + if issubclass(coltype._type_affinity, sqltypes.Integer): + autoincrement = True + # the default is related to a Sequence + if "." not in match.group(2) and schema is not None: + # unconditionally quote the schema name. this could + # later be enhanced to obey quoting rules / + # "quote schema" + default = ( + match.group(1) + + ('"%s"' % schema) + + "." + + match.group(2) + + match.group(3) + ) + + column_info = { + "name": name, + "type": coltype, + "nullable": nullable, + "default": default, + "autoincrement": autoincrement or identity is not None, + "comment": row_dict["comment"], + } + if computed is not None: + column_info["computed"] = computed + if identity is not None: + column_info["identity"] = identity + + table_cols.append(column_info) + + return columns + + @lru_cache() + def _table_oids_query(self, schema, has_filter_names, scope, kind): + relkinds = self._kind_to_relkinds(kind) + oid_q = select( + pg_catalog.pg_class.c.oid, pg_catalog.pg_class.c.relname + ).where(self._pg_class_relkind_condition(relkinds)) + oid_q = self._pg_class_filter_scope_schema(oid_q, schema, scope=scope) + + if has_filter_names: + oid_q = oid_q.where( + pg_catalog.pg_class.c.relname.in_(bindparam("filter_names")) + ) + return oid_q + + @reflection.flexi_cache( + ("schema", InternalTraversal.dp_string), + ("filter_names", InternalTraversal.dp_string_list), + ("kind", InternalTraversal.dp_plain_obj), + ("scope", InternalTraversal.dp_plain_obj), + ) + def _get_table_oids( + self, connection, schema, filter_names, scope, kind, **kw + ): + has_filter_names, params = self._prepare_filter_names(filter_names) + oid_q = self._table_oids_query(schema, has_filter_names, scope, kind) + result = connection.execute(oid_q, params) + return result.all() + + @lru_cache() + def _constraint_query(self, is_unique): + con_sq = ( + select( + pg_catalog.pg_constraint.c.conrelid, + pg_catalog.pg_constraint.c.conname, + pg_catalog.pg_constraint.c.conindid, + sql.func.unnest(pg_catalog.pg_constraint.c.conkey).label( + "attnum" + ), + sql.func.generate_subscripts( + pg_catalog.pg_constraint.c.conkey, 1 + ).label("ord"), + pg_catalog.pg_description.c.description, + ) + .outerjoin( + pg_catalog.pg_description, + pg_catalog.pg_description.c.objoid + == pg_catalog.pg_constraint.c.oid, + ) + .where( + pg_catalog.pg_constraint.c.contype == bindparam("contype"), + pg_catalog.pg_constraint.c.conrelid.in_(bindparam("oids")), + ) + .subquery("con") + ) + + attr_sq = ( + select( + con_sq.c.conrelid, + con_sq.c.conname, + con_sq.c.conindid, + con_sq.c.description, + con_sq.c.ord, + pg_catalog.pg_attribute.c.attname, + ) + .select_from(pg_catalog.pg_attribute) + .join( + con_sq, + sql.and_( + pg_catalog.pg_attribute.c.attnum == con_sq.c.attnum, + pg_catalog.pg_attribute.c.attrelid == con_sq.c.conrelid, + ), + ) + .where( + # NOTE: restate the condition here, since pg15 otherwise + # seems to get confused on pscopg2 sometimes, doing + # a sequential scan of pg_attribute. + # The condition in the con_sq subquery is not actually needed + # in pg15, but it may be needed in older versions. Keeping it + # does not seems to have any inpact in any case. + con_sq.c.conrelid.in_(bindparam("oids")) + ) + .subquery("attr") + ) + + constraint_query = ( + select( + attr_sq.c.conrelid, + sql.func.array_agg( + # NOTE: cast since some postgresql derivatives may + # not support array_agg on the name type + aggregate_order_by( + attr_sq.c.attname.cast(TEXT), attr_sq.c.ord + ) + ).label("cols"), + attr_sq.c.conname, + sql.func.min(attr_sq.c.description).label("description"), + ) + .group_by(attr_sq.c.conrelid, attr_sq.c.conname) + .order_by(attr_sq.c.conrelid, attr_sq.c.conname) + ) + + if is_unique: + if self.server_version_info >= (15,): + constraint_query = constraint_query.join( + pg_catalog.pg_index, + attr_sq.c.conindid == pg_catalog.pg_index.c.indexrelid, + ).add_columns( + sql.func.bool_and( + pg_catalog.pg_index.c.indnullsnotdistinct + ).label("indnullsnotdistinct") + ) + else: + constraint_query = constraint_query.add_columns( + sql.false().label("indnullsnotdistinct") + ) + else: + constraint_query = constraint_query.add_columns( + sql.null().label("extra") + ) + return constraint_query + + def _reflect_constraint( + self, connection, contype, schema, filter_names, scope, kind, **kw + ): + # used to reflect primary and unique constraint + table_oids = self._get_table_oids( + connection, schema, filter_names, scope, kind, **kw + ) + batches = list(table_oids) + is_unique = contype == "u" + + while batches: + batch = batches[0:3000] + batches[0:3000] = [] + + result = connection.execute( + self._constraint_query(is_unique), + {"oids": [r[0] for r in batch], "contype": contype}, + ) + + result_by_oid = defaultdict(list) + for oid, cols, constraint_name, comment, extra in result: + result_by_oid[oid].append( + (cols, constraint_name, comment, extra) + ) + + for oid, tablename in batch: + for_oid = result_by_oid.get(oid, ()) + if for_oid: + for cols, constraint, comment, extra in for_oid: + if is_unique: + yield tablename, cols, constraint, comment, { + "nullsnotdistinct": extra + } + else: + yield tablename, cols, constraint, comment, None + else: + yield tablename, None, None, None, None + + @reflection.cache + def get_pk_constraint(self, connection, table_name, schema=None, **kw): + data = self.get_multi_pk_constraint( + connection, + schema=schema, + filter_names=[table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + def get_multi_pk_constraint( + self, connection, schema, filter_names, scope, kind, **kw + ): + result = self._reflect_constraint( + connection, "p", schema, filter_names, scope, kind, **kw + ) + + # only a single pk can be present for each table. Return an entry + # even if a table has no primary key + default = ReflectionDefaults.pk_constraint + return ( + ( + (schema, table_name), + ( + { + "constrained_columns": [] if cols is None else cols, + "name": pk_name, + "comment": comment, + } + if pk_name is not None + else default() + ), + ) + for table_name, cols, pk_name, comment, _ in result + ) + + @reflection.cache + def get_foreign_keys( + self, + connection, + table_name, + schema=None, + postgresql_ignore_search_path=False, + **kw, + ): + data = self.get_multi_foreign_keys( + connection, + schema=schema, + filter_names=[table_name], + postgresql_ignore_search_path=postgresql_ignore_search_path, + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + @lru_cache() + def _foreing_key_query(self, schema, has_filter_names, scope, kind): + pg_class_ref = pg_catalog.pg_class.alias("cls_ref") + pg_namespace_ref = pg_catalog.pg_namespace.alias("nsp_ref") + relkinds = self._kind_to_relkinds(kind) + query = ( + select( + pg_catalog.pg_class.c.relname, + pg_catalog.pg_constraint.c.conname, + # NOTE: avoid calling pg_get_constraintdef when not needed + # to speed up the query + sql.case( + ( + pg_catalog.pg_constraint.c.oid.is_not(None), + pg_catalog.pg_get_constraintdef( + pg_catalog.pg_constraint.c.oid, True + ), + ), + else_=None, + ), + pg_namespace_ref.c.nspname, + pg_catalog.pg_description.c.description, + ) + .select_from(pg_catalog.pg_class) + .outerjoin( + pg_catalog.pg_constraint, + sql.and_( + pg_catalog.pg_class.c.oid + == pg_catalog.pg_constraint.c.conrelid, + pg_catalog.pg_constraint.c.contype == "f", + ), + ) + .outerjoin( + pg_class_ref, + pg_class_ref.c.oid == pg_catalog.pg_constraint.c.confrelid, + ) + .outerjoin( + pg_namespace_ref, + pg_class_ref.c.relnamespace == pg_namespace_ref.c.oid, + ) + .outerjoin( + pg_catalog.pg_description, + pg_catalog.pg_description.c.objoid + == pg_catalog.pg_constraint.c.oid, + ) + .order_by( + pg_catalog.pg_class.c.relname, + pg_catalog.pg_constraint.c.conname, + ) + .where(self._pg_class_relkind_condition(relkinds)) + ) + query = self._pg_class_filter_scope_schema(query, schema, scope) + if has_filter_names: + query = query.where( + pg_catalog.pg_class.c.relname.in_(bindparam("filter_names")) + ) + return query + + @util.memoized_property + def _fk_regex_pattern(self): + # optionally quoted token + qtoken = '(?:"[^"]+"|[A-Za-z0-9_]+?)' + + # https://www.postgresql.org/docs/current/static/sql-createtable.html + return re.compile( + r"FOREIGN KEY \((.*?)\) " + rf"REFERENCES (?:({qtoken})\.)?({qtoken})\(((?:{qtoken}(?: *, *)?)+)\)" # noqa: E501 + r"[\s]?(MATCH (FULL|PARTIAL|SIMPLE)+)?" + r"[\s]?(ON UPDATE " + r"(CASCADE|RESTRICT|NO ACTION|SET NULL|SET DEFAULT)+)?" + r"[\s]?(ON DELETE " + r"(CASCADE|RESTRICT|NO ACTION|SET NULL|SET DEFAULT)+)?" + r"[\s]?(DEFERRABLE|NOT DEFERRABLE)?" + r"[\s]?(INITIALLY (DEFERRED|IMMEDIATE)+)?" + ) + + def get_multi_foreign_keys( + self, + connection, + schema, + filter_names, + scope, + kind, + postgresql_ignore_search_path=False, + **kw, + ): + preparer = self.identifier_preparer + + has_filter_names, params = self._prepare_filter_names(filter_names) + query = self._foreing_key_query(schema, has_filter_names, scope, kind) + result = connection.execute(query, params) + + FK_REGEX = self._fk_regex_pattern + + fkeys = defaultdict(list) + default = ReflectionDefaults.foreign_keys + for table_name, conname, condef, conschema, comment in result: + # ensure that each table has an entry, even if it has + # no foreign keys + if conname is None: + fkeys[(schema, table_name)] = default() + continue + table_fks = fkeys[(schema, table_name)] + m = re.search(FK_REGEX, condef).groups() + + ( + constrained_columns, + referred_schema, + referred_table, + referred_columns, + _, + match, + _, + onupdate, + _, + ondelete, + deferrable, + _, + initially, + ) = m + + if deferrable is not None: + deferrable = True if deferrable == "DEFERRABLE" else False + constrained_columns = [ + preparer._unquote_identifier(x) + for x in re.split(r"\s*,\s*", constrained_columns) + ] + + if postgresql_ignore_search_path: + # when ignoring search path, we use the actual schema + # provided it isn't the "default" schema + if conschema != self.default_schema_name: + referred_schema = conschema + else: + referred_schema = schema + elif referred_schema: + # referred_schema is the schema that we regexp'ed from + # pg_get_constraintdef(). If the schema is in the search + # path, pg_get_constraintdef() will give us None. + referred_schema = preparer._unquote_identifier(referred_schema) + elif schema is not None and schema == conschema: + # If the actual schema matches the schema of the table + # we're reflecting, then we will use that. + referred_schema = schema + + referred_table = preparer._unquote_identifier(referred_table) + referred_columns = [ + preparer._unquote_identifier(x) + for x in re.split(r"\s*,\s", referred_columns) + ] + options = { + k: v + for k, v in [ + ("onupdate", onupdate), + ("ondelete", ondelete), + ("initially", initially), + ("deferrable", deferrable), + ("match", match), + ] + if v is not None and v != "NO ACTION" + } + fkey_d = { + "name": conname, + "constrained_columns": constrained_columns, + "referred_schema": referred_schema, + "referred_table": referred_table, + "referred_columns": referred_columns, + "options": options, + "comment": comment, + } + table_fks.append(fkey_d) + return fkeys.items() + + @reflection.cache + def get_indexes(self, connection, table_name, schema=None, **kw): + data = self.get_multi_indexes( + connection, + schema=schema, + filter_names=[table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + @util.memoized_property + def _index_query(self): + pg_class_index = pg_catalog.pg_class.alias("cls_idx") + # NOTE: repeating oids clause improve query performance + + # subquery to get the columns + idx_sq = ( + select( + pg_catalog.pg_index.c.indexrelid, + pg_catalog.pg_index.c.indrelid, + sql.func.unnest(pg_catalog.pg_index.c.indkey).label("attnum"), + sql.func.generate_subscripts( + pg_catalog.pg_index.c.indkey, 1 + ).label("ord"), + ) + .where( + ~pg_catalog.pg_index.c.indisprimary, + pg_catalog.pg_index.c.indrelid.in_(bindparam("oids")), + ) + .subquery("idx") + ) + + attr_sq = ( + select( + idx_sq.c.indexrelid, + idx_sq.c.indrelid, + idx_sq.c.ord, + # NOTE: always using pg_get_indexdef is too slow so just + # invoke when the element is an expression + sql.case( + ( + idx_sq.c.attnum == 0, + pg_catalog.pg_get_indexdef( + idx_sq.c.indexrelid, idx_sq.c.ord + 1, True + ), + ), + # NOTE: need to cast this since attname is of type "name" + # that's limited to 63 bytes, while pg_get_indexdef + # returns "text" so its output may get cut + else_=pg_catalog.pg_attribute.c.attname.cast(TEXT), + ).label("element"), + (idx_sq.c.attnum == 0).label("is_expr"), + ) + .select_from(idx_sq) + .outerjoin( + # do not remove rows where idx_sq.c.attnum is 0 + pg_catalog.pg_attribute, + sql.and_( + pg_catalog.pg_attribute.c.attnum == idx_sq.c.attnum, + pg_catalog.pg_attribute.c.attrelid == idx_sq.c.indrelid, + ), + ) + .where(idx_sq.c.indrelid.in_(bindparam("oids"))) + .subquery("idx_attr") + ) + + cols_sq = ( + select( + attr_sq.c.indexrelid, + sql.func.min(attr_sq.c.indrelid), + sql.func.array_agg( + aggregate_order_by(attr_sq.c.element, attr_sq.c.ord) + ).label("elements"), + sql.func.array_agg( + aggregate_order_by(attr_sq.c.is_expr, attr_sq.c.ord) + ).label("elements_is_expr"), + ) + .group_by(attr_sq.c.indexrelid) + .subquery("idx_cols") + ) + + if self.server_version_info >= (11, 0): + indnkeyatts = pg_catalog.pg_index.c.indnkeyatts + else: + indnkeyatts = sql.null().label("indnkeyatts") + + if self.server_version_info >= (15,): + nulls_not_distinct = pg_catalog.pg_index.c.indnullsnotdistinct + else: + nulls_not_distinct = sql.false().label("indnullsnotdistinct") + + return ( + select( + pg_catalog.pg_index.c.indrelid, + pg_class_index.c.relname.label("relname_index"), + pg_catalog.pg_index.c.indisunique, + pg_catalog.pg_constraint.c.conrelid.is_not(None).label( + "has_constraint" + ), + pg_catalog.pg_index.c.indoption, + pg_class_index.c.reloptions, + pg_catalog.pg_am.c.amname, + # NOTE: pg_get_expr is very fast so this case has almost no + # performance impact + sql.case( + ( + pg_catalog.pg_index.c.indpred.is_not(None), + pg_catalog.pg_get_expr( + pg_catalog.pg_index.c.indpred, + pg_catalog.pg_index.c.indrelid, + ), + ), + else_=None, + ).label("filter_definition"), + indnkeyatts, + nulls_not_distinct, + cols_sq.c.elements, + cols_sq.c.elements_is_expr, + ) + .select_from(pg_catalog.pg_index) + .where( + pg_catalog.pg_index.c.indrelid.in_(bindparam("oids")), + ~pg_catalog.pg_index.c.indisprimary, + ) + .join( + pg_class_index, + pg_catalog.pg_index.c.indexrelid == pg_class_index.c.oid, + ) + .join( + pg_catalog.pg_am, + pg_class_index.c.relam == pg_catalog.pg_am.c.oid, + ) + .outerjoin( + cols_sq, + pg_catalog.pg_index.c.indexrelid == cols_sq.c.indexrelid, + ) + .outerjoin( + pg_catalog.pg_constraint, + sql.and_( + pg_catalog.pg_index.c.indrelid + == pg_catalog.pg_constraint.c.conrelid, + pg_catalog.pg_index.c.indexrelid + == pg_catalog.pg_constraint.c.conindid, + pg_catalog.pg_constraint.c.contype + == sql.any_(_array.array(("p", "u", "x"))), + ), + ) + .order_by(pg_catalog.pg_index.c.indrelid, pg_class_index.c.relname) + ) + + def get_multi_indexes( + self, connection, schema, filter_names, scope, kind, **kw + ): + table_oids = self._get_table_oids( + connection, schema, filter_names, scope, kind, **kw + ) + + indexes = defaultdict(list) + default = ReflectionDefaults.indexes + + batches = list(table_oids) + + while batches: + batch = batches[0:3000] + batches[0:3000] = [] + + result = connection.execute( + self._index_query, {"oids": [r[0] for r in batch]} + ).mappings() + + result_by_oid = defaultdict(list) + for row_dict in result: + result_by_oid[row_dict["indrelid"]].append(row_dict) + + for oid, table_name in batch: + if oid not in result_by_oid: + # ensure that each table has an entry, even if reflection + # is skipped because not supported + indexes[(schema, table_name)] = default() + continue + + for row in result_by_oid[oid]: + index_name = row["relname_index"] + + table_indexes = indexes[(schema, table_name)] + + all_elements = row["elements"] + all_elements_is_expr = row["elements_is_expr"] + indnkeyatts = row["indnkeyatts"] + # "The number of key columns in the index, not counting any + # included columns, which are merely stored and do not + # participate in the index semantics" + if indnkeyatts and len(all_elements) > indnkeyatts: + # this is a "covering index" which has INCLUDE columns + # as well as regular index columns + inc_cols = all_elements[indnkeyatts:] + idx_elements = all_elements[:indnkeyatts] + idx_elements_is_expr = all_elements_is_expr[ + :indnkeyatts + ] + # postgresql does not support expression on included + # columns as of v14: "ERROR: expressions are not + # supported in included columns". + assert all( + not is_expr + for is_expr in all_elements_is_expr[indnkeyatts:] + ) + else: + idx_elements = all_elements + idx_elements_is_expr = all_elements_is_expr + inc_cols = [] + + index = {"name": index_name, "unique": row["indisunique"]} + if any(idx_elements_is_expr): + index["column_names"] = [ + None if is_expr else expr + for expr, is_expr in zip( + idx_elements, idx_elements_is_expr + ) + ] + index["expressions"] = idx_elements + else: + index["column_names"] = idx_elements + + sorting = {} + for col_index, col_flags in enumerate(row["indoption"]): + col_sorting = () + # try to set flags only if they differ from PG + # defaults... + if col_flags & 0x01: + col_sorting += ("desc",) + if not (col_flags & 0x02): + col_sorting += ("nulls_last",) + else: + if col_flags & 0x02: + col_sorting += ("nulls_first",) + if col_sorting: + sorting[idx_elements[col_index]] = col_sorting + if sorting: + index["column_sorting"] = sorting + if row["has_constraint"]: + index["duplicates_constraint"] = index_name + + dialect_options = {} + if row["reloptions"]: + dialect_options["postgresql_with"] = dict( + [option.split("=") for option in row["reloptions"]] + ) + # it *might* be nice to include that this is 'btree' in the + # reflection info. But we don't want an Index object + # to have a ``postgresql_using`` in it that is just the + # default, so for the moment leaving this out. + amname = row["amname"] + if amname != "btree": + dialect_options["postgresql_using"] = row["amname"] + if row["filter_definition"]: + dialect_options["postgresql_where"] = row[ + "filter_definition" + ] + if self.server_version_info >= (11,): + # NOTE: this is legacy, this is part of + # dialect_options now as of #7382 + index["include_columns"] = inc_cols + dialect_options["postgresql_include"] = inc_cols + if row["indnullsnotdistinct"]: + # the default is False, so ignore it. + dialect_options["postgresql_nulls_not_distinct"] = row[ + "indnullsnotdistinct" + ] + + if dialect_options: + index["dialect_options"] = dialect_options + + table_indexes.append(index) + return indexes.items() + + @reflection.cache + def get_unique_constraints( + self, connection, table_name, schema=None, **kw + ): + data = self.get_multi_unique_constraints( + connection, + schema=schema, + filter_names=[table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + def get_multi_unique_constraints( + self, + connection, + schema, + filter_names, + scope, + kind, + **kw, + ): + result = self._reflect_constraint( + connection, "u", schema, filter_names, scope, kind, **kw + ) + + # each table can have multiple unique constraints + uniques = defaultdict(list) + default = ReflectionDefaults.unique_constraints + for table_name, cols, con_name, comment, options in result: + # ensure a list is created for each table. leave it empty if + # the table has no unique cosntraint + if con_name is None: + uniques[(schema, table_name)] = default() + continue + + uc_dict = { + "column_names": cols, + "name": con_name, + "comment": comment, + } + if options: + if options["nullsnotdistinct"]: + uc_dict["dialect_options"] = { + "postgresql_nulls_not_distinct": options[ + "nullsnotdistinct" + ] + } + + uniques[(schema, table_name)].append(uc_dict) + return uniques.items() + + @reflection.cache + def get_table_comment(self, connection, table_name, schema=None, **kw): + data = self.get_multi_table_comment( + connection, + schema, + [table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + @lru_cache() + def _comment_query(self, schema, has_filter_names, scope, kind): + relkinds = self._kind_to_relkinds(kind) + query = ( + select( + pg_catalog.pg_class.c.relname, + pg_catalog.pg_description.c.description, + ) + .select_from(pg_catalog.pg_class) + .outerjoin( + pg_catalog.pg_description, + sql.and_( + pg_catalog.pg_class.c.oid + == pg_catalog.pg_description.c.objoid, + pg_catalog.pg_description.c.objsubid == 0, + ), + ) + .where(self._pg_class_relkind_condition(relkinds)) + ) + query = self._pg_class_filter_scope_schema(query, schema, scope) + if has_filter_names: + query = query.where( + pg_catalog.pg_class.c.relname.in_(bindparam("filter_names")) + ) + return query + + def get_multi_table_comment( + self, connection, schema, filter_names, scope, kind, **kw + ): + has_filter_names, params = self._prepare_filter_names(filter_names) + query = self._comment_query(schema, has_filter_names, scope, kind) + result = connection.execute(query, params) + + default = ReflectionDefaults.table_comment + return ( + ( + (schema, table), + {"text": comment} if comment is not None else default(), + ) + for table, comment in result + ) + + @reflection.cache + def get_check_constraints(self, connection, table_name, schema=None, **kw): + data = self.get_multi_check_constraints( + connection, + schema, + [table_name], + scope=ObjectScope.ANY, + kind=ObjectKind.ANY, + **kw, + ) + return self._value_or_raise(data, table_name, schema) + + @lru_cache() + def _check_constraint_query(self, schema, has_filter_names, scope, kind): + relkinds = self._kind_to_relkinds(kind) + query = ( + select( + pg_catalog.pg_class.c.relname, + pg_catalog.pg_constraint.c.conname, + # NOTE: avoid calling pg_get_constraintdef when not needed + # to speed up the query + sql.case( + ( + pg_catalog.pg_constraint.c.oid.is_not(None), + pg_catalog.pg_get_constraintdef( + pg_catalog.pg_constraint.c.oid, True + ), + ), + else_=None, + ), + pg_catalog.pg_description.c.description, + ) + .select_from(pg_catalog.pg_class) + .outerjoin( + pg_catalog.pg_constraint, + sql.and_( + pg_catalog.pg_class.c.oid + == pg_catalog.pg_constraint.c.conrelid, + pg_catalog.pg_constraint.c.contype == "c", + ), + ) + .outerjoin( + pg_catalog.pg_description, + pg_catalog.pg_description.c.objoid + == pg_catalog.pg_constraint.c.oid, + ) + .order_by( + pg_catalog.pg_class.c.relname, + pg_catalog.pg_constraint.c.conname, + ) + .where(self._pg_class_relkind_condition(relkinds)) + ) + query = self._pg_class_filter_scope_schema(query, schema, scope) + if has_filter_names: + query = query.where( + pg_catalog.pg_class.c.relname.in_(bindparam("filter_names")) + ) + return query + + def get_multi_check_constraints( + self, connection, schema, filter_names, scope, kind, **kw + ): + has_filter_names, params = self._prepare_filter_names(filter_names) + query = self._check_constraint_query( + schema, has_filter_names, scope, kind + ) + result = connection.execute(query, params) + + check_constraints = defaultdict(list) + default = ReflectionDefaults.check_constraints + for table_name, check_name, src, comment in result: + # only two cases for check_name and src: both null or both defined + if check_name is None and src is None: + check_constraints[(schema, table_name)] = default() + continue + # samples: + # "CHECK (((a > 1) AND (a < 5)))" + # "CHECK (((a = 1) OR ((a > 2) AND (a < 5))))" + # "CHECK (((a > 1) AND (a < 5))) NOT VALID" + # "CHECK (some_boolean_function(a))" + # "CHECK (((a\n < 1)\n OR\n (a\n >= 5))\n)" + # "CHECK (a NOT NULL) NO INHERIT" + # "CHECK (a NOT NULL) NO INHERIT NOT VALID" + + m = re.match( + r"^CHECK *\((.+)\)( NO INHERIT)?( NOT VALID)?$", + src, + flags=re.DOTALL, + ) + if not m: + util.warn("Could not parse CHECK constraint text: %r" % src) + sqltext = "" + else: + sqltext = re.compile( + r"^[\s\n]*\((.+)\)[\s\n]*$", flags=re.DOTALL + ).sub(r"\1", m.group(1)) + entry = { + "name": check_name, + "sqltext": sqltext, + "comment": comment, + } + if m: + do = {} + if " NOT VALID" in m.groups(): + do["not_valid"] = True + if " NO INHERIT" in m.groups(): + do["no_inherit"] = True + if do: + entry["dialect_options"] = do + + check_constraints[(schema, table_name)].append(entry) + return check_constraints.items() + + def _pg_type_filter_schema(self, query, schema): + if schema is None: + query = query.where( + pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid), + # ignore pg_catalog schema + pg_catalog.pg_namespace.c.nspname != "pg_catalog", + ) + elif schema != "*": + query = query.where(pg_catalog.pg_namespace.c.nspname == schema) + return query + + @lru_cache() + def _enum_query(self, schema): + lbl_agg_sq = ( + select( + pg_catalog.pg_enum.c.enumtypid, + sql.func.array_agg( + aggregate_order_by( + # NOTE: cast since some postgresql derivatives may + # not support array_agg on the name type + pg_catalog.pg_enum.c.enumlabel.cast(TEXT), + pg_catalog.pg_enum.c.enumsortorder, + ) + ).label("labels"), + ) + .group_by(pg_catalog.pg_enum.c.enumtypid) + .subquery("lbl_agg") + ) + + query = ( + select( + pg_catalog.pg_type.c.typname.label("name"), + pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid).label( + "visible" + ), + pg_catalog.pg_namespace.c.nspname.label("schema"), + lbl_agg_sq.c.labels.label("labels"), + ) + .join( + pg_catalog.pg_namespace, + pg_catalog.pg_namespace.c.oid + == pg_catalog.pg_type.c.typnamespace, + ) + .outerjoin( + lbl_agg_sq, pg_catalog.pg_type.c.oid == lbl_agg_sq.c.enumtypid + ) + .where(pg_catalog.pg_type.c.typtype == "e") + .order_by( + pg_catalog.pg_namespace.c.nspname, pg_catalog.pg_type.c.typname + ) + ) + + return self._pg_type_filter_schema(query, schema) + + @reflection.cache + def _load_enums(self, connection, schema=None, **kw): + if not self.supports_native_enum: + return [] + + result = connection.execute(self._enum_query(schema)) + + enums = [] + for name, visible, schema, labels in result: + enums.append( + { + "name": name, + "schema": schema, + "visible": visible, + "labels": [] if labels is None else labels, + } + ) + return enums + + @lru_cache() + def _domain_query(self, schema): + con_sq = ( + select( + pg_catalog.pg_constraint.c.contypid, + sql.func.array_agg( + pg_catalog.pg_get_constraintdef( + pg_catalog.pg_constraint.c.oid, True + ) + ).label("condefs"), + sql.func.array_agg( + # NOTE: cast since some postgresql derivatives may + # not support array_agg on the name type + pg_catalog.pg_constraint.c.conname.cast(TEXT) + ).label("connames"), + ) + # The domain this constraint is on; zero if not a domain constraint + .where(pg_catalog.pg_constraint.c.contypid != 0) + .group_by(pg_catalog.pg_constraint.c.contypid) + .subquery("domain_constraints") + ) + + query = ( + select( + pg_catalog.pg_type.c.typname.label("name"), + pg_catalog.format_type( + pg_catalog.pg_type.c.typbasetype, + pg_catalog.pg_type.c.typtypmod, + ).label("attype"), + (~pg_catalog.pg_type.c.typnotnull).label("nullable"), + pg_catalog.pg_type.c.typdefault.label("default"), + pg_catalog.pg_type_is_visible(pg_catalog.pg_type.c.oid).label( + "visible" + ), + pg_catalog.pg_namespace.c.nspname.label("schema"), + con_sq.c.condefs, + con_sq.c.connames, + pg_catalog.pg_collation.c.collname, + ) + .join( + pg_catalog.pg_namespace, + pg_catalog.pg_namespace.c.oid + == pg_catalog.pg_type.c.typnamespace, + ) + .outerjoin( + pg_catalog.pg_collation, + pg_catalog.pg_type.c.typcollation + == pg_catalog.pg_collation.c.oid, + ) + .outerjoin( + con_sq, + pg_catalog.pg_type.c.oid == con_sq.c.contypid, + ) + .where(pg_catalog.pg_type.c.typtype == "d") + .order_by( + pg_catalog.pg_namespace.c.nspname, pg_catalog.pg_type.c.typname + ) + ) + return self._pg_type_filter_schema(query, schema) + + @reflection.cache + def _load_domains(self, connection, schema=None, **kw): + result = connection.execute(self._domain_query(schema)) + + domains: List[ReflectedDomain] = [] + for domain in result.mappings(): + # strip (30) from character varying(30) + attype = re.search(r"([^\(]+)", domain["attype"]).group(1) + constraints: List[ReflectedDomainConstraint] = [] + if domain["connames"]: + # When a domain has multiple CHECK constraints, they will + # be tested in alphabetical order by name. + sorted_constraints = sorted( + zip(domain["connames"], domain["condefs"]), + key=lambda t: t[0], + ) + for name, def_ in sorted_constraints: + # constraint is in the form "CHECK (expression)". + # remove "CHECK (" and the tailing ")". + check = def_[7:-1] + constraints.append({"name": name, "check": check}) + + domain_rec: ReflectedDomain = { + "name": domain["name"], + "schema": domain["schema"], + "visible": domain["visible"], + "type": attype, + "nullable": domain["nullable"], + "default": domain["default"], + "constraints": constraints, + "collation": domain["collname"], + } + domains.append(domain_rec) + + return domains + + def _set_backslash_escapes(self, connection): + # this method is provided as an override hook for descendant + # dialects (e.g. Redshift), so removing it may break them + std_string = connection.exec_driver_sql( + "show standard_conforming_strings" + ).scalar() + self._backslash_escapes = std_string == "off" diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/dml.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/dml.py new file mode 100644 index 0000000..4404ecd --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/dml.py @@ -0,0 +1,310 @@ +# dialects/postgresql/dml.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +from __future__ import annotations + +from typing import Any +from typing import Optional + +from . import ext +from .._typing import _OnConflictConstraintT +from .._typing import _OnConflictIndexElementsT +from .._typing import _OnConflictIndexWhereT +from .._typing import _OnConflictSetT +from .._typing import _OnConflictWhereT +from ... import util +from ...sql import coercions +from ...sql import roles +from ...sql import schema +from ...sql._typing import _DMLTableArgument +from ...sql.base import _exclusive_against +from ...sql.base import _generative +from ...sql.base import ColumnCollection +from ...sql.base import ReadOnlyColumnCollection +from ...sql.dml import Insert as StandardInsert +from ...sql.elements import ClauseElement +from ...sql.elements import KeyedColumnElement +from ...sql.expression import alias +from ...util.typing import Self + + +__all__ = ("Insert", "insert") + + +def insert(table: _DMLTableArgument) -> Insert: + """Construct a PostgreSQL-specific variant :class:`_postgresql.Insert` + construct. + + .. container:: inherited_member + + The :func:`sqlalchemy.dialects.postgresql.insert` function creates + a :class:`sqlalchemy.dialects.postgresql.Insert`. This class is based + on the dialect-agnostic :class:`_sql.Insert` construct which may + be constructed using the :func:`_sql.insert` function in + SQLAlchemy Core. + + The :class:`_postgresql.Insert` construct includes additional methods + :meth:`_postgresql.Insert.on_conflict_do_update`, + :meth:`_postgresql.Insert.on_conflict_do_nothing`. + + """ + return Insert(table) + + +class Insert(StandardInsert): + """PostgreSQL-specific implementation of INSERT. + + Adds methods for PG-specific syntaxes such as ON CONFLICT. + + The :class:`_postgresql.Insert` object is created using the + :func:`sqlalchemy.dialects.postgresql.insert` function. + + """ + + stringify_dialect = "postgresql" + inherit_cache = False + + @util.memoized_property + def excluded( + self, + ) -> ReadOnlyColumnCollection[str, KeyedColumnElement[Any]]: + """Provide the ``excluded`` namespace for an ON CONFLICT statement + + PG's ON CONFLICT clause allows reference to the row that would + be inserted, known as ``excluded``. This attribute provides + all columns in this row to be referenceable. + + .. tip:: The :attr:`_postgresql.Insert.excluded` attribute is an + instance of :class:`_expression.ColumnCollection`, which provides + an interface the same as that of the :attr:`_schema.Table.c` + collection described at :ref:`metadata_tables_and_columns`. + With this collection, ordinary names are accessible like attributes + (e.g. ``stmt.excluded.some_column``), but special names and + dictionary method names should be accessed using indexed access, + such as ``stmt.excluded["column name"]`` or + ``stmt.excluded["values"]``. See the docstring for + :class:`_expression.ColumnCollection` for further examples. + + .. seealso:: + + :ref:`postgresql_insert_on_conflict` - example of how + to use :attr:`_expression.Insert.excluded` + + """ + return alias(self.table, name="excluded").columns + + _on_conflict_exclusive = _exclusive_against( + "_post_values_clause", + msgs={ + "_post_values_clause": "This Insert construct already has " + "an ON CONFLICT clause established" + }, + ) + + @_generative + @_on_conflict_exclusive + def on_conflict_do_update( + self, + constraint: _OnConflictConstraintT = None, + index_elements: _OnConflictIndexElementsT = None, + index_where: _OnConflictIndexWhereT = None, + set_: _OnConflictSetT = None, + where: _OnConflictWhereT = None, + ) -> Self: + r""" + Specifies a DO UPDATE SET action for ON CONFLICT clause. + + Either the ``constraint`` or ``index_elements`` argument is + required, but only one of these can be specified. + + :param constraint: + The name of a unique or exclusion constraint on the table, + or the constraint object itself if it has a .name attribute. + + :param index_elements: + A sequence consisting of string column names, :class:`_schema.Column` + objects, or other column expression objects that will be used + to infer a target index. + + :param index_where: + Additional WHERE criterion that can be used to infer a + conditional target index. + + :param set\_: + A dictionary or other mapping object + where the keys are either names of columns in the target table, + or :class:`_schema.Column` objects or other ORM-mapped columns + matching that of the target table, and expressions or literals + as values, specifying the ``SET`` actions to take. + + .. versionadded:: 1.4 The + :paramref:`_postgresql.Insert.on_conflict_do_update.set_` + parameter supports :class:`_schema.Column` objects from the target + :class:`_schema.Table` as keys. + + .. warning:: This dictionary does **not** take into account + Python-specified default UPDATE values or generation functions, + e.g. those specified using :paramref:`_schema.Column.onupdate`. + These values will not be exercised for an ON CONFLICT style of + UPDATE, unless they are manually specified in the + :paramref:`.Insert.on_conflict_do_update.set_` dictionary. + + :param where: + Optional argument. If present, can be a literal SQL + string or an acceptable expression for a ``WHERE`` clause + that restricts the rows affected by ``DO UPDATE SET``. Rows + not meeting the ``WHERE`` condition will not be updated + (effectively a ``DO NOTHING`` for those rows). + + + .. seealso:: + + :ref:`postgresql_insert_on_conflict` + + """ + self._post_values_clause = OnConflictDoUpdate( + constraint, index_elements, index_where, set_, where + ) + return self + + @_generative + @_on_conflict_exclusive + def on_conflict_do_nothing( + self, + constraint: _OnConflictConstraintT = None, + index_elements: _OnConflictIndexElementsT = None, + index_where: _OnConflictIndexWhereT = None, + ) -> Self: + """ + Specifies a DO NOTHING action for ON CONFLICT clause. + + The ``constraint`` and ``index_elements`` arguments + are optional, but only one of these can be specified. + + :param constraint: + The name of a unique or exclusion constraint on the table, + or the constraint object itself if it has a .name attribute. + + :param index_elements: + A sequence consisting of string column names, :class:`_schema.Column` + objects, or other column expression objects that will be used + to infer a target index. + + :param index_where: + Additional WHERE criterion that can be used to infer a + conditional target index. + + .. seealso:: + + :ref:`postgresql_insert_on_conflict` + + """ + self._post_values_clause = OnConflictDoNothing( + constraint, index_elements, index_where + ) + return self + + +class OnConflictClause(ClauseElement): + stringify_dialect = "postgresql" + + constraint_target: Optional[str] + inferred_target_elements: _OnConflictIndexElementsT + inferred_target_whereclause: _OnConflictIndexWhereT + + def __init__( + self, + constraint: _OnConflictConstraintT = None, + index_elements: _OnConflictIndexElementsT = None, + index_where: _OnConflictIndexWhereT = None, + ): + if constraint is not None: + if not isinstance(constraint, str) and isinstance( + constraint, + (schema.Constraint, ext.ExcludeConstraint), + ): + constraint = getattr(constraint, "name") or constraint + + if constraint is not None: + if index_elements is not None: + raise ValueError( + "'constraint' and 'index_elements' are mutually exclusive" + ) + + if isinstance(constraint, str): + self.constraint_target = constraint + self.inferred_target_elements = None + self.inferred_target_whereclause = None + elif isinstance(constraint, schema.Index): + index_elements = constraint.expressions + index_where = constraint.dialect_options["postgresql"].get( + "where" + ) + elif isinstance(constraint, ext.ExcludeConstraint): + index_elements = constraint.columns + index_where = constraint.where + else: + index_elements = constraint.columns + index_where = constraint.dialect_options["postgresql"].get( + "where" + ) + + if index_elements is not None: + self.constraint_target = None + self.inferred_target_elements = index_elements + self.inferred_target_whereclause = index_where + elif constraint is None: + self.constraint_target = self.inferred_target_elements = ( + self.inferred_target_whereclause + ) = None + + +class OnConflictDoNothing(OnConflictClause): + __visit_name__ = "on_conflict_do_nothing" + + +class OnConflictDoUpdate(OnConflictClause): + __visit_name__ = "on_conflict_do_update" + + def __init__( + self, + constraint: _OnConflictConstraintT = None, + index_elements: _OnConflictIndexElementsT = None, + index_where: _OnConflictIndexWhereT = None, + set_: _OnConflictSetT = None, + where: _OnConflictWhereT = None, + ): + super().__init__( + constraint=constraint, + index_elements=index_elements, + index_where=index_where, + ) + + if ( + self.inferred_target_elements is None + and self.constraint_target is None + ): + raise ValueError( + "Either constraint or index_elements, " + "but not both, must be specified unless DO NOTHING" + ) + + if isinstance(set_, dict): + if not set_: + raise ValueError("set parameter dictionary must not be empty") + elif isinstance(set_, ColumnCollection): + set_ = dict(set_) + else: + raise ValueError( + "set parameter must be a non-empty dictionary " + "or a ColumnCollection such as the `.c.` collection " + "of a Table object" + ) + self.update_values_to_set = [ + (coercions.expect(roles.DMLColumnRole, key), value) + for key, value in set_.items() + ] + self.update_whereclause = where diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ext.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ext.py new file mode 100644 index 0000000..7fc0895 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ext.py @@ -0,0 +1,496 @@ +# dialects/postgresql/ext.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors +from __future__ import annotations + +from typing import Any +from typing import TYPE_CHECKING +from typing import TypeVar + +from . import types +from .array import ARRAY +from ...sql import coercions +from ...sql import elements +from ...sql import expression +from ...sql import functions +from ...sql import roles +from ...sql import schema +from ...sql.schema import ColumnCollectionConstraint +from ...sql.sqltypes import TEXT +from ...sql.visitors import InternalTraversal + +_T = TypeVar("_T", bound=Any) + +if TYPE_CHECKING: + from ...sql.visitors import _TraverseInternalsType + + +class aggregate_order_by(expression.ColumnElement): + """Represent a PostgreSQL aggregate order by expression. + + E.g.:: + + from sqlalchemy.dialects.postgresql import aggregate_order_by + expr = func.array_agg(aggregate_order_by(table.c.a, table.c.b.desc())) + stmt = select(expr) + + would represent the expression:: + + SELECT array_agg(a ORDER BY b DESC) FROM table; + + Similarly:: + + expr = func.string_agg( + table.c.a, + aggregate_order_by(literal_column("','"), table.c.a) + ) + stmt = select(expr) + + Would represent:: + + SELECT string_agg(a, ',' ORDER BY a) FROM table; + + .. versionchanged:: 1.2.13 - the ORDER BY argument may be multiple terms + + .. seealso:: + + :class:`_functions.array_agg` + + """ + + __visit_name__ = "aggregate_order_by" + + stringify_dialect = "postgresql" + _traverse_internals: _TraverseInternalsType = [ + ("target", InternalTraversal.dp_clauseelement), + ("type", InternalTraversal.dp_type), + ("order_by", InternalTraversal.dp_clauseelement), + ] + + def __init__(self, target, *order_by): + self.target = coercions.expect(roles.ExpressionElementRole, target) + self.type = self.target.type + + _lob = len(order_by) + if _lob == 0: + raise TypeError("at least one ORDER BY element is required") + elif _lob == 1: + self.order_by = coercions.expect( + roles.ExpressionElementRole, order_by[0] + ) + else: + self.order_by = elements.ClauseList( + *order_by, _literal_as_text_role=roles.ExpressionElementRole + ) + + def self_group(self, against=None): + return self + + def get_children(self, **kwargs): + return self.target, self.order_by + + def _copy_internals(self, clone=elements._clone, **kw): + self.target = clone(self.target, **kw) + self.order_by = clone(self.order_by, **kw) + + @property + def _from_objects(self): + return self.target._from_objects + self.order_by._from_objects + + +class ExcludeConstraint(ColumnCollectionConstraint): + """A table-level EXCLUDE constraint. + + Defines an EXCLUDE constraint as described in the `PostgreSQL + documentation`__. + + __ https://www.postgresql.org/docs/current/static/sql-createtable.html#SQL-CREATETABLE-EXCLUDE + + """ # noqa + + __visit_name__ = "exclude_constraint" + + where = None + inherit_cache = False + + create_drop_stringify_dialect = "postgresql" + + @elements._document_text_coercion( + "where", + ":class:`.ExcludeConstraint`", + ":paramref:`.ExcludeConstraint.where`", + ) + def __init__(self, *elements, **kw): + r""" + Create an :class:`.ExcludeConstraint` object. + + E.g.:: + + const = ExcludeConstraint( + (Column('period'), '&&'), + (Column('group'), '='), + where=(Column('group') != 'some group'), + ops={'group': 'my_operator_class'} + ) + + The constraint is normally embedded into the :class:`_schema.Table` + construct + directly, or added later using :meth:`.append_constraint`:: + + some_table = Table( + 'some_table', metadata, + Column('id', Integer, primary_key=True), + Column('period', TSRANGE()), + Column('group', String) + ) + + some_table.append_constraint( + ExcludeConstraint( + (some_table.c.period, '&&'), + (some_table.c.group, '='), + where=some_table.c.group != 'some group', + name='some_table_excl_const', + ops={'group': 'my_operator_class'} + ) + ) + + The exclude constraint defined in this example requires the + ``btree_gist`` extension, that can be created using the + command ``CREATE EXTENSION btree_gist;``. + + :param \*elements: + + A sequence of two tuples of the form ``(column, operator)`` where + "column" is either a :class:`_schema.Column` object, or a SQL + expression element (e.g. ``func.int8range(table.from, table.to)``) + or the name of a column as string, and "operator" is a string + containing the operator to use (e.g. `"&&"` or `"="`). + + In order to specify a column name when a :class:`_schema.Column` + object is not available, while ensuring + that any necessary quoting rules take effect, an ad-hoc + :class:`_schema.Column` or :func:`_expression.column` + object should be used. + The ``column`` may also be a string SQL expression when + passed as :func:`_expression.literal_column` or + :func:`_expression.text` + + :param name: + Optional, the in-database name of this constraint. + + :param deferrable: + Optional bool. If set, emit DEFERRABLE or NOT DEFERRABLE when + issuing DDL for this constraint. + + :param initially: + Optional string. If set, emit INITIALLY <value> when issuing DDL + for this constraint. + + :param using: + Optional string. If set, emit USING <index_method> when issuing DDL + for this constraint. Defaults to 'gist'. + + :param where: + Optional SQL expression construct or literal SQL string. + If set, emit WHERE <predicate> when issuing DDL + for this constraint. + + :param ops: + Optional dictionary. Used to define operator classes for the + elements; works the same way as that of the + :ref:`postgresql_ops <postgresql_operator_classes>` + parameter specified to the :class:`_schema.Index` construct. + + .. versionadded:: 1.3.21 + + .. seealso:: + + :ref:`postgresql_operator_classes` - general description of how + PostgreSQL operator classes are specified. + + """ + columns = [] + render_exprs = [] + self.operators = {} + + expressions, operators = zip(*elements) + + for (expr, column, strname, add_element), operator in zip( + coercions.expect_col_expression_collection( + roles.DDLConstraintColumnRole, expressions + ), + operators, + ): + if add_element is not None: + columns.append(add_element) + + name = column.name if column is not None else strname + + if name is not None: + # backwards compat + self.operators[name] = operator + + render_exprs.append((expr, name, operator)) + + self._render_exprs = render_exprs + + ColumnCollectionConstraint.__init__( + self, + *columns, + name=kw.get("name"), + deferrable=kw.get("deferrable"), + initially=kw.get("initially"), + ) + self.using = kw.get("using", "gist") + where = kw.get("where") + if where is not None: + self.where = coercions.expect(roles.StatementOptionRole, where) + + self.ops = kw.get("ops", {}) + + def _set_parent(self, table, **kw): + super()._set_parent(table) + + self._render_exprs = [ + ( + expr if not isinstance(expr, str) else table.c[expr], + name, + operator, + ) + for expr, name, operator in (self._render_exprs) + ] + + def _copy(self, target_table=None, **kw): + elements = [ + ( + schema._copy_expression(expr, self.parent, target_table), + operator, + ) + for expr, _, operator in self._render_exprs + ] + c = self.__class__( + *elements, + name=self.name, + deferrable=self.deferrable, + initially=self.initially, + where=self.where, + using=self.using, + ) + c.dispatch._update(self.dispatch) + return c + + +def array_agg(*arg, **kw): + """PostgreSQL-specific form of :class:`_functions.array_agg`, ensures + return type is :class:`_postgresql.ARRAY` and not + the plain :class:`_types.ARRAY`, unless an explicit ``type_`` + is passed. + + """ + kw["_default_array_type"] = ARRAY + return functions.func.array_agg(*arg, **kw) + + +class _regconfig_fn(functions.GenericFunction[_T]): + inherit_cache = True + + def __init__(self, *args, **kwargs): + args = list(args) + if len(args) > 1: + initial_arg = coercions.expect( + roles.ExpressionElementRole, + args.pop(0), + name=getattr(self, "name", None), + apply_propagate_attrs=self, + type_=types.REGCONFIG, + ) + initial_arg = [initial_arg] + else: + initial_arg = [] + + addtl_args = [ + coercions.expect( + roles.ExpressionElementRole, + c, + name=getattr(self, "name", None), + apply_propagate_attrs=self, + ) + for c in args + ] + super().__init__(*(initial_arg + addtl_args), **kwargs) + + +class to_tsvector(_regconfig_fn): + """The PostgreSQL ``to_tsvector`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_postgresql.TSVECTOR`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.to_tsvector` will be used automatically when invoking + ``sqlalchemy.func.to_tsvector()``, ensuring the correct argument and return + type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = types.TSVECTOR + + +class to_tsquery(_regconfig_fn): + """The PostgreSQL ``to_tsquery`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_postgresql.TSQUERY`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.to_tsquery` will be used automatically when invoking + ``sqlalchemy.func.to_tsquery()``, ensuring the correct argument and return + type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = types.TSQUERY + + +class plainto_tsquery(_regconfig_fn): + """The PostgreSQL ``plainto_tsquery`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_postgresql.TSQUERY`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.plainto_tsquery` will be used automatically when + invoking ``sqlalchemy.func.plainto_tsquery()``, ensuring the correct + argument and return type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = types.TSQUERY + + +class phraseto_tsquery(_regconfig_fn): + """The PostgreSQL ``phraseto_tsquery`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_postgresql.TSQUERY`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.phraseto_tsquery` will be used automatically when + invoking ``sqlalchemy.func.phraseto_tsquery()``, ensuring the correct + argument and return type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = types.TSQUERY + + +class websearch_to_tsquery(_regconfig_fn): + """The PostgreSQL ``websearch_to_tsquery`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_postgresql.TSQUERY`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.websearch_to_tsquery` will be used automatically when + invoking ``sqlalchemy.func.websearch_to_tsquery()``, ensuring the correct + argument and return type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = types.TSQUERY + + +class ts_headline(_regconfig_fn): + """The PostgreSQL ``ts_headline`` SQL function. + + This function applies automatic casting of the REGCONFIG argument + to use the :class:`_postgresql.REGCONFIG` datatype automatically, + and applies a return type of :class:`_types.TEXT`. + + Assuming the PostgreSQL dialect has been imported, either by invoking + ``from sqlalchemy.dialects import postgresql``, or by creating a PostgreSQL + engine using ``create_engine("postgresql...")``, + :class:`_postgresql.ts_headline` will be used automatically when invoking + ``sqlalchemy.func.ts_headline()``, ensuring the correct argument and return + type handlers are used at compile and execution time. + + .. versionadded:: 2.0.0rc1 + + """ + + inherit_cache = True + type = TEXT + + def __init__(self, *args, **kwargs): + args = list(args) + + # parse types according to + # https://www.postgresql.org/docs/current/textsearch-controls.html#TEXTSEARCH-HEADLINE + if len(args) < 2: + # invalid args; don't do anything + has_regconfig = False + elif ( + isinstance(args[1], elements.ColumnElement) + and args[1].type._type_affinity is types.TSQUERY + ): + # tsquery is second argument, no regconfig argument + has_regconfig = False + else: + has_regconfig = True + + if has_regconfig: + initial_arg = coercions.expect( + roles.ExpressionElementRole, + args.pop(0), + apply_propagate_attrs=self, + name=getattr(self, "name", None), + type_=types.REGCONFIG, + ) + initial_arg = [initial_arg] + else: + initial_arg = [] + + addtl_args = [ + coercions.expect( + roles.ExpressionElementRole, + c, + name=getattr(self, "name", None), + apply_propagate_attrs=self, + ) + for c in args + ] + super().__init__(*(initial_arg + addtl_args), **kwargs) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/hstore.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/hstore.py new file mode 100644 index 0000000..04c8cf1 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/hstore.py @@ -0,0 +1,397 @@ +# dialects/postgresql/hstore.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + + +import re + +from .array import ARRAY +from .operators import CONTAINED_BY +from .operators import CONTAINS +from .operators import GETITEM +from .operators import HAS_ALL +from .operators import HAS_ANY +from .operators import HAS_KEY +from ... import types as sqltypes +from ...sql import functions as sqlfunc + + +__all__ = ("HSTORE", "hstore") + + +class HSTORE(sqltypes.Indexable, sqltypes.Concatenable, sqltypes.TypeEngine): + """Represent the PostgreSQL HSTORE type. + + The :class:`.HSTORE` type stores dictionaries containing strings, e.g.:: + + data_table = Table('data_table', metadata, + Column('id', Integer, primary_key=True), + Column('data', HSTORE) + ) + + with engine.connect() as conn: + conn.execute( + data_table.insert(), + data = {"key1": "value1", "key2": "value2"} + ) + + :class:`.HSTORE` provides for a wide range of operations, including: + + * Index operations:: + + data_table.c.data['some key'] == 'some value' + + * Containment operations:: + + data_table.c.data.has_key('some key') + + data_table.c.data.has_all(['one', 'two', 'three']) + + * Concatenation:: + + data_table.c.data + {"k1": "v1"} + + For a full list of special methods see + :class:`.HSTORE.comparator_factory`. + + .. container:: topic + + **Detecting Changes in HSTORE columns when using the ORM** + + For usage with the SQLAlchemy ORM, it may be desirable to combine the + usage of :class:`.HSTORE` with :class:`.MutableDict` dictionary now + part of the :mod:`sqlalchemy.ext.mutable` extension. This extension + will allow "in-place" changes to the dictionary, e.g. addition of new + keys or replacement/removal of existing keys to/from the current + dictionary, to produce events which will be detected by the unit of + work:: + + from sqlalchemy.ext.mutable import MutableDict + + class MyClass(Base): + __tablename__ = 'data_table' + + id = Column(Integer, primary_key=True) + data = Column(MutableDict.as_mutable(HSTORE)) + + my_object = session.query(MyClass).one() + + # in-place mutation, requires Mutable extension + # in order for the ORM to detect + my_object.data['some_key'] = 'some value' + + session.commit() + + When the :mod:`sqlalchemy.ext.mutable` extension is not used, the ORM + will not be alerted to any changes to the contents of an existing + dictionary, unless that dictionary value is re-assigned to the + HSTORE-attribute itself, thus generating a change event. + + .. seealso:: + + :class:`.hstore` - render the PostgreSQL ``hstore()`` function. + + + """ + + __visit_name__ = "HSTORE" + hashable = False + text_type = sqltypes.Text() + + def __init__(self, text_type=None): + """Construct a new :class:`.HSTORE`. + + :param text_type: the type that should be used for indexed values. + Defaults to :class:`_types.Text`. + + """ + if text_type is not None: + self.text_type = text_type + + class Comparator( + sqltypes.Indexable.Comparator, sqltypes.Concatenable.Comparator + ): + """Define comparison operations for :class:`.HSTORE`.""" + + def has_key(self, other): + """Boolean expression. Test for presence of a key. Note that the + key may be a SQLA expression. + """ + return self.operate(HAS_KEY, other, result_type=sqltypes.Boolean) + + def has_all(self, other): + """Boolean expression. Test for presence of all keys in jsonb""" + return self.operate(HAS_ALL, other, result_type=sqltypes.Boolean) + + def has_any(self, other): + """Boolean expression. Test for presence of any key in jsonb""" + return self.operate(HAS_ANY, other, result_type=sqltypes.Boolean) + + def contains(self, other, **kwargs): + """Boolean expression. Test if keys (or array) are a superset + of/contained the keys of the argument jsonb expression. + + kwargs may be ignored by this operator but are required for API + conformance. + """ + return self.operate(CONTAINS, other, result_type=sqltypes.Boolean) + + def contained_by(self, other): + """Boolean expression. Test if keys are a proper subset of the + keys of the argument jsonb expression. + """ + return self.operate( + CONTAINED_BY, other, result_type=sqltypes.Boolean + ) + + def _setup_getitem(self, index): + return GETITEM, index, self.type.text_type + + def defined(self, key): + """Boolean expression. Test for presence of a non-NULL value for + the key. Note that the key may be a SQLA expression. + """ + return _HStoreDefinedFunction(self.expr, key) + + def delete(self, key): + """HStore expression. Returns the contents of this hstore with the + given key deleted. Note that the key may be a SQLA expression. + """ + if isinstance(key, dict): + key = _serialize_hstore(key) + return _HStoreDeleteFunction(self.expr, key) + + def slice(self, array): + """HStore expression. Returns a subset of an hstore defined by + array of keys. + """ + return _HStoreSliceFunction(self.expr, array) + + def keys(self): + """Text array expression. Returns array of keys.""" + return _HStoreKeysFunction(self.expr) + + def vals(self): + """Text array expression. Returns array of values.""" + return _HStoreValsFunction(self.expr) + + def array(self): + """Text array expression. Returns array of alternating keys and + values. + """ + return _HStoreArrayFunction(self.expr) + + def matrix(self): + """Text array expression. Returns array of [key, value] pairs.""" + return _HStoreMatrixFunction(self.expr) + + comparator_factory = Comparator + + def bind_processor(self, dialect): + def process(value): + if isinstance(value, dict): + return _serialize_hstore(value) + else: + return value + + return process + + def result_processor(self, dialect, coltype): + def process(value): + if value is not None: + return _parse_hstore(value) + else: + return value + + return process + + +class hstore(sqlfunc.GenericFunction): + """Construct an hstore value within a SQL expression using the + PostgreSQL ``hstore()`` function. + + The :class:`.hstore` function accepts one or two arguments as described + in the PostgreSQL documentation. + + E.g.:: + + from sqlalchemy.dialects.postgresql import array, hstore + + select(hstore('key1', 'value1')) + + select( + hstore( + array(['key1', 'key2', 'key3']), + array(['value1', 'value2', 'value3']) + ) + ) + + .. seealso:: + + :class:`.HSTORE` - the PostgreSQL ``HSTORE`` datatype. + + """ + + type = HSTORE + name = "hstore" + inherit_cache = True + + +class _HStoreDefinedFunction(sqlfunc.GenericFunction): + type = sqltypes.Boolean + name = "defined" + inherit_cache = True + + +class _HStoreDeleteFunction(sqlfunc.GenericFunction): + type = HSTORE + name = "delete" + inherit_cache = True + + +class _HStoreSliceFunction(sqlfunc.GenericFunction): + type = HSTORE + name = "slice" + inherit_cache = True + + +class _HStoreKeysFunction(sqlfunc.GenericFunction): + type = ARRAY(sqltypes.Text) + name = "akeys" + inherit_cache = True + + +class _HStoreValsFunction(sqlfunc.GenericFunction): + type = ARRAY(sqltypes.Text) + name = "avals" + inherit_cache = True + + +class _HStoreArrayFunction(sqlfunc.GenericFunction): + type = ARRAY(sqltypes.Text) + name = "hstore_to_array" + inherit_cache = True + + +class _HStoreMatrixFunction(sqlfunc.GenericFunction): + type = ARRAY(sqltypes.Text) + name = "hstore_to_matrix" + inherit_cache = True + + +# +# parsing. note that none of this is used with the psycopg2 backend, +# which provides its own native extensions. +# + +# My best guess at the parsing rules of hstore literals, since no formal +# grammar is given. This is mostly reverse engineered from PG's input parser +# behavior. +HSTORE_PAIR_RE = re.compile( + r""" +( + "(?P<key> (\\ . | [^"])* )" # Quoted key +) +[ ]* => [ ]* # Pair operator, optional adjoining whitespace +( + (?P<value_null> NULL ) # NULL value + | "(?P<value> (\\ . | [^"])* )" # Quoted value +) +""", + re.VERBOSE, +) + +HSTORE_DELIMITER_RE = re.compile( + r""" +[ ]* , [ ]* +""", + re.VERBOSE, +) + + +def _parse_error(hstore_str, pos): + """format an unmarshalling error.""" + + ctx = 20 + hslen = len(hstore_str) + + parsed_tail = hstore_str[max(pos - ctx - 1, 0) : min(pos, hslen)] + residual = hstore_str[min(pos, hslen) : min(pos + ctx + 1, hslen)] + + if len(parsed_tail) > ctx: + parsed_tail = "[...]" + parsed_tail[1:] + if len(residual) > ctx: + residual = residual[:-1] + "[...]" + + return "After %r, could not parse residual at position %d: %r" % ( + parsed_tail, + pos, + residual, + ) + + +def _parse_hstore(hstore_str): + """Parse an hstore from its literal string representation. + + Attempts to approximate PG's hstore input parsing rules as closely as + possible. Although currently this is not strictly necessary, since the + current implementation of hstore's output syntax is stricter than what it + accepts as input, the documentation makes no guarantees that will always + be the case. + + + + """ + result = {} + pos = 0 + pair_match = HSTORE_PAIR_RE.match(hstore_str) + + while pair_match is not None: + key = pair_match.group("key").replace(r"\"", '"').replace("\\\\", "\\") + if pair_match.group("value_null"): + value = None + else: + value = ( + pair_match.group("value") + .replace(r"\"", '"') + .replace("\\\\", "\\") + ) + result[key] = value + + pos += pair_match.end() + + delim_match = HSTORE_DELIMITER_RE.match(hstore_str[pos:]) + if delim_match is not None: + pos += delim_match.end() + + pair_match = HSTORE_PAIR_RE.match(hstore_str[pos:]) + + if pos != len(hstore_str): + raise ValueError(_parse_error(hstore_str, pos)) + + return result + + +def _serialize_hstore(val): + """Serialize a dictionary into an hstore literal. Keys and values must + both be strings (except None for values). + + """ + + def esc(s, position): + if position == "value" and s is None: + return "NULL" + elif isinstance(s, str): + return '"%s"' % s.replace("\\", "\\\\").replace('"', r"\"") + else: + raise ValueError( + "%r in %s position is not a string." % (s, position) + ) + + return ", ".join( + "%s=>%s" % (esc(k, "key"), esc(v, "value")) for k, v in val.items() + ) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/json.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/json.py new file mode 100644 index 0000000..3790fa3 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/json.py @@ -0,0 +1,325 @@ +# dialects/postgresql/json.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + + +from .array import ARRAY +from .array import array as _pg_array +from .operators import ASTEXT +from .operators import CONTAINED_BY +from .operators import CONTAINS +from .operators import DELETE_PATH +from .operators import HAS_ALL +from .operators import HAS_ANY +from .operators import HAS_KEY +from .operators import JSONPATH_ASTEXT +from .operators import PATH_EXISTS +from .operators import PATH_MATCH +from ... import types as sqltypes +from ...sql import cast + +__all__ = ("JSON", "JSONB") + + +class JSONPathType(sqltypes.JSON.JSONPathType): + def _processor(self, dialect, super_proc): + def process(value): + if isinstance(value, str): + # If it's already a string assume that it's in json path + # format. This allows using cast with json paths literals + return value + elif value: + # If it's already a string assume that it's in json path + # format. This allows using cast with json paths literals + value = "{%s}" % (", ".join(map(str, value))) + else: + value = "{}" + if super_proc: + value = super_proc(value) + return value + + return process + + def bind_processor(self, dialect): + return self._processor(dialect, self.string_bind_processor(dialect)) + + def literal_processor(self, dialect): + return self._processor(dialect, self.string_literal_processor(dialect)) + + +class JSONPATH(JSONPathType): + """JSON Path Type. + + This is usually required to cast literal values to json path when using + json search like function, such as ``jsonb_path_query_array`` or + ``jsonb_path_exists``:: + + stmt = sa.select( + sa.func.jsonb_path_query_array( + table.c.jsonb_col, cast("$.address.id", JSONPATH) + ) + ) + + """ + + __visit_name__ = "JSONPATH" + + +class JSON(sqltypes.JSON): + """Represent the PostgreSQL JSON type. + + :class:`_postgresql.JSON` is used automatically whenever the base + :class:`_types.JSON` datatype is used against a PostgreSQL backend, + however base :class:`_types.JSON` datatype does not provide Python + accessors for PostgreSQL-specific comparison methods such as + :meth:`_postgresql.JSON.Comparator.astext`; additionally, to use + PostgreSQL ``JSONB``, the :class:`_postgresql.JSONB` datatype should + be used explicitly. + + .. seealso:: + + :class:`_types.JSON` - main documentation for the generic + cross-platform JSON datatype. + + The operators provided by the PostgreSQL version of :class:`_types.JSON` + include: + + * Index operations (the ``->`` operator):: + + data_table.c.data['some key'] + + data_table.c.data[5] + + + * Index operations returning text (the ``->>`` operator):: + + data_table.c.data['some key'].astext == 'some value' + + Note that equivalent functionality is available via the + :attr:`.JSON.Comparator.as_string` accessor. + + * Index operations with CAST + (equivalent to ``CAST(col ->> ['some key'] AS <type>)``):: + + data_table.c.data['some key'].astext.cast(Integer) == 5 + + Note that equivalent functionality is available via the + :attr:`.JSON.Comparator.as_integer` and similar accessors. + + * Path index operations (the ``#>`` operator):: + + data_table.c.data[('key_1', 'key_2', 5, ..., 'key_n')] + + * Path index operations returning text (the ``#>>`` operator):: + + data_table.c.data[('key_1', 'key_2', 5, ..., 'key_n')].astext == 'some value' + + Index operations return an expression object whose type defaults to + :class:`_types.JSON` by default, + so that further JSON-oriented instructions + may be called upon the result type. + + Custom serializers and deserializers are specified at the dialect level, + that is using :func:`_sa.create_engine`. The reason for this is that when + using psycopg2, the DBAPI only allows serializers at the per-cursor + or per-connection level. E.g.:: + + engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test", + json_serializer=my_serialize_fn, + json_deserializer=my_deserialize_fn + ) + + When using the psycopg2 dialect, the json_deserializer is registered + against the database using ``psycopg2.extras.register_default_json``. + + .. seealso:: + + :class:`_types.JSON` - Core level JSON type + + :class:`_postgresql.JSONB` + + """ # noqa + + astext_type = sqltypes.Text() + + def __init__(self, none_as_null=False, astext_type=None): + """Construct a :class:`_types.JSON` type. + + :param none_as_null: if True, persist the value ``None`` as a + SQL NULL value, not the JSON encoding of ``null``. Note that + when this flag is False, the :func:`.null` construct can still + be used to persist a NULL value:: + + from sqlalchemy import null + conn.execute(table.insert(), {"data": null()}) + + .. seealso:: + + :attr:`_types.JSON.NULL` + + :param astext_type: the type to use for the + :attr:`.JSON.Comparator.astext` + accessor on indexed attributes. Defaults to :class:`_types.Text`. + + """ + super().__init__(none_as_null=none_as_null) + if astext_type is not None: + self.astext_type = astext_type + + class Comparator(sqltypes.JSON.Comparator): + """Define comparison operations for :class:`_types.JSON`.""" + + @property + def astext(self): + """On an indexed expression, use the "astext" (e.g. "->>") + conversion when rendered in SQL. + + E.g.:: + + select(data_table.c.data['some key'].astext) + + .. seealso:: + + :meth:`_expression.ColumnElement.cast` + + """ + if isinstance(self.expr.right.type, sqltypes.JSON.JSONPathType): + return self.expr.left.operate( + JSONPATH_ASTEXT, + self.expr.right, + result_type=self.type.astext_type, + ) + else: + return self.expr.left.operate( + ASTEXT, self.expr.right, result_type=self.type.astext_type + ) + + comparator_factory = Comparator + + +class JSONB(JSON): + """Represent the PostgreSQL JSONB type. + + The :class:`_postgresql.JSONB` type stores arbitrary JSONB format data, + e.g.:: + + data_table = Table('data_table', metadata, + Column('id', Integer, primary_key=True), + Column('data', JSONB) + ) + + with engine.connect() as conn: + conn.execute( + data_table.insert(), + data = {"key1": "value1", "key2": "value2"} + ) + + The :class:`_postgresql.JSONB` type includes all operations provided by + :class:`_types.JSON`, including the same behaviors for indexing + operations. + It also adds additional operators specific to JSONB, including + :meth:`.JSONB.Comparator.has_key`, :meth:`.JSONB.Comparator.has_all`, + :meth:`.JSONB.Comparator.has_any`, :meth:`.JSONB.Comparator.contains`, + :meth:`.JSONB.Comparator.contained_by`, + :meth:`.JSONB.Comparator.delete_path`, + :meth:`.JSONB.Comparator.path_exists` and + :meth:`.JSONB.Comparator.path_match`. + + Like the :class:`_types.JSON` type, the :class:`_postgresql.JSONB` + type does not detect + in-place changes when used with the ORM, unless the + :mod:`sqlalchemy.ext.mutable` extension is used. + + Custom serializers and deserializers + are shared with the :class:`_types.JSON` class, + using the ``json_serializer`` + and ``json_deserializer`` keyword arguments. These must be specified + at the dialect level using :func:`_sa.create_engine`. When using + psycopg2, the serializers are associated with the jsonb type using + ``psycopg2.extras.register_default_jsonb`` on a per-connection basis, + in the same way that ``psycopg2.extras.register_default_json`` is used + to register these handlers with the json type. + + .. seealso:: + + :class:`_types.JSON` + + """ + + __visit_name__ = "JSONB" + + class Comparator(JSON.Comparator): + """Define comparison operations for :class:`_types.JSON`.""" + + def has_key(self, other): + """Boolean expression. Test for presence of a key. Note that the + key may be a SQLA expression. + """ + return self.operate(HAS_KEY, other, result_type=sqltypes.Boolean) + + def has_all(self, other): + """Boolean expression. Test for presence of all keys in jsonb""" + return self.operate(HAS_ALL, other, result_type=sqltypes.Boolean) + + def has_any(self, other): + """Boolean expression. Test for presence of any key in jsonb""" + return self.operate(HAS_ANY, other, result_type=sqltypes.Boolean) + + def contains(self, other, **kwargs): + """Boolean expression. Test if keys (or array) are a superset + of/contained the keys of the argument jsonb expression. + + kwargs may be ignored by this operator but are required for API + conformance. + """ + return self.operate(CONTAINS, other, result_type=sqltypes.Boolean) + + def contained_by(self, other): + """Boolean expression. Test if keys are a proper subset of the + keys of the argument jsonb expression. + """ + return self.operate( + CONTAINED_BY, other, result_type=sqltypes.Boolean + ) + + def delete_path(self, array): + """JSONB expression. Deletes field or array element specified in + the argument array. + + The input may be a list of strings that will be coerced to an + ``ARRAY`` or an instance of :meth:`_postgres.array`. + + .. versionadded:: 2.0 + """ + if not isinstance(array, _pg_array): + array = _pg_array(array) + right_side = cast(array, ARRAY(sqltypes.TEXT)) + return self.operate(DELETE_PATH, right_side, result_type=JSONB) + + def path_exists(self, other): + """Boolean expression. Test for presence of item given by the + argument JSONPath expression. + + .. versionadded:: 2.0 + """ + return self.operate( + PATH_EXISTS, other, result_type=sqltypes.Boolean + ) + + def path_match(self, other): + """Boolean expression. Test if JSONPath predicate given by the + argument JSONPath expression matches. + + Only the first item of the result is taken into account. + + .. versionadded:: 2.0 + """ + return self.operate( + PATH_MATCH, other, result_type=sqltypes.Boolean + ) + + comparator_factory = Comparator diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/named_types.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/named_types.py new file mode 100644 index 0000000..16e5c86 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/named_types.py @@ -0,0 +1,509 @@ +# dialects/postgresql/named_types.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors +from __future__ import annotations + +from typing import Any +from typing import Optional +from typing import Type +from typing import TYPE_CHECKING +from typing import Union + +from ... import schema +from ... import util +from ...sql import coercions +from ...sql import elements +from ...sql import roles +from ...sql import sqltypes +from ...sql import type_api +from ...sql.base import _NoArg +from ...sql.ddl import InvokeCreateDDLBase +from ...sql.ddl import InvokeDropDDLBase + +if TYPE_CHECKING: + from ...sql._typing import _TypeEngineArgument + + +class NamedType(sqltypes.TypeEngine): + """Base for named types.""" + + __abstract__ = True + DDLGenerator: Type[NamedTypeGenerator] + DDLDropper: Type[NamedTypeDropper] + create_type: bool + + def create(self, bind, checkfirst=True, **kw): + """Emit ``CREATE`` DDL for this type. + + :param bind: a connectable :class:`_engine.Engine`, + :class:`_engine.Connection`, or similar object to emit + SQL. + :param checkfirst: if ``True``, a query against + the PG catalog will be first performed to see + if the type does not exist already before + creating. + + """ + bind._run_ddl_visitor(self.DDLGenerator, self, checkfirst=checkfirst) + + def drop(self, bind, checkfirst=True, **kw): + """Emit ``DROP`` DDL for this type. + + :param bind: a connectable :class:`_engine.Engine`, + :class:`_engine.Connection`, or similar object to emit + SQL. + :param checkfirst: if ``True``, a query against + the PG catalog will be first performed to see + if the type actually exists before dropping. + + """ + bind._run_ddl_visitor(self.DDLDropper, self, checkfirst=checkfirst) + + def _check_for_name_in_memos(self, checkfirst, kw): + """Look in the 'ddl runner' for 'memos', then + note our name in that collection. + + This to ensure a particular named type is operated + upon only once within any kind of create/drop + sequence without relying upon "checkfirst". + + """ + if not self.create_type: + return True + if "_ddl_runner" in kw: + ddl_runner = kw["_ddl_runner"] + type_name = f"pg_{self.__visit_name__}" + if type_name in ddl_runner.memo: + existing = ddl_runner.memo[type_name] + else: + existing = ddl_runner.memo[type_name] = set() + present = (self.schema, self.name) in existing + existing.add((self.schema, self.name)) + return present + else: + return False + + def _on_table_create(self, target, bind, checkfirst=False, **kw): + if ( + checkfirst + or ( + not self.metadata + and not kw.get("_is_metadata_operation", False) + ) + ) and not self._check_for_name_in_memos(checkfirst, kw): + self.create(bind=bind, checkfirst=checkfirst) + + def _on_table_drop(self, target, bind, checkfirst=False, **kw): + if ( + not self.metadata + and not kw.get("_is_metadata_operation", False) + and not self._check_for_name_in_memos(checkfirst, kw) + ): + self.drop(bind=bind, checkfirst=checkfirst) + + def _on_metadata_create(self, target, bind, checkfirst=False, **kw): + if not self._check_for_name_in_memos(checkfirst, kw): + self.create(bind=bind, checkfirst=checkfirst) + + def _on_metadata_drop(self, target, bind, checkfirst=False, **kw): + if not self._check_for_name_in_memos(checkfirst, kw): + self.drop(bind=bind, checkfirst=checkfirst) + + +class NamedTypeGenerator(InvokeCreateDDLBase): + def __init__(self, dialect, connection, checkfirst=False, **kwargs): + super().__init__(connection, **kwargs) + self.checkfirst = checkfirst + + def _can_create_type(self, type_): + if not self.checkfirst: + return True + + effective_schema = self.connection.schema_for_object(type_) + return not self.connection.dialect.has_type( + self.connection, type_.name, schema=effective_schema + ) + + +class NamedTypeDropper(InvokeDropDDLBase): + def __init__(self, dialect, connection, checkfirst=False, **kwargs): + super().__init__(connection, **kwargs) + self.checkfirst = checkfirst + + def _can_drop_type(self, type_): + if not self.checkfirst: + return True + + effective_schema = self.connection.schema_for_object(type_) + return self.connection.dialect.has_type( + self.connection, type_.name, schema=effective_schema + ) + + +class EnumGenerator(NamedTypeGenerator): + def visit_enum(self, enum): + if not self._can_create_type(enum): + return + + with self.with_ddl_events(enum): + self.connection.execute(CreateEnumType(enum)) + + +class EnumDropper(NamedTypeDropper): + def visit_enum(self, enum): + if not self._can_drop_type(enum): + return + + with self.with_ddl_events(enum): + self.connection.execute(DropEnumType(enum)) + + +class ENUM(NamedType, type_api.NativeForEmulated, sqltypes.Enum): + """PostgreSQL ENUM type. + + This is a subclass of :class:`_types.Enum` which includes + support for PG's ``CREATE TYPE`` and ``DROP TYPE``. + + When the builtin type :class:`_types.Enum` is used and the + :paramref:`.Enum.native_enum` flag is left at its default of + True, the PostgreSQL backend will use a :class:`_postgresql.ENUM` + type as the implementation, so the special create/drop rules + will be used. + + The create/drop behavior of ENUM is necessarily intricate, due to the + awkward relationship the ENUM type has in relationship to the + parent table, in that it may be "owned" by just a single table, or + may be shared among many tables. + + When using :class:`_types.Enum` or :class:`_postgresql.ENUM` + in an "inline" fashion, the ``CREATE TYPE`` and ``DROP TYPE`` is emitted + corresponding to when the :meth:`_schema.Table.create` and + :meth:`_schema.Table.drop` + methods are called:: + + table = Table('sometable', metadata, + Column('some_enum', ENUM('a', 'b', 'c', name='myenum')) + ) + + table.create(engine) # will emit CREATE ENUM and CREATE TABLE + table.drop(engine) # will emit DROP TABLE and DROP ENUM + + To use a common enumerated type between multiple tables, the best + practice is to declare the :class:`_types.Enum` or + :class:`_postgresql.ENUM` independently, and associate it with the + :class:`_schema.MetaData` object itself:: + + my_enum = ENUM('a', 'b', 'c', name='myenum', metadata=metadata) + + t1 = Table('sometable_one', metadata, + Column('some_enum', myenum) + ) + + t2 = Table('sometable_two', metadata, + Column('some_enum', myenum) + ) + + When this pattern is used, care must still be taken at the level + of individual table creates. Emitting CREATE TABLE without also + specifying ``checkfirst=True`` will still cause issues:: + + t1.create(engine) # will fail: no such type 'myenum' + + If we specify ``checkfirst=True``, the individual table-level create + operation will check for the ``ENUM`` and create if not exists:: + + # will check if enum exists, and emit CREATE TYPE if not + t1.create(engine, checkfirst=True) + + When using a metadata-level ENUM type, the type will always be created + and dropped if either the metadata-wide create/drop is called:: + + metadata.create_all(engine) # will emit CREATE TYPE + metadata.drop_all(engine) # will emit DROP TYPE + + The type can also be created and dropped directly:: + + my_enum.create(engine) + my_enum.drop(engine) + + """ + + native_enum = True + DDLGenerator = EnumGenerator + DDLDropper = EnumDropper + + def __init__( + self, + *enums, + name: Union[str, _NoArg, None] = _NoArg.NO_ARG, + create_type: bool = True, + **kw, + ): + """Construct an :class:`_postgresql.ENUM`. + + Arguments are the same as that of + :class:`_types.Enum`, but also including + the following parameters. + + :param create_type: Defaults to True. + Indicates that ``CREATE TYPE`` should be + emitted, after optionally checking for the + presence of the type, when the parent + table is being created; and additionally + that ``DROP TYPE`` is called when the table + is dropped. When ``False``, no check + will be performed and no ``CREATE TYPE`` + or ``DROP TYPE`` is emitted, unless + :meth:`~.postgresql.ENUM.create` + or :meth:`~.postgresql.ENUM.drop` + are called directly. + Setting to ``False`` is helpful + when invoking a creation scheme to a SQL file + without access to the actual database - + the :meth:`~.postgresql.ENUM.create` and + :meth:`~.postgresql.ENUM.drop` methods can + be used to emit SQL to a target bind. + + """ + native_enum = kw.pop("native_enum", None) + if native_enum is False: + util.warn( + "the native_enum flag does not apply to the " + "sqlalchemy.dialects.postgresql.ENUM datatype; this type " + "always refers to ENUM. Use sqlalchemy.types.Enum for " + "non-native enum." + ) + self.create_type = create_type + if name is not _NoArg.NO_ARG: + kw["name"] = name + super().__init__(*enums, **kw) + + def coerce_compared_value(self, op, value): + super_coerced_type = super().coerce_compared_value(op, value) + if ( + super_coerced_type._type_affinity + is type_api.STRINGTYPE._type_affinity + ): + return self + else: + return super_coerced_type + + @classmethod + def __test_init__(cls): + return cls(name="name") + + @classmethod + def adapt_emulated_to_native(cls, impl, **kw): + """Produce a PostgreSQL native :class:`_postgresql.ENUM` from plain + :class:`.Enum`. + + """ + kw.setdefault("validate_strings", impl.validate_strings) + kw.setdefault("name", impl.name) + kw.setdefault("schema", impl.schema) + kw.setdefault("inherit_schema", impl.inherit_schema) + kw.setdefault("metadata", impl.metadata) + kw.setdefault("_create_events", False) + kw.setdefault("values_callable", impl.values_callable) + kw.setdefault("omit_aliases", impl._omit_aliases) + kw.setdefault("_adapted_from", impl) + if type_api._is_native_for_emulated(impl.__class__): + kw.setdefault("create_type", impl.create_type) + + return cls(**kw) + + def create(self, bind=None, checkfirst=True): + """Emit ``CREATE TYPE`` for this + :class:`_postgresql.ENUM`. + + If the underlying dialect does not support + PostgreSQL CREATE TYPE, no action is taken. + + :param bind: a connectable :class:`_engine.Engine`, + :class:`_engine.Connection`, or similar object to emit + SQL. + :param checkfirst: if ``True``, a query against + the PG catalog will be first performed to see + if the type does not exist already before + creating. + + """ + if not bind.dialect.supports_native_enum: + return + + super().create(bind, checkfirst=checkfirst) + + def drop(self, bind=None, checkfirst=True): + """Emit ``DROP TYPE`` for this + :class:`_postgresql.ENUM`. + + If the underlying dialect does not support + PostgreSQL DROP TYPE, no action is taken. + + :param bind: a connectable :class:`_engine.Engine`, + :class:`_engine.Connection`, or similar object to emit + SQL. + :param checkfirst: if ``True``, a query against + the PG catalog will be first performed to see + if the type actually exists before dropping. + + """ + if not bind.dialect.supports_native_enum: + return + + super().drop(bind, checkfirst=checkfirst) + + def get_dbapi_type(self, dbapi): + """dont return dbapi.STRING for ENUM in PostgreSQL, since that's + a different type""" + + return None + + +class DomainGenerator(NamedTypeGenerator): + def visit_DOMAIN(self, domain): + if not self._can_create_type(domain): + return + with self.with_ddl_events(domain): + self.connection.execute(CreateDomainType(domain)) + + +class DomainDropper(NamedTypeDropper): + def visit_DOMAIN(self, domain): + if not self._can_drop_type(domain): + return + + with self.with_ddl_events(domain): + self.connection.execute(DropDomainType(domain)) + + +class DOMAIN(NamedType, sqltypes.SchemaType): + r"""Represent the DOMAIN PostgreSQL type. + + A domain is essentially a data type with optional constraints + that restrict the allowed set of values. E.g.:: + + PositiveInt = DOMAIN( + "pos_int", Integer, check="VALUE > 0", not_null=True + ) + + UsPostalCode = DOMAIN( + "us_postal_code", + Text, + check="VALUE ~ '^\d{5}$' OR VALUE ~ '^\d{5}-\d{4}$'" + ) + + See the `PostgreSQL documentation`__ for additional details + + __ https://www.postgresql.org/docs/current/sql-createdomain.html + + .. versionadded:: 2.0 + + """ + + DDLGenerator = DomainGenerator + DDLDropper = DomainDropper + + __visit_name__ = "DOMAIN" + + def __init__( + self, + name: str, + data_type: _TypeEngineArgument[Any], + *, + collation: Optional[str] = None, + default: Union[elements.TextClause, str, None] = None, + constraint_name: Optional[str] = None, + not_null: Optional[bool] = None, + check: Union[elements.TextClause, str, None] = None, + create_type: bool = True, + **kw: Any, + ): + """ + Construct a DOMAIN. + + :param name: the name of the domain + :param data_type: The underlying data type of the domain. + This can include array specifiers. + :param collation: An optional collation for the domain. + If no collation is specified, the underlying data type's default + collation is used. The underlying type must be collatable if + ``collation`` is specified. + :param default: The DEFAULT clause specifies a default value for + columns of the domain data type. The default should be a string + or a :func:`_expression.text` value. + If no default value is specified, then the default value is + the null value. + :param constraint_name: An optional name for a constraint. + If not specified, the backend generates a name. + :param not_null: Values of this domain are prevented from being null. + By default domain are allowed to be null. If not specified + no nullability clause will be emitted. + :param check: CHECK clause specify integrity constraint or test + which values of the domain must satisfy. A constraint must be + an expression producing a Boolean result that can use the key + word VALUE to refer to the value being tested. + Differently from PostgreSQL, only a single check clause is + currently allowed in SQLAlchemy. + :param schema: optional schema name + :param metadata: optional :class:`_schema.MetaData` object which + this :class:`_postgresql.DOMAIN` will be directly associated + :param create_type: Defaults to True. + Indicates that ``CREATE TYPE`` should be emitted, after optionally + checking for the presence of the type, when the parent table is + being created; and additionally that ``DROP TYPE`` is called + when the table is dropped. + + """ + self.data_type = type_api.to_instance(data_type) + self.default = default + self.collation = collation + self.constraint_name = constraint_name + self.not_null = bool(not_null) + if check is not None: + check = coercions.expect(roles.DDLExpressionRole, check) + self.check = check + self.create_type = create_type + super().__init__(name=name, **kw) + + @classmethod + def __test_init__(cls): + return cls("name", sqltypes.Integer) + + def adapt(self, impl, **kw): + if self.default: + kw["default"] = self.default + if self.constraint_name is not None: + kw["constraint_name"] = self.constraint_name + if self.not_null: + kw["not_null"] = self.not_null + if self.check is not None: + kw["check"] = str(self.check) + if self.create_type: + kw["create_type"] = self.create_type + + return super().adapt(impl, **kw) + + +class CreateEnumType(schema._CreateDropBase): + __visit_name__ = "create_enum_type" + + +class DropEnumType(schema._CreateDropBase): + __visit_name__ = "drop_enum_type" + + +class CreateDomainType(schema._CreateDropBase): + """Represent a CREATE DOMAIN statement.""" + + __visit_name__ = "create_domain_type" + + +class DropDomainType(schema._CreateDropBase): + """Represent a DROP DOMAIN statement.""" + + __visit_name__ = "drop_domain_type" diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/operators.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/operators.py new file mode 100644 index 0000000..53e175f --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/operators.py @@ -0,0 +1,129 @@ +# dialects/postgresql/operators.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors +from ...sql import operators + + +_getitem_precedence = operators._PRECEDENCE[operators.json_getitem_op] +_eq_precedence = operators._PRECEDENCE[operators.eq] + +# JSON + JSONB +ASTEXT = operators.custom_op( + "->>", + precedence=_getitem_precedence, + natural_self_precedent=True, + eager_grouping=True, +) + +JSONPATH_ASTEXT = operators.custom_op( + "#>>", + precedence=_getitem_precedence, + natural_self_precedent=True, + eager_grouping=True, +) + +# JSONB + HSTORE +HAS_KEY = operators.custom_op( + "?", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +HAS_ALL = operators.custom_op( + "?&", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +HAS_ANY = operators.custom_op( + "?|", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +# JSONB +DELETE_PATH = operators.custom_op( + "#-", + precedence=_getitem_precedence, + natural_self_precedent=True, + eager_grouping=True, +) + +PATH_EXISTS = operators.custom_op( + "@?", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +PATH_MATCH = operators.custom_op( + "@@", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +# JSONB + ARRAY + HSTORE + RANGE +CONTAINS = operators.custom_op( + "@>", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +CONTAINED_BY = operators.custom_op( + "<@", + precedence=_eq_precedence, + natural_self_precedent=True, + eager_grouping=True, + is_comparison=True, +) + +# ARRAY + RANGE +OVERLAP = operators.custom_op( + "&&", + precedence=_eq_precedence, + is_comparison=True, +) + +# RANGE +STRICTLY_LEFT_OF = operators.custom_op( + "<<", precedence=_eq_precedence, is_comparison=True +) + +STRICTLY_RIGHT_OF = operators.custom_op( + ">>", precedence=_eq_precedence, is_comparison=True +) + +NOT_EXTEND_RIGHT_OF = operators.custom_op( + "&<", precedence=_eq_precedence, is_comparison=True +) + +NOT_EXTEND_LEFT_OF = operators.custom_op( + "&>", precedence=_eq_precedence, is_comparison=True +) + +ADJACENT_TO = operators.custom_op( + "-|-", precedence=_eq_precedence, is_comparison=True +) + +# HSTORE +GETITEM = operators.custom_op( + "->", + precedence=_getitem_precedence, + natural_self_precedent=True, + eager_grouping=True, +) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg8000.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg8000.py new file mode 100644 index 0000000..0151be0 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg8000.py @@ -0,0 +1,662 @@ +# dialects/postgresql/pg8000.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors <see AUTHORS +# file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql+pg8000 + :name: pg8000 + :dbapi: pg8000 + :connectstring: postgresql+pg8000://user:password@host:port/dbname[?key=value&key=value...] + :url: https://pypi.org/project/pg8000/ + +.. versionchanged:: 1.4 The pg8000 dialect has been updated for version + 1.16.6 and higher, and is again part of SQLAlchemy's continuous integration + with full feature support. + +.. _pg8000_unicode: + +Unicode +------- + +pg8000 will encode / decode string values between it and the server using the +PostgreSQL ``client_encoding`` parameter; by default this is the value in +the ``postgresql.conf`` file, which often defaults to ``SQL_ASCII``. +Typically, this can be changed to ``utf-8``, as a more useful default:: + + #client_encoding = sql_ascii # actually, defaults to database + # encoding + client_encoding = utf8 + +The ``client_encoding`` can be overridden for a session by executing the SQL: + +SET CLIENT_ENCODING TO 'utf8'; + +SQLAlchemy will execute this SQL on all new connections based on the value +passed to :func:`_sa.create_engine` using the ``client_encoding`` parameter:: + + engine = create_engine( + "postgresql+pg8000://user:pass@host/dbname", client_encoding='utf8') + +.. _pg8000_ssl: + +SSL Connections +--------------- + +pg8000 accepts a Python ``SSLContext`` object which may be specified using the +:paramref:`_sa.create_engine.connect_args` dictionary:: + + import ssl + ssl_context = ssl.create_default_context() + engine = sa.create_engine( + "postgresql+pg8000://scott:tiger@192.168.0.199/test", + connect_args={"ssl_context": ssl_context}, + ) + +If the server uses an automatically-generated certificate that is self-signed +or does not match the host name (as seen from the client), it may also be +necessary to disable hostname checking:: + + import ssl + ssl_context = ssl.create_default_context() + ssl_context.check_hostname = False + ssl_context.verify_mode = ssl.CERT_NONE + engine = sa.create_engine( + "postgresql+pg8000://scott:tiger@192.168.0.199/test", + connect_args={"ssl_context": ssl_context}, + ) + +.. _pg8000_isolation_level: + +pg8000 Transaction Isolation Level +------------------------------------- + +The pg8000 dialect offers the same isolation level settings as that +of the :ref:`psycopg2 <psycopg2_isolation_level>` dialect: + +* ``READ COMMITTED`` +* ``READ UNCOMMITTED`` +* ``REPEATABLE READ`` +* ``SERIALIZABLE`` +* ``AUTOCOMMIT`` + +.. seealso:: + + :ref:`postgresql_isolation_level` + + :ref:`psycopg2_isolation_level` + + +""" # noqa +import decimal +import re + +from . import ranges +from .array import ARRAY as PGARRAY +from .base import _DECIMAL_TYPES +from .base import _FLOAT_TYPES +from .base import _INT_TYPES +from .base import ENUM +from .base import INTERVAL +from .base import PGCompiler +from .base import PGDialect +from .base import PGExecutionContext +from .base import PGIdentifierPreparer +from .json import JSON +from .json import JSONB +from .json import JSONPathType +from .pg_catalog import _SpaceVector +from .pg_catalog import OIDVECTOR +from .types import CITEXT +from ... import exc +from ... import util +from ...engine import processors +from ...sql import sqltypes +from ...sql.elements import quoted_name + + +class _PGString(sqltypes.String): + render_bind_cast = True + + +class _PGNumeric(sqltypes.Numeric): + render_bind_cast = True + + def result_processor(self, dialect, coltype): + if self.asdecimal: + if coltype in _FLOAT_TYPES: + return processors.to_decimal_processor_factory( + decimal.Decimal, self._effective_decimal_return_scale + ) + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + # pg8000 returns Decimal natively for 1700 + return None + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + else: + if coltype in _FLOAT_TYPES: + # pg8000 returns float natively for 701 + return None + elif coltype in _DECIMAL_TYPES or coltype in _INT_TYPES: + return processors.to_float + else: + raise exc.InvalidRequestError( + "Unknown PG numeric type: %d" % coltype + ) + + +class _PGFloat(_PGNumeric, sqltypes.Float): + __visit_name__ = "float" + render_bind_cast = True + + +class _PGNumericNoBind(_PGNumeric): + def bind_processor(self, dialect): + return None + + +class _PGJSON(JSON): + render_bind_cast = True + + def result_processor(self, dialect, coltype): + return None + + +class _PGJSONB(JSONB): + render_bind_cast = True + + def result_processor(self, dialect, coltype): + return None + + +class _PGJSONIndexType(sqltypes.JSON.JSONIndexType): + def get_dbapi_type(self, dbapi): + raise NotImplementedError("should not be here") + + +class _PGJSONIntIndexType(sqltypes.JSON.JSONIntIndexType): + __visit_name__ = "json_int_index" + + render_bind_cast = True + + +class _PGJSONStrIndexType(sqltypes.JSON.JSONStrIndexType): + __visit_name__ = "json_str_index" + + render_bind_cast = True + + +class _PGJSONPathType(JSONPathType): + pass + + # DBAPI type 1009 + + +class _PGEnum(ENUM): + def get_dbapi_type(self, dbapi): + return dbapi.UNKNOWN + + +class _PGInterval(INTERVAL): + render_bind_cast = True + + def get_dbapi_type(self, dbapi): + return dbapi.INTERVAL + + @classmethod + def adapt_emulated_to_native(cls, interval, **kw): + return _PGInterval(precision=interval.second_precision) + + +class _PGTimeStamp(sqltypes.DateTime): + render_bind_cast = True + + +class _PGDate(sqltypes.Date): + render_bind_cast = True + + +class _PGTime(sqltypes.Time): + render_bind_cast = True + + +class _PGInteger(sqltypes.Integer): + render_bind_cast = True + + +class _PGSmallInteger(sqltypes.SmallInteger): + render_bind_cast = True + + +class _PGNullType(sqltypes.NullType): + pass + + +class _PGBigInteger(sqltypes.BigInteger): + render_bind_cast = True + + +class _PGBoolean(sqltypes.Boolean): + render_bind_cast = True + + +class _PGARRAY(PGARRAY): + render_bind_cast = True + + +class _PGOIDVECTOR(_SpaceVector, OIDVECTOR): + pass + + +class _Pg8000Range(ranges.AbstractSingleRangeImpl): + def bind_processor(self, dialect): + pg8000_Range = dialect.dbapi.Range + + def to_range(value): + if isinstance(value, ranges.Range): + value = pg8000_Range( + value.lower, value.upper, value.bounds, value.empty + ) + return value + + return to_range + + def result_processor(self, dialect, coltype): + def to_range(value): + if value is not None: + value = ranges.Range( + value.lower, + value.upper, + bounds=value.bounds, + empty=value.is_empty, + ) + return value + + return to_range + + +class _Pg8000MultiRange(ranges.AbstractMultiRangeImpl): + def bind_processor(self, dialect): + pg8000_Range = dialect.dbapi.Range + + def to_multirange(value): + if isinstance(value, list): + mr = [] + for v in value: + if isinstance(v, ranges.Range): + mr.append( + pg8000_Range(v.lower, v.upper, v.bounds, v.empty) + ) + else: + mr.append(v) + return mr + else: + return value + + return to_multirange + + def result_processor(self, dialect, coltype): + def to_multirange(value): + if value is None: + return None + else: + return ranges.MultiRange( + ranges.Range( + v.lower, v.upper, bounds=v.bounds, empty=v.is_empty + ) + for v in value + ) + + return to_multirange + + +_server_side_id = util.counter() + + +class PGExecutionContext_pg8000(PGExecutionContext): + def create_server_side_cursor(self): + ident = "c_%s_%s" % (hex(id(self))[2:], hex(_server_side_id())[2:]) + return ServerSideCursor(self._dbapi_connection.cursor(), ident) + + def pre_exec(self): + if not self.compiled: + return + + +class ServerSideCursor: + server_side = True + + def __init__(self, cursor, ident): + self.ident = ident + self.cursor = cursor + + @property + def connection(self): + return self.cursor.connection + + @property + def rowcount(self): + return self.cursor.rowcount + + @property + def description(self): + return self.cursor.description + + def execute(self, operation, args=(), stream=None): + op = "DECLARE " + self.ident + " NO SCROLL CURSOR FOR " + operation + self.cursor.execute(op, args, stream=stream) + return self + + def executemany(self, operation, param_sets): + self.cursor.executemany(operation, param_sets) + return self + + def fetchone(self): + self.cursor.execute("FETCH FORWARD 1 FROM " + self.ident) + return self.cursor.fetchone() + + def fetchmany(self, num=None): + if num is None: + return self.fetchall() + else: + self.cursor.execute( + "FETCH FORWARD " + str(int(num)) + " FROM " + self.ident + ) + return self.cursor.fetchall() + + def fetchall(self): + self.cursor.execute("FETCH FORWARD ALL FROM " + self.ident) + return self.cursor.fetchall() + + def close(self): + self.cursor.execute("CLOSE " + self.ident) + self.cursor.close() + + def setinputsizes(self, *sizes): + self.cursor.setinputsizes(*sizes) + + def setoutputsize(self, size, column=None): + pass + + +class PGCompiler_pg8000(PGCompiler): + def visit_mod_binary(self, binary, operator, **kw): + return ( + self.process(binary.left, **kw) + + " %% " + + self.process(binary.right, **kw) + ) + + +class PGIdentifierPreparer_pg8000(PGIdentifierPreparer): + def __init__(self, *args, **kwargs): + PGIdentifierPreparer.__init__(self, *args, **kwargs) + self._double_percents = False + + +class PGDialect_pg8000(PGDialect): + driver = "pg8000" + supports_statement_cache = True + + supports_unicode_statements = True + + supports_unicode_binds = True + + default_paramstyle = "format" + supports_sane_multi_rowcount = True + execution_ctx_cls = PGExecutionContext_pg8000 + statement_compiler = PGCompiler_pg8000 + preparer = PGIdentifierPreparer_pg8000 + supports_server_side_cursors = True + + render_bind_cast = True + + # reversed as of pg8000 1.16.6. 1.16.5 and lower + # are no longer compatible + description_encoding = None + # description_encoding = "use_encoding" + + colspecs = util.update_copy( + PGDialect.colspecs, + { + sqltypes.String: _PGString, + sqltypes.Numeric: _PGNumericNoBind, + sqltypes.Float: _PGFloat, + sqltypes.JSON: _PGJSON, + sqltypes.Boolean: _PGBoolean, + sqltypes.NullType: _PGNullType, + JSONB: _PGJSONB, + CITEXT: CITEXT, + sqltypes.JSON.JSONPathType: _PGJSONPathType, + sqltypes.JSON.JSONIndexType: _PGJSONIndexType, + sqltypes.JSON.JSONIntIndexType: _PGJSONIntIndexType, + sqltypes.JSON.JSONStrIndexType: _PGJSONStrIndexType, + sqltypes.Interval: _PGInterval, + INTERVAL: _PGInterval, + sqltypes.DateTime: _PGTimeStamp, + sqltypes.DateTime: _PGTimeStamp, + sqltypes.Date: _PGDate, + sqltypes.Time: _PGTime, + sqltypes.Integer: _PGInteger, + sqltypes.SmallInteger: _PGSmallInteger, + sqltypes.BigInteger: _PGBigInteger, + sqltypes.Enum: _PGEnum, + sqltypes.ARRAY: _PGARRAY, + OIDVECTOR: _PGOIDVECTOR, + ranges.INT4RANGE: _Pg8000Range, + ranges.INT8RANGE: _Pg8000Range, + ranges.NUMRANGE: _Pg8000Range, + ranges.DATERANGE: _Pg8000Range, + ranges.TSRANGE: _Pg8000Range, + ranges.TSTZRANGE: _Pg8000Range, + ranges.INT4MULTIRANGE: _Pg8000MultiRange, + ranges.INT8MULTIRANGE: _Pg8000MultiRange, + ranges.NUMMULTIRANGE: _Pg8000MultiRange, + ranges.DATEMULTIRANGE: _Pg8000MultiRange, + ranges.TSMULTIRANGE: _Pg8000MultiRange, + ranges.TSTZMULTIRANGE: _Pg8000MultiRange, + }, + ) + + def __init__(self, client_encoding=None, **kwargs): + PGDialect.__init__(self, **kwargs) + self.client_encoding = client_encoding + + if self._dbapi_version < (1, 16, 6): + raise NotImplementedError("pg8000 1.16.6 or greater is required") + + if self._native_inet_types: + raise NotImplementedError( + "The pg8000 dialect does not fully implement " + "ipaddress type handling; INET is supported by default, " + "CIDR is not" + ) + + @util.memoized_property + def _dbapi_version(self): + if self.dbapi and hasattr(self.dbapi, "__version__"): + return tuple( + [ + int(x) + for x in re.findall( + r"(\d+)(?:[-\.]?|$)", self.dbapi.__version__ + ) + ] + ) + else: + return (99, 99, 99) + + @classmethod + def import_dbapi(cls): + return __import__("pg8000") + + def create_connect_args(self, url): + opts = url.translate_connect_args(username="user") + if "port" in opts: + opts["port"] = int(opts["port"]) + opts.update(url.query) + return ([], opts) + + def is_disconnect(self, e, connection, cursor): + if isinstance(e, self.dbapi.InterfaceError) and "network error" in str( + e + ): + # new as of pg8000 1.19.0 for broken connections + return True + + # connection was closed normally + return "connection is closed" in str(e) + + def get_isolation_level_values(self, dbapi_connection): + return ( + "AUTOCOMMIT", + "READ COMMITTED", + "READ UNCOMMITTED", + "REPEATABLE READ", + "SERIALIZABLE", + ) + + def set_isolation_level(self, dbapi_connection, level): + level = level.replace("_", " ") + + if level == "AUTOCOMMIT": + dbapi_connection.autocommit = True + else: + dbapi_connection.autocommit = False + cursor = dbapi_connection.cursor() + cursor.execute( + "SET SESSION CHARACTERISTICS AS TRANSACTION " + f"ISOLATION LEVEL {level}" + ) + cursor.execute("COMMIT") + cursor.close() + + def set_readonly(self, connection, value): + cursor = connection.cursor() + try: + cursor.execute( + "SET SESSION CHARACTERISTICS AS TRANSACTION %s" + % ("READ ONLY" if value else "READ WRITE") + ) + cursor.execute("COMMIT") + finally: + cursor.close() + + def get_readonly(self, connection): + cursor = connection.cursor() + try: + cursor.execute("show transaction_read_only") + val = cursor.fetchone()[0] + finally: + cursor.close() + + return val == "on" + + def set_deferrable(self, connection, value): + cursor = connection.cursor() + try: + cursor.execute( + "SET SESSION CHARACTERISTICS AS TRANSACTION %s" + % ("DEFERRABLE" if value else "NOT DEFERRABLE") + ) + cursor.execute("COMMIT") + finally: + cursor.close() + + def get_deferrable(self, connection): + cursor = connection.cursor() + try: + cursor.execute("show transaction_deferrable") + val = cursor.fetchone()[0] + finally: + cursor.close() + + return val == "on" + + def _set_client_encoding(self, dbapi_connection, client_encoding): + cursor = dbapi_connection.cursor() + cursor.execute( + f"""SET CLIENT_ENCODING TO '{ + client_encoding.replace("'", "''") + }'""" + ) + cursor.execute("COMMIT") + cursor.close() + + def do_begin_twophase(self, connection, xid): + connection.connection.tpc_begin((0, xid, "")) + + def do_prepare_twophase(self, connection, xid): + connection.connection.tpc_prepare() + + def do_rollback_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + connection.connection.tpc_rollback((0, xid, "")) + + def do_commit_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + connection.connection.tpc_commit((0, xid, "")) + + def do_recover_twophase(self, connection): + return [row[1] for row in connection.connection.tpc_recover()] + + def on_connect(self): + fns = [] + + def on_connect(conn): + conn.py_types[quoted_name] = conn.py_types[str] + + fns.append(on_connect) + + if self.client_encoding is not None: + + def on_connect(conn): + self._set_client_encoding(conn, self.client_encoding) + + fns.append(on_connect) + + if self._native_inet_types is False: + + def on_connect(conn): + # inet + conn.register_in_adapter(869, lambda s: s) + + # cidr + conn.register_in_adapter(650, lambda s: s) + + fns.append(on_connect) + + if self._json_deserializer: + + def on_connect(conn): + # json + conn.register_in_adapter(114, self._json_deserializer) + + # jsonb + conn.register_in_adapter(3802, self._json_deserializer) + + fns.append(on_connect) + + if len(fns) > 0: + + def on_connect(conn): + for fn in fns: + fn(conn) + + return on_connect + else: + return None + + @util.memoized_property + def _dialect_specific_select_one(self): + return ";" + + +dialect = PGDialect_pg8000 diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py new file mode 100644 index 0000000..9b5562c --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/pg_catalog.py @@ -0,0 +1,300 @@ +# dialects/postgresql/pg_catalog.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +from .array import ARRAY +from .types import OID +from .types import REGCLASS +from ... import Column +from ... import func +from ... import MetaData +from ... import Table +from ...types import BigInteger +from ...types import Boolean +from ...types import CHAR +from ...types import Float +from ...types import Integer +from ...types import SmallInteger +from ...types import String +from ...types import Text +from ...types import TypeDecorator + + +# types +class NAME(TypeDecorator): + impl = String(64, collation="C") + cache_ok = True + + +class PG_NODE_TREE(TypeDecorator): + impl = Text(collation="C") + cache_ok = True + + +class INT2VECTOR(TypeDecorator): + impl = ARRAY(SmallInteger) + cache_ok = True + + +class OIDVECTOR(TypeDecorator): + impl = ARRAY(OID) + cache_ok = True + + +class _SpaceVector: + def result_processor(self, dialect, coltype): + def process(value): + if value is None: + return value + return [int(p) for p in value.split(" ")] + + return process + + +REGPROC = REGCLASS # seems an alias + +# functions +_pg_cat = func.pg_catalog +quote_ident = _pg_cat.quote_ident +pg_table_is_visible = _pg_cat.pg_table_is_visible +pg_type_is_visible = _pg_cat.pg_type_is_visible +pg_get_viewdef = _pg_cat.pg_get_viewdef +pg_get_serial_sequence = _pg_cat.pg_get_serial_sequence +format_type = _pg_cat.format_type +pg_get_expr = _pg_cat.pg_get_expr +pg_get_constraintdef = _pg_cat.pg_get_constraintdef +pg_get_indexdef = _pg_cat.pg_get_indexdef + +# constants +RELKINDS_TABLE_NO_FOREIGN = ("r", "p") +RELKINDS_TABLE = RELKINDS_TABLE_NO_FOREIGN + ("f",) +RELKINDS_VIEW = ("v",) +RELKINDS_MAT_VIEW = ("m",) +RELKINDS_ALL_TABLE_LIKE = RELKINDS_TABLE + RELKINDS_VIEW + RELKINDS_MAT_VIEW + +# tables +pg_catalog_meta = MetaData(schema="pg_catalog") + +pg_namespace = Table( + "pg_namespace", + pg_catalog_meta, + Column("oid", OID), + Column("nspname", NAME), + Column("nspowner", OID), +) + +pg_class = Table( + "pg_class", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("relname", NAME), + Column("relnamespace", OID), + Column("reltype", OID), + Column("reloftype", OID), + Column("relowner", OID), + Column("relam", OID), + Column("relfilenode", OID), + Column("reltablespace", OID), + Column("relpages", Integer), + Column("reltuples", Float), + Column("relallvisible", Integer, info={"server_version": (9, 2)}), + Column("reltoastrelid", OID), + Column("relhasindex", Boolean), + Column("relisshared", Boolean), + Column("relpersistence", CHAR, info={"server_version": (9, 1)}), + Column("relkind", CHAR), + Column("relnatts", SmallInteger), + Column("relchecks", SmallInteger), + Column("relhasrules", Boolean), + Column("relhastriggers", Boolean), + Column("relhassubclass", Boolean), + Column("relrowsecurity", Boolean), + Column("relforcerowsecurity", Boolean, info={"server_version": (9, 5)}), + Column("relispopulated", Boolean, info={"server_version": (9, 3)}), + Column("relreplident", CHAR, info={"server_version": (9, 4)}), + Column("relispartition", Boolean, info={"server_version": (10,)}), + Column("relrewrite", OID, info={"server_version": (11,)}), + Column("reloptions", ARRAY(Text)), +) + +pg_type = Table( + "pg_type", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("typname", NAME), + Column("typnamespace", OID), + Column("typowner", OID), + Column("typlen", SmallInteger), + Column("typbyval", Boolean), + Column("typtype", CHAR), + Column("typcategory", CHAR), + Column("typispreferred", Boolean), + Column("typisdefined", Boolean), + Column("typdelim", CHAR), + Column("typrelid", OID), + Column("typelem", OID), + Column("typarray", OID), + Column("typinput", REGPROC), + Column("typoutput", REGPROC), + Column("typreceive", REGPROC), + Column("typsend", REGPROC), + Column("typmodin", REGPROC), + Column("typmodout", REGPROC), + Column("typanalyze", REGPROC), + Column("typalign", CHAR), + Column("typstorage", CHAR), + Column("typnotnull", Boolean), + Column("typbasetype", OID), + Column("typtypmod", Integer), + Column("typndims", Integer), + Column("typcollation", OID, info={"server_version": (9, 1)}), + Column("typdefault", Text), +) + +pg_index = Table( + "pg_index", + pg_catalog_meta, + Column("indexrelid", OID), + Column("indrelid", OID), + Column("indnatts", SmallInteger), + Column("indnkeyatts", SmallInteger, info={"server_version": (11,)}), + Column("indisunique", Boolean), + Column("indnullsnotdistinct", Boolean, info={"server_version": (15,)}), + Column("indisprimary", Boolean), + Column("indisexclusion", Boolean, info={"server_version": (9, 1)}), + Column("indimmediate", Boolean), + Column("indisclustered", Boolean), + Column("indisvalid", Boolean), + Column("indcheckxmin", Boolean), + Column("indisready", Boolean), + Column("indislive", Boolean, info={"server_version": (9, 3)}), # 9.3 + Column("indisreplident", Boolean), + Column("indkey", INT2VECTOR), + Column("indcollation", OIDVECTOR, info={"server_version": (9, 1)}), # 9.1 + Column("indclass", OIDVECTOR), + Column("indoption", INT2VECTOR), + Column("indexprs", PG_NODE_TREE), + Column("indpred", PG_NODE_TREE), +) + +pg_attribute = Table( + "pg_attribute", + pg_catalog_meta, + Column("attrelid", OID), + Column("attname", NAME), + Column("atttypid", OID), + Column("attstattarget", Integer), + Column("attlen", SmallInteger), + Column("attnum", SmallInteger), + Column("attndims", Integer), + Column("attcacheoff", Integer), + Column("atttypmod", Integer), + Column("attbyval", Boolean), + Column("attstorage", CHAR), + Column("attalign", CHAR), + Column("attnotnull", Boolean), + Column("atthasdef", Boolean), + Column("atthasmissing", Boolean, info={"server_version": (11,)}), + Column("attidentity", CHAR, info={"server_version": (10,)}), + Column("attgenerated", CHAR, info={"server_version": (12,)}), + Column("attisdropped", Boolean), + Column("attislocal", Boolean), + Column("attinhcount", Integer), + Column("attcollation", OID, info={"server_version": (9, 1)}), +) + +pg_constraint = Table( + "pg_constraint", + pg_catalog_meta, + Column("oid", OID), # 9.3 + Column("conname", NAME), + Column("connamespace", OID), + Column("contype", CHAR), + Column("condeferrable", Boolean), + Column("condeferred", Boolean), + Column("convalidated", Boolean, info={"server_version": (9, 1)}), + Column("conrelid", OID), + Column("contypid", OID), + Column("conindid", OID), + Column("conparentid", OID, info={"server_version": (11,)}), + Column("confrelid", OID), + Column("confupdtype", CHAR), + Column("confdeltype", CHAR), + Column("confmatchtype", CHAR), + Column("conislocal", Boolean), + Column("coninhcount", Integer), + Column("connoinherit", Boolean, info={"server_version": (9, 2)}), + Column("conkey", ARRAY(SmallInteger)), + Column("confkey", ARRAY(SmallInteger)), +) + +pg_sequence = Table( + "pg_sequence", + pg_catalog_meta, + Column("seqrelid", OID), + Column("seqtypid", OID), + Column("seqstart", BigInteger), + Column("seqincrement", BigInteger), + Column("seqmax", BigInteger), + Column("seqmin", BigInteger), + Column("seqcache", BigInteger), + Column("seqcycle", Boolean), + info={"server_version": (10,)}, +) + +pg_attrdef = Table( + "pg_attrdef", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("adrelid", OID), + Column("adnum", SmallInteger), + Column("adbin", PG_NODE_TREE), +) + +pg_description = Table( + "pg_description", + pg_catalog_meta, + Column("objoid", OID), + Column("classoid", OID), + Column("objsubid", Integer), + Column("description", Text(collation="C")), +) + +pg_enum = Table( + "pg_enum", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("enumtypid", OID), + Column("enumsortorder", Float(), info={"server_version": (9, 1)}), + Column("enumlabel", NAME), +) + +pg_am = Table( + "pg_am", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("amname", NAME), + Column("amhandler", REGPROC, info={"server_version": (9, 6)}), + Column("amtype", CHAR, info={"server_version": (9, 6)}), +) + +pg_collation = Table( + "pg_collation", + pg_catalog_meta, + Column("oid", OID, info={"server_version": (9, 3)}), + Column("collname", NAME), + Column("collnamespace", OID), + Column("collowner", OID), + Column("collprovider", CHAR, info={"server_version": (10,)}), + Column("collisdeterministic", Boolean, info={"server_version": (12,)}), + Column("collencoding", Integer), + Column("collcollate", Text), + Column("collctype", Text), + Column("colliculocale", Text), + Column("collicurules", Text, info={"server_version": (16,)}), + Column("collversion", Text, info={"server_version": (10,)}), +) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/provision.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/provision.py new file mode 100644 index 0000000..a87bb93 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/provision.py @@ -0,0 +1,175 @@ +# dialects/postgresql/provision.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +import time + +from ... import exc +from ... import inspect +from ... import text +from ...testing import warn_test_suite +from ...testing.provision import create_db +from ...testing.provision import drop_all_schema_objects_post_tables +from ...testing.provision import drop_all_schema_objects_pre_tables +from ...testing.provision import drop_db +from ...testing.provision import log +from ...testing.provision import post_configure_engine +from ...testing.provision import prepare_for_drop_tables +from ...testing.provision import set_default_schema_on_connection +from ...testing.provision import temp_table_keyword_args +from ...testing.provision import upsert + + +@create_db.for_db("postgresql") +def _pg_create_db(cfg, eng, ident): + template_db = cfg.options.postgresql_templatedb + + with eng.execution_options(isolation_level="AUTOCOMMIT").begin() as conn: + if not template_db: + template_db = conn.exec_driver_sql( + "select current_database()" + ).scalar() + + attempt = 0 + while True: + try: + conn.exec_driver_sql( + "CREATE DATABASE %s TEMPLATE %s" % (ident, template_db) + ) + except exc.OperationalError as err: + attempt += 1 + if attempt >= 3: + raise + if "accessed by other users" in str(err): + log.info( + "Waiting to create %s, URI %r, " + "template DB %s is in use sleeping for .5", + ident, + eng.url, + template_db, + ) + time.sleep(0.5) + except: + raise + else: + break + + +@drop_db.for_db("postgresql") +def _pg_drop_db(cfg, eng, ident): + with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn: + with conn.begin(): + conn.execute( + text( + "select pg_terminate_backend(pid) from pg_stat_activity " + "where usename=current_user and pid != pg_backend_pid() " + "and datname=:dname" + ), + dict(dname=ident), + ) + conn.exec_driver_sql("DROP DATABASE %s" % ident) + + +@temp_table_keyword_args.for_db("postgresql") +def _postgresql_temp_table_keyword_args(cfg, eng): + return {"prefixes": ["TEMPORARY"]} + + +@set_default_schema_on_connection.for_db("postgresql") +def _postgresql_set_default_schema_on_connection( + cfg, dbapi_connection, schema_name +): + existing_autocommit = dbapi_connection.autocommit + dbapi_connection.autocommit = True + cursor = dbapi_connection.cursor() + cursor.execute("SET SESSION search_path='%s'" % schema_name) + cursor.close() + dbapi_connection.autocommit = existing_autocommit + + +@drop_all_schema_objects_pre_tables.for_db("postgresql") +def drop_all_schema_objects_pre_tables(cfg, eng): + with eng.connect().execution_options(isolation_level="AUTOCOMMIT") as conn: + for xid in conn.exec_driver_sql( + "select gid from pg_prepared_xacts" + ).scalars(): + conn.execute("ROLLBACK PREPARED '%s'" % xid) + + +@drop_all_schema_objects_post_tables.for_db("postgresql") +def drop_all_schema_objects_post_tables(cfg, eng): + from sqlalchemy.dialects import postgresql + + inspector = inspect(eng) + with eng.begin() as conn: + for enum in inspector.get_enums("*"): + conn.execute( + postgresql.DropEnumType( + postgresql.ENUM(name=enum["name"], schema=enum["schema"]) + ) + ) + + +@prepare_for_drop_tables.for_db("postgresql") +def prepare_for_drop_tables(config, connection): + """Ensure there are no locks on the current username/database.""" + + result = connection.exec_driver_sql( + "select pid, state, wait_event_type, query " + # "select pg_terminate_backend(pid), state, wait_event_type " + "from pg_stat_activity where " + "usename=current_user " + "and datname=current_database() and state='idle in transaction' " + "and pid != pg_backend_pid()" + ) + rows = result.all() # noqa + if rows: + warn_test_suite( + "PostgreSQL may not be able to DROP tables due to " + "idle in transaction: %s" + % ("; ".join(row._mapping["query"] for row in rows)) + ) + + +@upsert.for_db("postgresql") +def _upsert( + cfg, table, returning, *, set_lambda=None, sort_by_parameter_order=False +): + from sqlalchemy.dialects.postgresql import insert + + stmt = insert(table) + + table_pk = inspect(table).selectable + + if set_lambda: + stmt = stmt.on_conflict_do_update( + index_elements=table_pk.primary_key, set_=set_lambda(stmt.excluded) + ) + else: + stmt = stmt.on_conflict_do_nothing() + + stmt = stmt.returning( + *returning, sort_by_parameter_order=sort_by_parameter_order + ) + return stmt + + +_extensions = [ + ("citext", (13,)), + ("hstore", (13,)), +] + + +@post_configure_engine.for_db("postgresql") +def _create_citext_extension(url, engine, follower_ident): + with engine.connect() as conn: + for extension, min_version in _extensions: + if conn.dialect.server_version_info >= min_version: + conn.execute( + text(f"CREATE EXTENSION IF NOT EXISTS {extension}") + ) + conn.commit() diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg.py new file mode 100644 index 0000000..90177a4 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg.py @@ -0,0 +1,749 @@ +# dialects/postgresql/psycopg.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql+psycopg + :name: psycopg (a.k.a. psycopg 3) + :dbapi: psycopg + :connectstring: postgresql+psycopg://user:password@host:port/dbname[?key=value&key=value...] + :url: https://pypi.org/project/psycopg/ + +``psycopg`` is the package and module name for version 3 of the ``psycopg`` +database driver, formerly known as ``psycopg2``. This driver is different +enough from its ``psycopg2`` predecessor that SQLAlchemy supports it +via a totally separate dialect; support for ``psycopg2`` is expected to remain +for as long as that package continues to function for modern Python versions, +and also remains the default dialect for the ``postgresql://`` dialect +series. + +The SQLAlchemy ``psycopg`` dialect provides both a sync and an async +implementation under the same dialect name. The proper version is +selected depending on how the engine is created: + +* calling :func:`_sa.create_engine` with ``postgresql+psycopg://...`` will + automatically select the sync version, e.g.:: + + from sqlalchemy import create_engine + sync_engine = create_engine("postgresql+psycopg://scott:tiger@localhost/test") + +* calling :func:`_asyncio.create_async_engine` with + ``postgresql+psycopg://...`` will automatically select the async version, + e.g.:: + + from sqlalchemy.ext.asyncio import create_async_engine + asyncio_engine = create_async_engine("postgresql+psycopg://scott:tiger@localhost/test") + +The asyncio version of the dialect may also be specified explicitly using the +``psycopg_async`` suffix, as:: + + from sqlalchemy.ext.asyncio import create_async_engine + asyncio_engine = create_async_engine("postgresql+psycopg_async://scott:tiger@localhost/test") + +.. seealso:: + + :ref:`postgresql_psycopg2` - The SQLAlchemy ``psycopg`` + dialect shares most of its behavior with the ``psycopg2`` dialect. + Further documentation is available there. + +""" # noqa +from __future__ import annotations + +import logging +import re +from typing import cast +from typing import TYPE_CHECKING + +from . import ranges +from ._psycopg_common import _PGDialect_common_psycopg +from ._psycopg_common import _PGExecutionContext_common_psycopg +from .base import INTERVAL +from .base import PGCompiler +from .base import PGIdentifierPreparer +from .base import REGCONFIG +from .json import JSON +from .json import JSONB +from .json import JSONPathType +from .types import CITEXT +from ... import pool +from ... import util +from ...engine import AdaptedConnection +from ...sql import sqltypes +from ...util.concurrency import await_fallback +from ...util.concurrency import await_only + +if TYPE_CHECKING: + from typing import Iterable + + from psycopg import AsyncConnection + +logger = logging.getLogger("sqlalchemy.dialects.postgresql") + + +class _PGString(sqltypes.String): + render_bind_cast = True + + +class _PGREGCONFIG(REGCONFIG): + render_bind_cast = True + + +class _PGJSON(JSON): + render_bind_cast = True + + def bind_processor(self, dialect): + return self._make_bind_processor(None, dialect._psycopg_Json) + + def result_processor(self, dialect, coltype): + return None + + +class _PGJSONB(JSONB): + render_bind_cast = True + + def bind_processor(self, dialect): + return self._make_bind_processor(None, dialect._psycopg_Jsonb) + + def result_processor(self, dialect, coltype): + return None + + +class _PGJSONIntIndexType(sqltypes.JSON.JSONIntIndexType): + __visit_name__ = "json_int_index" + + render_bind_cast = True + + +class _PGJSONStrIndexType(sqltypes.JSON.JSONStrIndexType): + __visit_name__ = "json_str_index" + + render_bind_cast = True + + +class _PGJSONPathType(JSONPathType): + pass + + +class _PGInterval(INTERVAL): + render_bind_cast = True + + +class _PGTimeStamp(sqltypes.DateTime): + render_bind_cast = True + + +class _PGDate(sqltypes.Date): + render_bind_cast = True + + +class _PGTime(sqltypes.Time): + render_bind_cast = True + + +class _PGInteger(sqltypes.Integer): + render_bind_cast = True + + +class _PGSmallInteger(sqltypes.SmallInteger): + render_bind_cast = True + + +class _PGNullType(sqltypes.NullType): + render_bind_cast = True + + +class _PGBigInteger(sqltypes.BigInteger): + render_bind_cast = True + + +class _PGBoolean(sqltypes.Boolean): + render_bind_cast = True + + +class _PsycopgRange(ranges.AbstractSingleRangeImpl): + def bind_processor(self, dialect): + psycopg_Range = cast(PGDialect_psycopg, dialect)._psycopg_Range + + def to_range(value): + if isinstance(value, ranges.Range): + value = psycopg_Range( + value.lower, value.upper, value.bounds, value.empty + ) + return value + + return to_range + + def result_processor(self, dialect, coltype): + def to_range(value): + if value is not None: + value = ranges.Range( + value._lower, + value._upper, + bounds=value._bounds if value._bounds else "[)", + empty=not value._bounds, + ) + return value + + return to_range + + +class _PsycopgMultiRange(ranges.AbstractMultiRangeImpl): + def bind_processor(self, dialect): + psycopg_Range = cast(PGDialect_psycopg, dialect)._psycopg_Range + psycopg_Multirange = cast( + PGDialect_psycopg, dialect + )._psycopg_Multirange + + NoneType = type(None) + + def to_range(value): + if isinstance(value, (str, NoneType, psycopg_Multirange)): + return value + + return psycopg_Multirange( + [ + psycopg_Range( + element.lower, + element.upper, + element.bounds, + element.empty, + ) + for element in cast("Iterable[ranges.Range]", value) + ] + ) + + return to_range + + def result_processor(self, dialect, coltype): + def to_range(value): + if value is None: + return None + else: + return ranges.MultiRange( + ranges.Range( + elem._lower, + elem._upper, + bounds=elem._bounds if elem._bounds else "[)", + empty=not elem._bounds, + ) + for elem in value + ) + + return to_range + + +class PGExecutionContext_psycopg(_PGExecutionContext_common_psycopg): + pass + + +class PGCompiler_psycopg(PGCompiler): + pass + + +class PGIdentifierPreparer_psycopg(PGIdentifierPreparer): + pass + + +def _log_notices(diagnostic): + logger.info("%s: %s", diagnostic.severity, diagnostic.message_primary) + + +class PGDialect_psycopg(_PGDialect_common_psycopg): + driver = "psycopg" + + supports_statement_cache = True + supports_server_side_cursors = True + default_paramstyle = "pyformat" + supports_sane_multi_rowcount = True + + execution_ctx_cls = PGExecutionContext_psycopg + statement_compiler = PGCompiler_psycopg + preparer = PGIdentifierPreparer_psycopg + psycopg_version = (0, 0) + + _has_native_hstore = True + _psycopg_adapters_map = None + + colspecs = util.update_copy( + _PGDialect_common_psycopg.colspecs, + { + sqltypes.String: _PGString, + REGCONFIG: _PGREGCONFIG, + JSON: _PGJSON, + CITEXT: CITEXT, + sqltypes.JSON: _PGJSON, + JSONB: _PGJSONB, + sqltypes.JSON.JSONPathType: _PGJSONPathType, + sqltypes.JSON.JSONIntIndexType: _PGJSONIntIndexType, + sqltypes.JSON.JSONStrIndexType: _PGJSONStrIndexType, + sqltypes.Interval: _PGInterval, + INTERVAL: _PGInterval, + sqltypes.Date: _PGDate, + sqltypes.DateTime: _PGTimeStamp, + sqltypes.Time: _PGTime, + sqltypes.Integer: _PGInteger, + sqltypes.SmallInteger: _PGSmallInteger, + sqltypes.BigInteger: _PGBigInteger, + ranges.AbstractSingleRange: _PsycopgRange, + ranges.AbstractMultiRange: _PsycopgMultiRange, + }, + ) + + def __init__(self, **kwargs): + super().__init__(**kwargs) + + if self.dbapi: + m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", self.dbapi.__version__) + if m: + self.psycopg_version = tuple( + int(x) for x in m.group(1, 2, 3) if x is not None + ) + + if self.psycopg_version < (3, 0, 2): + raise ImportError( + "psycopg version 3.0.2 or higher is required." + ) + + from psycopg.adapt import AdaptersMap + + self._psycopg_adapters_map = adapters_map = AdaptersMap( + self.dbapi.adapters + ) + + if self._native_inet_types is False: + import psycopg.types.string + + adapters_map.register_loader( + "inet", psycopg.types.string.TextLoader + ) + adapters_map.register_loader( + "cidr", psycopg.types.string.TextLoader + ) + + if self._json_deserializer: + from psycopg.types.json import set_json_loads + + set_json_loads(self._json_deserializer, adapters_map) + + if self._json_serializer: + from psycopg.types.json import set_json_dumps + + set_json_dumps(self._json_serializer, adapters_map) + + def create_connect_args(self, url): + # see https://github.com/psycopg/psycopg/issues/83 + cargs, cparams = super().create_connect_args(url) + + if self._psycopg_adapters_map: + cparams["context"] = self._psycopg_adapters_map + if self.client_encoding is not None: + cparams["client_encoding"] = self.client_encoding + return cargs, cparams + + def _type_info_fetch(self, connection, name): + from psycopg.types import TypeInfo + + return TypeInfo.fetch(connection.connection.driver_connection, name) + + def initialize(self, connection): + super().initialize(connection) + + # PGDialect.initialize() checks server version for <= 8.2 and sets + # this flag to False if so + if not self.insert_returning: + self.insert_executemany_returning = False + + # HSTORE can't be registered until we have a connection so that + # we can look up its OID, so we set up this adapter in + # initialize() + if self.use_native_hstore: + info = self._type_info_fetch(connection, "hstore") + self._has_native_hstore = info is not None + if self._has_native_hstore: + from psycopg.types.hstore import register_hstore + + # register the adapter for connections made subsequent to + # this one + register_hstore(info, self._psycopg_adapters_map) + + # register the adapter for this connection + register_hstore(info, connection.connection) + + @classmethod + def import_dbapi(cls): + import psycopg + + return psycopg + + @classmethod + def get_async_dialect_cls(cls, url): + return PGDialectAsync_psycopg + + @util.memoized_property + def _isolation_lookup(self): + return { + "READ COMMITTED": self.dbapi.IsolationLevel.READ_COMMITTED, + "READ UNCOMMITTED": self.dbapi.IsolationLevel.READ_UNCOMMITTED, + "REPEATABLE READ": self.dbapi.IsolationLevel.REPEATABLE_READ, + "SERIALIZABLE": self.dbapi.IsolationLevel.SERIALIZABLE, + } + + @util.memoized_property + def _psycopg_Json(self): + from psycopg.types import json + + return json.Json + + @util.memoized_property + def _psycopg_Jsonb(self): + from psycopg.types import json + + return json.Jsonb + + @util.memoized_property + def _psycopg_TransactionStatus(self): + from psycopg.pq import TransactionStatus + + return TransactionStatus + + @util.memoized_property + def _psycopg_Range(self): + from psycopg.types.range import Range + + return Range + + @util.memoized_property + def _psycopg_Multirange(self): + from psycopg.types.multirange import Multirange + + return Multirange + + def _do_isolation_level(self, connection, autocommit, isolation_level): + connection.autocommit = autocommit + connection.isolation_level = isolation_level + + def get_isolation_level(self, dbapi_connection): + status_before = dbapi_connection.info.transaction_status + value = super().get_isolation_level(dbapi_connection) + + # don't rely on psycopg providing enum symbols, compare with + # eq/ne + if status_before == self._psycopg_TransactionStatus.IDLE: + dbapi_connection.rollback() + return value + + def set_isolation_level(self, dbapi_connection, level): + if level == "AUTOCOMMIT": + self._do_isolation_level( + dbapi_connection, autocommit=True, isolation_level=None + ) + else: + self._do_isolation_level( + dbapi_connection, + autocommit=False, + isolation_level=self._isolation_lookup[level], + ) + + def set_readonly(self, connection, value): + connection.read_only = value + + def get_readonly(self, connection): + return connection.read_only + + def on_connect(self): + def notices(conn): + conn.add_notice_handler(_log_notices) + + fns = [notices] + + if self.isolation_level is not None: + + def on_connect(conn): + self.set_isolation_level(conn, self.isolation_level) + + fns.append(on_connect) + + # fns always has the notices function + def on_connect(conn): + for fn in fns: + fn(conn) + + return on_connect + + def is_disconnect(self, e, connection, cursor): + if isinstance(e, self.dbapi.Error) and connection is not None: + if connection.closed or connection.broken: + return True + return False + + def _do_prepared_twophase(self, connection, command, recover=False): + dbapi_conn = connection.connection.dbapi_connection + if ( + recover + # don't rely on psycopg providing enum symbols, compare with + # eq/ne + or dbapi_conn.info.transaction_status + != self._psycopg_TransactionStatus.IDLE + ): + dbapi_conn.rollback() + before_autocommit = dbapi_conn.autocommit + try: + if not before_autocommit: + self._do_autocommit(dbapi_conn, True) + dbapi_conn.execute(command) + finally: + if not before_autocommit: + self._do_autocommit(dbapi_conn, before_autocommit) + + def do_rollback_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + if is_prepared: + self._do_prepared_twophase( + connection, f"ROLLBACK PREPARED '{xid}'", recover=recover + ) + else: + self.do_rollback(connection.connection) + + def do_commit_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + if is_prepared: + self._do_prepared_twophase( + connection, f"COMMIT PREPARED '{xid}'", recover=recover + ) + else: + self.do_commit(connection.connection) + + @util.memoized_property + def _dialect_specific_select_one(self): + return ";" + + +class AsyncAdapt_psycopg_cursor: + __slots__ = ("_cursor", "await_", "_rows") + + _psycopg_ExecStatus = None + + def __init__(self, cursor, await_) -> None: + self._cursor = cursor + self.await_ = await_ + self._rows = [] + + def __getattr__(self, name): + return getattr(self._cursor, name) + + @property + def arraysize(self): + return self._cursor.arraysize + + @arraysize.setter + def arraysize(self, value): + self._cursor.arraysize = value + + def close(self): + self._rows.clear() + # Normal cursor just call _close() in a non-sync way. + self._cursor._close() + + def execute(self, query, params=None, **kw): + result = self.await_(self._cursor.execute(query, params, **kw)) + # sqlalchemy result is not async, so need to pull all rows here + res = self._cursor.pgresult + + # don't rely on psycopg providing enum symbols, compare with + # eq/ne + if res and res.status == self._psycopg_ExecStatus.TUPLES_OK: + rows = self.await_(self._cursor.fetchall()) + if not isinstance(rows, list): + self._rows = list(rows) + else: + self._rows = rows + return result + + def executemany(self, query, params_seq): + return self.await_(self._cursor.executemany(query, params_seq)) + + def __iter__(self): + # TODO: try to avoid pop(0) on a list + while self._rows: + yield self._rows.pop(0) + + def fetchone(self): + if self._rows: + # TODO: try to avoid pop(0) on a list + return self._rows.pop(0) + else: + return None + + def fetchmany(self, size=None): + if size is None: + size = self._cursor.arraysize + + retval = self._rows[0:size] + self._rows = self._rows[size:] + return retval + + def fetchall(self): + retval = self._rows + self._rows = [] + return retval + + +class AsyncAdapt_psycopg_ss_cursor(AsyncAdapt_psycopg_cursor): + def execute(self, query, params=None, **kw): + self.await_(self._cursor.execute(query, params, **kw)) + return self + + def close(self): + self.await_(self._cursor.close()) + + def fetchone(self): + return self.await_(self._cursor.fetchone()) + + def fetchmany(self, size=0): + return self.await_(self._cursor.fetchmany(size)) + + def fetchall(self): + return self.await_(self._cursor.fetchall()) + + def __iter__(self): + iterator = self._cursor.__aiter__() + while True: + try: + yield self.await_(iterator.__anext__()) + except StopAsyncIteration: + break + + +class AsyncAdapt_psycopg_connection(AdaptedConnection): + _connection: AsyncConnection + __slots__ = () + await_ = staticmethod(await_only) + + def __init__(self, connection) -> None: + self._connection = connection + + def __getattr__(self, name): + return getattr(self._connection, name) + + def execute(self, query, params=None, **kw): + cursor = self.await_(self._connection.execute(query, params, **kw)) + return AsyncAdapt_psycopg_cursor(cursor, self.await_) + + def cursor(self, *args, **kw): + cursor = self._connection.cursor(*args, **kw) + if hasattr(cursor, "name"): + return AsyncAdapt_psycopg_ss_cursor(cursor, self.await_) + else: + return AsyncAdapt_psycopg_cursor(cursor, self.await_) + + def commit(self): + self.await_(self._connection.commit()) + + def rollback(self): + self.await_(self._connection.rollback()) + + def close(self): + self.await_(self._connection.close()) + + @property + def autocommit(self): + return self._connection.autocommit + + @autocommit.setter + def autocommit(self, value): + self.set_autocommit(value) + + def set_autocommit(self, value): + self.await_(self._connection.set_autocommit(value)) + + def set_isolation_level(self, value): + self.await_(self._connection.set_isolation_level(value)) + + def set_read_only(self, value): + self.await_(self._connection.set_read_only(value)) + + def set_deferrable(self, value): + self.await_(self._connection.set_deferrable(value)) + + +class AsyncAdaptFallback_psycopg_connection(AsyncAdapt_psycopg_connection): + __slots__ = () + await_ = staticmethod(await_fallback) + + +class PsycopgAdaptDBAPI: + def __init__(self, psycopg) -> None: + self.psycopg = psycopg + + for k, v in self.psycopg.__dict__.items(): + if k != "connect": + self.__dict__[k] = v + + def connect(self, *arg, **kw): + async_fallback = kw.pop("async_fallback", False) + creator_fn = kw.pop( + "async_creator_fn", self.psycopg.AsyncConnection.connect + ) + if util.asbool(async_fallback): + return AsyncAdaptFallback_psycopg_connection( + await_fallback(creator_fn(*arg, **kw)) + ) + else: + return AsyncAdapt_psycopg_connection( + await_only(creator_fn(*arg, **kw)) + ) + + +class PGDialectAsync_psycopg(PGDialect_psycopg): + is_async = True + supports_statement_cache = True + + @classmethod + def import_dbapi(cls): + import psycopg + from psycopg.pq import ExecStatus + + AsyncAdapt_psycopg_cursor._psycopg_ExecStatus = ExecStatus + + return PsycopgAdaptDBAPI(psycopg) + + @classmethod + def get_pool_class(cls, url): + async_fallback = url.query.get("async_fallback", False) + + if util.asbool(async_fallback): + return pool.FallbackAsyncAdaptedQueuePool + else: + return pool.AsyncAdaptedQueuePool + + def _type_info_fetch(self, connection, name): + from psycopg.types import TypeInfo + + adapted = connection.connection + return adapted.await_(TypeInfo.fetch(adapted.driver_connection, name)) + + def _do_isolation_level(self, connection, autocommit, isolation_level): + connection.set_autocommit(autocommit) + connection.set_isolation_level(isolation_level) + + def _do_autocommit(self, connection, value): + connection.set_autocommit(value) + + def set_readonly(self, connection, value): + connection.set_read_only(value) + + def set_deferrable(self, connection, value): + connection.set_deferrable(value) + + def get_driver_connection(self, connection): + return connection._connection + + +dialect = PGDialect_psycopg +dialect_async = PGDialectAsync_psycopg diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py new file mode 100644 index 0000000..9bf2e49 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2.py @@ -0,0 +1,876 @@ +# dialects/postgresql/psycopg2.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql+psycopg2 + :name: psycopg2 + :dbapi: psycopg2 + :connectstring: postgresql+psycopg2://user:password@host:port/dbname[?key=value&key=value...] + :url: https://pypi.org/project/psycopg2/ + +.. _psycopg2_toplevel: + +psycopg2 Connect Arguments +-------------------------- + +Keyword arguments that are specific to the SQLAlchemy psycopg2 dialect +may be passed to :func:`_sa.create_engine()`, and include the following: + + +* ``isolation_level``: This option, available for all PostgreSQL dialects, + includes the ``AUTOCOMMIT`` isolation level when using the psycopg2 + dialect. This option sets the **default** isolation level for the + connection that is set immediately upon connection to the database before + the connection is pooled. This option is generally superseded by the more + modern :paramref:`_engine.Connection.execution_options.isolation_level` + execution option, detailed at :ref:`dbapi_autocommit`. + + .. seealso:: + + :ref:`psycopg2_isolation_level` + + :ref:`dbapi_autocommit` + + +* ``client_encoding``: sets the client encoding in a libpq-agnostic way, + using psycopg2's ``set_client_encoding()`` method. + + .. seealso:: + + :ref:`psycopg2_unicode` + + +* ``executemany_mode``, ``executemany_batch_page_size``, + ``executemany_values_page_size``: Allows use of psycopg2 + extensions for optimizing "executemany"-style queries. See the referenced + section below for details. + + .. seealso:: + + :ref:`psycopg2_executemany_mode` + +.. tip:: + + The above keyword arguments are **dialect** keyword arguments, meaning + that they are passed as explicit keyword arguments to :func:`_sa.create_engine()`:: + + engine = create_engine( + "postgresql+psycopg2://scott:tiger@localhost/test", + isolation_level="SERIALIZABLE", + ) + + These should not be confused with **DBAPI** connect arguments, which + are passed as part of the :paramref:`_sa.create_engine.connect_args` + dictionary and/or are passed in the URL query string, as detailed in + the section :ref:`custom_dbapi_args`. + +.. _psycopg2_ssl: + +SSL Connections +--------------- + +The psycopg2 module has a connection argument named ``sslmode`` for +controlling its behavior regarding secure (SSL) connections. The default is +``sslmode=prefer``; it will attempt an SSL connection and if that fails it +will fall back to an unencrypted connection. ``sslmode=require`` may be used +to ensure that only secure connections are established. Consult the +psycopg2 / libpq documentation for further options that are available. + +Note that ``sslmode`` is specific to psycopg2 so it is included in the +connection URI:: + + engine = sa.create_engine( + "postgresql+psycopg2://scott:tiger@192.168.0.199:5432/test?sslmode=require" + ) + + +Unix Domain Connections +------------------------ + +psycopg2 supports connecting via Unix domain connections. When the ``host`` +portion of the URL is omitted, SQLAlchemy passes ``None`` to psycopg2, +which specifies Unix-domain communication rather than TCP/IP communication:: + + create_engine("postgresql+psycopg2://user:password@/dbname") + +By default, the socket file used is to connect to a Unix-domain socket +in ``/tmp``, or whatever socket directory was specified when PostgreSQL +was built. This value can be overridden by passing a pathname to psycopg2, +using ``host`` as an additional keyword argument:: + + create_engine("postgresql+psycopg2://user:password@/dbname?host=/var/lib/postgresql") + +.. warning:: The format accepted here allows for a hostname in the main URL + in addition to the "host" query string argument. **When using this URL + format, the initial host is silently ignored**. That is, this URL:: + + engine = create_engine("postgresql+psycopg2://user:password@myhost1/dbname?host=myhost2") + + Above, the hostname ``myhost1`` is **silently ignored and discarded.** The + host which is connected is the ``myhost2`` host. + + This is to maintain some degree of compatibility with PostgreSQL's own URL + format which has been tested to behave the same way and for which tools like + PifPaf hardcode two hostnames. + +.. seealso:: + + `PQconnectdbParams \ + <https://www.postgresql.org/docs/current/static/libpq-connect.html#LIBPQ-PQCONNECTDBPARAMS>`_ + +.. _psycopg2_multi_host: + +Specifying multiple fallback hosts +----------------------------------- + +psycopg2 supports multiple connection points in the connection string. +When the ``host`` parameter is used multiple times in the query section of +the URL, SQLAlchemy will create a single string of the host and port +information provided to make the connections. Tokens may consist of +``host::port`` or just ``host``; in the latter case, the default port +is selected by libpq. In the example below, three host connections +are specified, for ``HostA::PortA``, ``HostB`` connecting to the default port, +and ``HostC::PortC``:: + + create_engine( + "postgresql+psycopg2://user:password@/dbname?host=HostA:PortA&host=HostB&host=HostC:PortC" + ) + +As an alternative, libpq query string format also may be used; this specifies +``host`` and ``port`` as single query string arguments with comma-separated +lists - the default port can be chosen by indicating an empty value +in the comma separated list:: + + create_engine( + "postgresql+psycopg2://user:password@/dbname?host=HostA,HostB,HostC&port=PortA,,PortC" + ) + +With either URL style, connections to each host is attempted based on a +configurable strategy, which may be configured using the libpq +``target_session_attrs`` parameter. Per libpq this defaults to ``any`` +which indicates a connection to each host is then attempted until a connection is successful. +Other strategies include ``primary``, ``prefer-standby``, etc. The complete +list is documented by PostgreSQL at +`libpq connection strings <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNSTRING>`_. + +For example, to indicate two hosts using the ``primary`` strategy:: + + create_engine( + "postgresql+psycopg2://user:password@/dbname?host=HostA:PortA&host=HostB&host=HostC:PortC&target_session_attrs=primary" + ) + +.. versionchanged:: 1.4.40 Port specification in psycopg2 multiple host format + is repaired, previously ports were not correctly interpreted in this context. + libpq comma-separated format is also now supported. + +.. versionadded:: 1.3.20 Support for multiple hosts in PostgreSQL connection + string. + +.. seealso:: + + `libpq connection strings <https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-CONNSTRING>`_ - please refer + to this section in the libpq documentation for complete background on multiple host support. + + +Empty DSN Connections / Environment Variable Connections +--------------------------------------------------------- + +The psycopg2 DBAPI can connect to PostgreSQL by passing an empty DSN to the +libpq client library, which by default indicates to connect to a localhost +PostgreSQL database that is open for "trust" connections. This behavior can be +further tailored using a particular set of environment variables which are +prefixed with ``PG_...``, which are consumed by ``libpq`` to take the place of +any or all elements of the connection string. + +For this form, the URL can be passed without any elements other than the +initial scheme:: + + engine = create_engine('postgresql+psycopg2://') + +In the above form, a blank "dsn" string is passed to the ``psycopg2.connect()`` +function which in turn represents an empty DSN passed to libpq. + +.. versionadded:: 1.3.2 support for parameter-less connections with psycopg2. + +.. seealso:: + + `Environment Variables\ + <https://www.postgresql.org/docs/current/libpq-envars.html>`_ - + PostgreSQL documentation on how to use ``PG_...`` + environment variables for connections. + +.. _psycopg2_execution_options: + +Per-Statement/Connection Execution Options +------------------------------------------- + +The following DBAPI-specific options are respected when used with +:meth:`_engine.Connection.execution_options`, +:meth:`.Executable.execution_options`, +:meth:`_query.Query.execution_options`, +in addition to those not specific to DBAPIs: + +* ``isolation_level`` - Set the transaction isolation level for the lifespan + of a :class:`_engine.Connection` (can only be set on a connection, + not a statement + or query). See :ref:`psycopg2_isolation_level`. + +* ``stream_results`` - Enable or disable usage of psycopg2 server side + cursors - this feature makes use of "named" cursors in combination with + special result handling methods so that result rows are not fully buffered. + Defaults to False, meaning cursors are buffered by default. + +* ``max_row_buffer`` - when using ``stream_results``, an integer value that + specifies the maximum number of rows to buffer at a time. This is + interpreted by the :class:`.BufferedRowCursorResult`, and if omitted the + buffer will grow to ultimately store 1000 rows at a time. + + .. versionchanged:: 1.4 The ``max_row_buffer`` size can now be greater than + 1000, and the buffer will grow to that size. + +.. _psycopg2_batch_mode: + +.. _psycopg2_executemany_mode: + +Psycopg2 Fast Execution Helpers +------------------------------- + +Modern versions of psycopg2 include a feature known as +`Fast Execution Helpers \ +<https://initd.org/psycopg/docs/extras.html#fast-execution-helpers>`_, which +have been shown in benchmarking to improve psycopg2's executemany() +performance, primarily with INSERT statements, by at least +an order of magnitude. + +SQLAlchemy implements a native form of the "insert many values" +handler that will rewrite a single-row INSERT statement to accommodate for +many values at once within an extended VALUES clause; this handler is +equivalent to psycopg2's ``execute_values()`` handler; an overview of this +feature and its configuration are at :ref:`engine_insertmanyvalues`. + +.. versionadded:: 2.0 Replaced psycopg2's ``execute_values()`` fast execution + helper with a native SQLAlchemy mechanism known as + :ref:`insertmanyvalues <engine_insertmanyvalues>`. + +The psycopg2 dialect retains the ability to use the psycopg2-specific +``execute_batch()`` feature, although it is not expected that this is a widely +used feature. The use of this extension may be enabled using the +``executemany_mode`` flag which may be passed to :func:`_sa.create_engine`:: + + engine = create_engine( + "postgresql+psycopg2://scott:tiger@host/dbname", + executemany_mode='values_plus_batch') + + +Possible options for ``executemany_mode`` include: + +* ``values_only`` - this is the default value. SQLAlchemy's native + :ref:`insertmanyvalues <engine_insertmanyvalues>` handler is used for qualifying + INSERT statements, assuming + :paramref:`_sa.create_engine.use_insertmanyvalues` is left at + its default value of ``True``. This handler rewrites simple + INSERT statements to include multiple VALUES clauses so that many + parameter sets can be inserted with one statement. + +* ``'values_plus_batch'``- SQLAlchemy's native + :ref:`insertmanyvalues <engine_insertmanyvalues>` handler is used for qualifying + INSERT statements, assuming + :paramref:`_sa.create_engine.use_insertmanyvalues` is left at its default + value of ``True``. Then, psycopg2's ``execute_batch()`` handler is used for + qualifying UPDATE and DELETE statements when executed with multiple parameter + sets. When using this mode, the :attr:`_engine.CursorResult.rowcount` + attribute will not contain a value for executemany-style executions against + UPDATE and DELETE statements. + +.. versionchanged:: 2.0 Removed the ``'batch'`` and ``'None'`` options + from psycopg2 ``executemany_mode``. Control over batching for INSERT + statements is now configured via the + :paramref:`_sa.create_engine.use_insertmanyvalues` engine-level parameter. + +The term "qualifying statements" refers to the statement being executed +being a Core :func:`_expression.insert`, :func:`_expression.update` +or :func:`_expression.delete` construct, and **not** a plain textual SQL +string or one constructed using :func:`_expression.text`. It also may **not** be +a special "extension" statement such as an "ON CONFLICT" "upsert" statement. +When using the ORM, all insert/update/delete statements used by the ORM flush process +are qualifying. + +The "page size" for the psycopg2 "batch" strategy can be affected +by using the ``executemany_batch_page_size`` parameter, which defaults to +100. + +For the "insertmanyvalues" feature, the page size can be controlled using the +:paramref:`_sa.create_engine.insertmanyvalues_page_size` parameter, +which defaults to 1000. An example of modifying both parameters +is below:: + + engine = create_engine( + "postgresql+psycopg2://scott:tiger@host/dbname", + executemany_mode='values_plus_batch', + insertmanyvalues_page_size=5000, executemany_batch_page_size=500) + +.. seealso:: + + :ref:`engine_insertmanyvalues` - background on "insertmanyvalues" + + :ref:`tutorial_multiple_parameters` - General information on using the + :class:`_engine.Connection` + object to execute statements in such a way as to make + use of the DBAPI ``.executemany()`` method. + + +.. _psycopg2_unicode: + +Unicode with Psycopg2 +---------------------- + +The psycopg2 DBAPI driver supports Unicode data transparently. + +The client character encoding can be controlled for the psycopg2 dialect +in the following ways: + +* For PostgreSQL 9.1 and above, the ``client_encoding`` parameter may be + passed in the database URL; this parameter is consumed by the underlying + ``libpq`` PostgreSQL client library:: + + engine = create_engine("postgresql+psycopg2://user:pass@host/dbname?client_encoding=utf8") + + Alternatively, the above ``client_encoding`` value may be passed using + :paramref:`_sa.create_engine.connect_args` for programmatic establishment with + ``libpq``:: + + engine = create_engine( + "postgresql+psycopg2://user:pass@host/dbname", + connect_args={'client_encoding': 'utf8'} + ) + +* For all PostgreSQL versions, psycopg2 supports a client-side encoding + value that will be passed to database connections when they are first + established. The SQLAlchemy psycopg2 dialect supports this using the + ``client_encoding`` parameter passed to :func:`_sa.create_engine`:: + + engine = create_engine( + "postgresql+psycopg2://user:pass@host/dbname", + client_encoding="utf8" + ) + + .. tip:: The above ``client_encoding`` parameter admittedly is very similar + in appearance to usage of the parameter within the + :paramref:`_sa.create_engine.connect_args` dictionary; the difference + above is that the parameter is consumed by psycopg2 and is + passed to the database connection using ``SET client_encoding TO + 'utf8'``; in the previously mentioned style, the parameter is instead + passed through psycopg2 and consumed by the ``libpq`` library. + +* A common way to set up client encoding with PostgreSQL databases is to + ensure it is configured within the server-side postgresql.conf file; + this is the recommended way to set encoding for a server that is + consistently of one encoding in all databases:: + + # postgresql.conf file + + # client_encoding = sql_ascii # actually, defaults to database + # encoding + client_encoding = utf8 + + + +Transactions +------------ + +The psycopg2 dialect fully supports SAVEPOINT and two-phase commit operations. + +.. _psycopg2_isolation_level: + +Psycopg2 Transaction Isolation Level +------------------------------------- + +As discussed in :ref:`postgresql_isolation_level`, +all PostgreSQL dialects support setting of transaction isolation level +both via the ``isolation_level`` parameter passed to :func:`_sa.create_engine` +, +as well as the ``isolation_level`` argument used by +:meth:`_engine.Connection.execution_options`. When using the psycopg2 dialect +, these +options make use of psycopg2's ``set_isolation_level()`` connection method, +rather than emitting a PostgreSQL directive; this is because psycopg2's +API-level setting is always emitted at the start of each transaction in any +case. + +The psycopg2 dialect supports these constants for isolation level: + +* ``READ COMMITTED`` +* ``READ UNCOMMITTED`` +* ``REPEATABLE READ`` +* ``SERIALIZABLE`` +* ``AUTOCOMMIT`` + +.. seealso:: + + :ref:`postgresql_isolation_level` + + :ref:`pg8000_isolation_level` + + +NOTICE logging +--------------- + +The psycopg2 dialect will log PostgreSQL NOTICE messages +via the ``sqlalchemy.dialects.postgresql`` logger. When this logger +is set to the ``logging.INFO`` level, notice messages will be logged:: + + import logging + + logging.getLogger('sqlalchemy.dialects.postgresql').setLevel(logging.INFO) + +Above, it is assumed that logging is configured externally. If this is not +the case, configuration such as ``logging.basicConfig()`` must be utilized:: + + import logging + + logging.basicConfig() # log messages to stdout + logging.getLogger('sqlalchemy.dialects.postgresql').setLevel(logging.INFO) + +.. seealso:: + + `Logging HOWTO <https://docs.python.org/3/howto/logging.html>`_ - on the python.org website + +.. _psycopg2_hstore: + +HSTORE type +------------ + +The ``psycopg2`` DBAPI includes an extension to natively handle marshalling of +the HSTORE type. The SQLAlchemy psycopg2 dialect will enable this extension +by default when psycopg2 version 2.4 or greater is used, and +it is detected that the target database has the HSTORE type set up for use. +In other words, when the dialect makes the first +connection, a sequence like the following is performed: + +1. Request the available HSTORE oids using + ``psycopg2.extras.HstoreAdapter.get_oids()``. + If this function returns a list of HSTORE identifiers, we then determine + that the ``HSTORE`` extension is present. + This function is **skipped** if the version of psycopg2 installed is + less than version 2.4. + +2. If the ``use_native_hstore`` flag is at its default of ``True``, and + we've detected that ``HSTORE`` oids are available, the + ``psycopg2.extensions.register_hstore()`` extension is invoked for all + connections. + +The ``register_hstore()`` extension has the effect of **all Python +dictionaries being accepted as parameters regardless of the type of target +column in SQL**. The dictionaries are converted by this extension into a +textual HSTORE expression. If this behavior is not desired, disable the +use of the hstore extension by setting ``use_native_hstore`` to ``False`` as +follows:: + + engine = create_engine("postgresql+psycopg2://scott:tiger@localhost/test", + use_native_hstore=False) + +The ``HSTORE`` type is **still supported** when the +``psycopg2.extensions.register_hstore()`` extension is not used. It merely +means that the coercion between Python dictionaries and the HSTORE +string format, on both the parameter side and the result side, will take +place within SQLAlchemy's own marshalling logic, and not that of ``psycopg2`` +which may be more performant. + +""" # noqa +from __future__ import annotations + +import collections.abc as collections_abc +import logging +import re +from typing import cast + +from . import ranges +from ._psycopg_common import _PGDialect_common_psycopg +from ._psycopg_common import _PGExecutionContext_common_psycopg +from .base import PGIdentifierPreparer +from .json import JSON +from .json import JSONB +from ... import types as sqltypes +from ... import util +from ...util import FastIntFlag +from ...util import parse_user_argument_for_enum + +logger = logging.getLogger("sqlalchemy.dialects.postgresql") + + +class _PGJSON(JSON): + def result_processor(self, dialect, coltype): + return None + + +class _PGJSONB(JSONB): + def result_processor(self, dialect, coltype): + return None + + +class _Psycopg2Range(ranges.AbstractSingleRangeImpl): + _psycopg2_range_cls = "none" + + def bind_processor(self, dialect): + psycopg2_Range = getattr( + cast(PGDialect_psycopg2, dialect)._psycopg2_extras, + self._psycopg2_range_cls, + ) + + def to_range(value): + if isinstance(value, ranges.Range): + value = psycopg2_Range( + value.lower, value.upper, value.bounds, value.empty + ) + return value + + return to_range + + def result_processor(self, dialect, coltype): + def to_range(value): + if value is not None: + value = ranges.Range( + value._lower, + value._upper, + bounds=value._bounds if value._bounds else "[)", + empty=not value._bounds, + ) + return value + + return to_range + + +class _Psycopg2NumericRange(_Psycopg2Range): + _psycopg2_range_cls = "NumericRange" + + +class _Psycopg2DateRange(_Psycopg2Range): + _psycopg2_range_cls = "DateRange" + + +class _Psycopg2DateTimeRange(_Psycopg2Range): + _psycopg2_range_cls = "DateTimeRange" + + +class _Psycopg2DateTimeTZRange(_Psycopg2Range): + _psycopg2_range_cls = "DateTimeTZRange" + + +class PGExecutionContext_psycopg2(_PGExecutionContext_common_psycopg): + _psycopg2_fetched_rows = None + + def post_exec(self): + self._log_notices(self.cursor) + + def _log_notices(self, cursor): + # check also that notices is an iterable, after it's already + # established that we will be iterating through it. This is to get + # around test suites such as SQLAlchemy's using a Mock object for + # cursor + if not cursor.connection.notices or not isinstance( + cursor.connection.notices, collections_abc.Iterable + ): + return + + for notice in cursor.connection.notices: + # NOTICE messages have a + # newline character at the end + logger.info(notice.rstrip()) + + cursor.connection.notices[:] = [] + + +class PGIdentifierPreparer_psycopg2(PGIdentifierPreparer): + pass + + +class ExecutemanyMode(FastIntFlag): + EXECUTEMANY_VALUES = 0 + EXECUTEMANY_VALUES_PLUS_BATCH = 1 + + +( + EXECUTEMANY_VALUES, + EXECUTEMANY_VALUES_PLUS_BATCH, +) = ExecutemanyMode.__members__.values() + + +class PGDialect_psycopg2(_PGDialect_common_psycopg): + driver = "psycopg2" + + supports_statement_cache = True + supports_server_side_cursors = True + + default_paramstyle = "pyformat" + # set to true based on psycopg2 version + supports_sane_multi_rowcount = False + execution_ctx_cls = PGExecutionContext_psycopg2 + preparer = PGIdentifierPreparer_psycopg2 + psycopg2_version = (0, 0) + use_insertmanyvalues_wo_returning = True + + returns_native_bytes = False + + _has_native_hstore = True + + colspecs = util.update_copy( + _PGDialect_common_psycopg.colspecs, + { + JSON: _PGJSON, + sqltypes.JSON: _PGJSON, + JSONB: _PGJSONB, + ranges.INT4RANGE: _Psycopg2NumericRange, + ranges.INT8RANGE: _Psycopg2NumericRange, + ranges.NUMRANGE: _Psycopg2NumericRange, + ranges.DATERANGE: _Psycopg2DateRange, + ranges.TSRANGE: _Psycopg2DateTimeRange, + ranges.TSTZRANGE: _Psycopg2DateTimeTZRange, + }, + ) + + def __init__( + self, + executemany_mode="values_only", + executemany_batch_page_size=100, + **kwargs, + ): + _PGDialect_common_psycopg.__init__(self, **kwargs) + + if self._native_inet_types: + raise NotImplementedError( + "The psycopg2 dialect does not implement " + "ipaddress type handling; native_inet_types cannot be set " + "to ``True`` when using this dialect." + ) + + # Parse executemany_mode argument, allowing it to be only one of the + # symbol names + self.executemany_mode = parse_user_argument_for_enum( + executemany_mode, + { + EXECUTEMANY_VALUES: ["values_only"], + EXECUTEMANY_VALUES_PLUS_BATCH: ["values_plus_batch"], + }, + "executemany_mode", + ) + + self.executemany_batch_page_size = executemany_batch_page_size + + if self.dbapi and hasattr(self.dbapi, "__version__"): + m = re.match(r"(\d+)\.(\d+)(?:\.(\d+))?", self.dbapi.__version__) + if m: + self.psycopg2_version = tuple( + int(x) for x in m.group(1, 2, 3) if x is not None + ) + + if self.psycopg2_version < (2, 7): + raise ImportError( + "psycopg2 version 2.7 or higher is required." + ) + + def initialize(self, connection): + super().initialize(connection) + self._has_native_hstore = ( + self.use_native_hstore + and self._hstore_oids(connection.connection.dbapi_connection) + is not None + ) + + self.supports_sane_multi_rowcount = ( + self.executemany_mode is not EXECUTEMANY_VALUES_PLUS_BATCH + ) + + @classmethod + def import_dbapi(cls): + import psycopg2 + + return psycopg2 + + @util.memoized_property + def _psycopg2_extensions(cls): + from psycopg2 import extensions + + return extensions + + @util.memoized_property + def _psycopg2_extras(cls): + from psycopg2 import extras + + return extras + + @util.memoized_property + def _isolation_lookup(self): + extensions = self._psycopg2_extensions + return { + "AUTOCOMMIT": extensions.ISOLATION_LEVEL_AUTOCOMMIT, + "READ COMMITTED": extensions.ISOLATION_LEVEL_READ_COMMITTED, + "READ UNCOMMITTED": extensions.ISOLATION_LEVEL_READ_UNCOMMITTED, + "REPEATABLE READ": extensions.ISOLATION_LEVEL_REPEATABLE_READ, + "SERIALIZABLE": extensions.ISOLATION_LEVEL_SERIALIZABLE, + } + + def set_isolation_level(self, dbapi_connection, level): + dbapi_connection.set_isolation_level(self._isolation_lookup[level]) + + def set_readonly(self, connection, value): + connection.readonly = value + + def get_readonly(self, connection): + return connection.readonly + + def set_deferrable(self, connection, value): + connection.deferrable = value + + def get_deferrable(self, connection): + return connection.deferrable + + def on_connect(self): + extras = self._psycopg2_extras + + fns = [] + if self.client_encoding is not None: + + def on_connect(dbapi_conn): + dbapi_conn.set_client_encoding(self.client_encoding) + + fns.append(on_connect) + + if self.dbapi: + + def on_connect(dbapi_conn): + extras.register_uuid(None, dbapi_conn) + + fns.append(on_connect) + + if self.dbapi and self.use_native_hstore: + + def on_connect(dbapi_conn): + hstore_oids = self._hstore_oids(dbapi_conn) + if hstore_oids is not None: + oid, array_oid = hstore_oids + kw = {"oid": oid} + kw["array_oid"] = array_oid + extras.register_hstore(dbapi_conn, **kw) + + fns.append(on_connect) + + if self.dbapi and self._json_deserializer: + + def on_connect(dbapi_conn): + extras.register_default_json( + dbapi_conn, loads=self._json_deserializer + ) + extras.register_default_jsonb( + dbapi_conn, loads=self._json_deserializer + ) + + fns.append(on_connect) + + if fns: + + def on_connect(dbapi_conn): + for fn in fns: + fn(dbapi_conn) + + return on_connect + else: + return None + + def do_executemany(self, cursor, statement, parameters, context=None): + if self.executemany_mode is EXECUTEMANY_VALUES_PLUS_BATCH: + if self.executemany_batch_page_size: + kwargs = {"page_size": self.executemany_batch_page_size} + else: + kwargs = {} + self._psycopg2_extras.execute_batch( + cursor, statement, parameters, **kwargs + ) + else: + cursor.executemany(statement, parameters) + + def do_begin_twophase(self, connection, xid): + connection.connection.tpc_begin(xid) + + def do_prepare_twophase(self, connection, xid): + connection.connection.tpc_prepare() + + def _do_twophase(self, dbapi_conn, operation, xid, recover=False): + if recover: + if dbapi_conn.status != self._psycopg2_extensions.STATUS_READY: + dbapi_conn.rollback() + operation(xid) + else: + operation() + + def do_rollback_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + dbapi_conn = connection.connection.dbapi_connection + self._do_twophase( + dbapi_conn, dbapi_conn.tpc_rollback, xid, recover=recover + ) + + def do_commit_twophase( + self, connection, xid, is_prepared=True, recover=False + ): + dbapi_conn = connection.connection.dbapi_connection + self._do_twophase( + dbapi_conn, dbapi_conn.tpc_commit, xid, recover=recover + ) + + @util.memoized_instancemethod + def _hstore_oids(self, dbapi_connection): + extras = self._psycopg2_extras + oids = extras.HstoreAdapter.get_oids(dbapi_connection) + if oids is not None and oids[0]: + return oids[0:2] + else: + return None + + def is_disconnect(self, e, connection, cursor): + if isinstance(e, self.dbapi.Error): + # check the "closed" flag. this might not be + # present on old psycopg2 versions. Also, + # this flag doesn't actually help in a lot of disconnect + # situations, so don't rely on it. + if getattr(connection, "closed", False): + return True + + # checks based on strings. in the case that .closed + # didn't cut it, fall back onto these. + str_e = str(e).partition("\n")[0] + for msg in [ + # these error messages from libpq: interfaces/libpq/fe-misc.c + # and interfaces/libpq/fe-secure.c. + "terminating connection", + "closed the connection", + "connection not open", + "could not receive data from server", + "could not send data to server", + # psycopg2 client errors, psycopg2/connection.h, + # psycopg2/cursor.h + "connection already closed", + "cursor already closed", + # not sure where this path is originally from, it may + # be obsolete. It really says "losed", not "closed". + "losed the connection unexpectedly", + # these can occur in newer SSL + "connection has been closed unexpectedly", + "SSL error: decryption failed or bad record mac", + "SSL SYSCALL error: Bad file descriptor", + "SSL SYSCALL error: EOF detected", + "SSL SYSCALL error: Operation timed out", + "SSL SYSCALL error: Bad address", + ]: + idx = str_e.find(msg) + if idx >= 0 and '"' not in str_e[:idx]: + return True + return False + + +dialect = PGDialect_psycopg2 diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py new file mode 100644 index 0000000..3cc3b69 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/psycopg2cffi.py @@ -0,0 +1,61 @@ +# dialects/postgresql/psycopg2cffi.py +# Copyright (C) 2005-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +# mypy: ignore-errors + +r""" +.. dialect:: postgresql+psycopg2cffi + :name: psycopg2cffi + :dbapi: psycopg2cffi + :connectstring: postgresql+psycopg2cffi://user:password@host:port/dbname[?key=value&key=value...] + :url: https://pypi.org/project/psycopg2cffi/ + +``psycopg2cffi`` is an adaptation of ``psycopg2``, using CFFI for the C +layer. This makes it suitable for use in e.g. PyPy. Documentation +is as per ``psycopg2``. + +.. seealso:: + + :mod:`sqlalchemy.dialects.postgresql.psycopg2` + +""" # noqa +from .psycopg2 import PGDialect_psycopg2 +from ... import util + + +class PGDialect_psycopg2cffi(PGDialect_psycopg2): + driver = "psycopg2cffi" + supports_unicode_statements = True + supports_statement_cache = True + + # psycopg2cffi's first release is 2.5.0, but reports + # __version__ as 2.4.4. Subsequent releases seem to have + # fixed this. + + FEATURE_VERSION_MAP = dict( + native_json=(2, 4, 4), + native_jsonb=(2, 7, 1), + sane_multi_rowcount=(2, 4, 4), + array_oid=(2, 4, 4), + hstore_adapter=(2, 4, 4), + ) + + @classmethod + def import_dbapi(cls): + return __import__("psycopg2cffi") + + @util.memoized_property + def _psycopg2_extensions(cls): + root = __import__("psycopg2cffi", fromlist=["extensions"]) + return root.extensions + + @util.memoized_property + def _psycopg2_extras(cls): + root = __import__("psycopg2cffi", fromlist=["extras"]) + return root.extras + + +dialect = PGDialect_psycopg2cffi diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ranges.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ranges.py new file mode 100644 index 0000000..b793ca4 --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/ranges.py @@ -0,0 +1,1029 @@ +# dialects/postgresql/ranges.py +# Copyright (C) 2013-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php + +from __future__ import annotations + +import dataclasses +from datetime import date +from datetime import datetime +from datetime import timedelta +from decimal import Decimal +from typing import Any +from typing import cast +from typing import Generic +from typing import List +from typing import Optional +from typing import overload +from typing import Sequence +from typing import Tuple +from typing import Type +from typing import TYPE_CHECKING +from typing import TypeVar +from typing import Union + +from .operators import ADJACENT_TO +from .operators import CONTAINED_BY +from .operators import CONTAINS +from .operators import NOT_EXTEND_LEFT_OF +from .operators import NOT_EXTEND_RIGHT_OF +from .operators import OVERLAP +from .operators import STRICTLY_LEFT_OF +from .operators import STRICTLY_RIGHT_OF +from ... import types as sqltypes +from ...sql import operators +from ...sql.type_api import TypeEngine +from ...util import py310 +from ...util.typing import Literal + +if TYPE_CHECKING: + from ...sql.elements import ColumnElement + from ...sql.type_api import _TE + from ...sql.type_api import TypeEngineMixin + +_T = TypeVar("_T", bound=Any) + +_BoundsType = Literal["()", "[)", "(]", "[]"] + +if py310: + dc_slots = {"slots": True} + dc_kwonly = {"kw_only": True} +else: + dc_slots = {} + dc_kwonly = {} + + +@dataclasses.dataclass(frozen=True, **dc_slots) +class Range(Generic[_T]): + """Represent a PostgreSQL range. + + E.g.:: + + r = Range(10, 50, bounds="()") + + The calling style is similar to that of psycopg and psycopg2, in part + to allow easier migration from previous SQLAlchemy versions that used + these objects directly. + + :param lower: Lower bound value, or None + :param upper: Upper bound value, or None + :param bounds: keyword-only, optional string value that is one of + ``"()"``, ``"[)"``, ``"(]"``, ``"[]"``. Defaults to ``"[)"``. + :param empty: keyword-only, optional bool indicating this is an "empty" + range + + .. versionadded:: 2.0 + + """ + + lower: Optional[_T] = None + """the lower bound""" + + upper: Optional[_T] = None + """the upper bound""" + + if TYPE_CHECKING: + bounds: _BoundsType = dataclasses.field(default="[)") + empty: bool = dataclasses.field(default=False) + else: + bounds: _BoundsType = dataclasses.field(default="[)", **dc_kwonly) + empty: bool = dataclasses.field(default=False, **dc_kwonly) + + if not py310: + + def __init__( + self, + lower: Optional[_T] = None, + upper: Optional[_T] = None, + *, + bounds: _BoundsType = "[)", + empty: bool = False, + ): + # no __slots__ either so we can update dict + self.__dict__.update( + { + "lower": lower, + "upper": upper, + "bounds": bounds, + "empty": empty, + } + ) + + def __bool__(self) -> bool: + return not self.empty + + @property + def isempty(self) -> bool: + "A synonym for the 'empty' attribute." + + return self.empty + + @property + def is_empty(self) -> bool: + "A synonym for the 'empty' attribute." + + return self.empty + + @property + def lower_inc(self) -> bool: + """Return True if the lower bound is inclusive.""" + + return self.bounds[0] == "[" + + @property + def lower_inf(self) -> bool: + """Return True if this range is non-empty and lower bound is + infinite.""" + + return not self.empty and self.lower is None + + @property + def upper_inc(self) -> bool: + """Return True if the upper bound is inclusive.""" + + return self.bounds[1] == "]" + + @property + def upper_inf(self) -> bool: + """Return True if this range is non-empty and the upper bound is + infinite.""" + + return not self.empty and self.upper is None + + @property + def __sa_type_engine__(self) -> AbstractSingleRange[_T]: + return AbstractSingleRange() + + def _contains_value(self, value: _T) -> bool: + """Return True if this range contains the given value.""" + + if self.empty: + return False + + if self.lower is None: + return self.upper is None or ( + value < self.upper + if self.bounds[1] == ")" + else value <= self.upper + ) + + if self.upper is None: + return ( # type: ignore + value > self.lower + if self.bounds[0] == "(" + else value >= self.lower + ) + + return ( # type: ignore + value > self.lower + if self.bounds[0] == "(" + else value >= self.lower + ) and ( + value < self.upper + if self.bounds[1] == ")" + else value <= self.upper + ) + + def _get_discrete_step(self) -> Any: + "Determine the “step” for this range, if it is a discrete one." + + # See + # https://www.postgresql.org/docs/current/rangetypes.html#RANGETYPES-DISCRETE + # for the rationale + + if isinstance(self.lower, int) or isinstance(self.upper, int): + return 1 + elif isinstance(self.lower, datetime) or isinstance( + self.upper, datetime + ): + # This is required, because a `isinstance(datetime.now(), date)` + # is True + return None + elif isinstance(self.lower, date) or isinstance(self.upper, date): + return timedelta(days=1) + else: + return None + + def _compare_edges( + self, + value1: Optional[_T], + bound1: str, + value2: Optional[_T], + bound2: str, + only_values: bool = False, + ) -> int: + """Compare two range bounds. + + Return -1, 0 or 1 respectively when `value1` is less than, + equal to or greater than `value2`. + + When `only_value` is ``True``, do not consider the *inclusivity* + of the edges, just their values. + """ + + value1_is_lower_bound = bound1 in {"[", "("} + value2_is_lower_bound = bound2 in {"[", "("} + + # Infinite edges are equal when they are on the same side, + # otherwise a lower edge is considered less than the upper end + if value1 is value2 is None: + if value1_is_lower_bound == value2_is_lower_bound: + return 0 + else: + return -1 if value1_is_lower_bound else 1 + elif value1 is None: + return -1 if value1_is_lower_bound else 1 + elif value2 is None: + return 1 if value2_is_lower_bound else -1 + + # Short path for trivial case + if bound1 == bound2 and value1 == value2: + return 0 + + value1_inc = bound1 in {"[", "]"} + value2_inc = bound2 in {"[", "]"} + step = self._get_discrete_step() + + if step is not None: + # "Normalize" the two edges as '[)', to simplify successive + # logic when the range is discrete: otherwise we would need + # to handle the comparison between ``(0`` and ``[1`` that + # are equal when dealing with integers while for floats the + # former is lesser than the latter + + if value1_is_lower_bound: + if not value1_inc: + value1 += step + value1_inc = True + else: + if value1_inc: + value1 += step + value1_inc = False + if value2_is_lower_bound: + if not value2_inc: + value2 += step + value2_inc = True + else: + if value2_inc: + value2 += step + value2_inc = False + + if value1 < value2: # type: ignore + return -1 + elif value1 > value2: # type: ignore + return 1 + elif only_values: + return 0 + else: + # Neither one is infinite but are equal, so we + # need to consider the respective inclusive/exclusive + # flag + + if value1_inc and value2_inc: + return 0 + elif not value1_inc and not value2_inc: + if value1_is_lower_bound == value2_is_lower_bound: + return 0 + else: + return 1 if value1_is_lower_bound else -1 + elif not value1_inc: + return 1 if value1_is_lower_bound else -1 + elif not value2_inc: + return -1 if value2_is_lower_bound else 1 + else: + return 0 + + def __eq__(self, other: Any) -> bool: + """Compare this range to the `other` taking into account + bounds inclusivity, returning ``True`` if they are equal. + """ + + if not isinstance(other, Range): + return NotImplemented + + if self.empty and other.empty: + return True + elif self.empty != other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + olower = other.lower + olower_b = other.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + oupper = other.upper + oupper_b = other.bounds[1] + + return ( + self._compare_edges(slower, slower_b, olower, olower_b) == 0 + and self._compare_edges(supper, supper_b, oupper, oupper_b) == 0 + ) + + def contained_by(self, other: Range[_T]) -> bool: + "Determine whether this range is a contained by `other`." + + # Any range contains the empty one + if self.empty: + return True + + # An empty range does not contain any range except the empty one + if other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + olower = other.lower + olower_b = other.bounds[0] + + if self._compare_edges(slower, slower_b, olower, olower_b) < 0: + return False + + supper = self.upper + supper_b = self.bounds[1] + oupper = other.upper + oupper_b = other.bounds[1] + + if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0: + return False + + return True + + def contains(self, value: Union[_T, Range[_T]]) -> bool: + "Determine whether this range contains `value`." + + if isinstance(value, Range): + return value.contained_by(self) + else: + return self._contains_value(value) + + def overlaps(self, other: Range[_T]) -> bool: + "Determine whether this range overlaps with `other`." + + # Empty ranges never overlap with any other range + if self.empty or other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + # Check whether this lower bound is contained in the other range + if ( + self._compare_edges(slower, slower_b, olower, olower_b) >= 0 + and self._compare_edges(slower, slower_b, oupper, oupper_b) <= 0 + ): + return True + + # Check whether other lower bound is contained in this range + if ( + self._compare_edges(olower, olower_b, slower, slower_b) >= 0 + and self._compare_edges(olower, olower_b, supper, supper_b) <= 0 + ): + return True + + return False + + def strictly_left_of(self, other: Range[_T]) -> bool: + "Determine whether this range is completely to the left of `other`." + + # Empty ranges are neither to left nor to the right of any other range + if self.empty or other.empty: + return False + + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + + # Check whether this upper edge is less than other's lower end + return self._compare_edges(supper, supper_b, olower, olower_b) < 0 + + __lshift__ = strictly_left_of + + def strictly_right_of(self, other: Range[_T]) -> bool: + "Determine whether this range is completely to the right of `other`." + + # Empty ranges are neither to left nor to the right of any other range + if self.empty or other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + # Check whether this lower edge is greater than other's upper end + return self._compare_edges(slower, slower_b, oupper, oupper_b) > 0 + + __rshift__ = strictly_right_of + + def not_extend_left_of(self, other: Range[_T]) -> bool: + "Determine whether this does not extend to the left of `other`." + + # Empty ranges are neither to left nor to the right of any other range + if self.empty or other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + olower = other.lower + olower_b = other.bounds[0] + + # Check whether this lower edge is not less than other's lower end + return self._compare_edges(slower, slower_b, olower, olower_b) >= 0 + + def not_extend_right_of(self, other: Range[_T]) -> bool: + "Determine whether this does not extend to the right of `other`." + + # Empty ranges are neither to left nor to the right of any other range + if self.empty or other.empty: + return False + + supper = self.upper + supper_b = self.bounds[1] + oupper = other.upper + oupper_b = other.bounds[1] + + # Check whether this upper edge is not greater than other's upper end + return self._compare_edges(supper, supper_b, oupper, oupper_b) <= 0 + + def _upper_edge_adjacent_to_lower( + self, + value1: Optional[_T], + bound1: str, + value2: Optional[_T], + bound2: str, + ) -> bool: + """Determine whether an upper bound is immediately successive to a + lower bound.""" + + # Since we need a peculiar way to handle the bounds inclusivity, + # just do a comparison by value here + res = self._compare_edges(value1, bound1, value2, bound2, True) + if res == -1: + step = self._get_discrete_step() + if step is None: + return False + if bound1 == "]": + if bound2 == "[": + return value1 == value2 - step # type: ignore + else: + return value1 == value2 + else: + if bound2 == "[": + return value1 == value2 + else: + return value1 == value2 - step # type: ignore + elif res == 0: + # Cover cases like [0,0] -|- [1,] and [0,2) -|- (1,3] + if ( + bound1 == "]" + and bound2 == "[" + or bound1 == ")" + and bound2 == "(" + ): + step = self._get_discrete_step() + if step is not None: + return True + return ( + bound1 == ")" + and bound2 == "[" + or bound1 == "]" + and bound2 == "(" + ) + else: + return False + + def adjacent_to(self, other: Range[_T]) -> bool: + "Determine whether this range is adjacent to the `other`." + + # Empty ranges are not adjacent to any other range + if self.empty or other.empty: + return False + + slower = self.lower + slower_b = self.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + return self._upper_edge_adjacent_to_lower( + supper, supper_b, olower, olower_b + ) or self._upper_edge_adjacent_to_lower( + oupper, oupper_b, slower, slower_b + ) + + def union(self, other: Range[_T]) -> Range[_T]: + """Compute the union of this range with the `other`. + + This raises a ``ValueError`` exception if the two ranges are + "disjunct", that is neither adjacent nor overlapping. + """ + + # Empty ranges are "additive identities" + if self.empty: + return other + if other.empty: + return self + + if not self.overlaps(other) and not self.adjacent_to(other): + raise ValueError( + "Adding non-overlapping and non-adjacent" + " ranges is not implemented" + ) + + slower = self.lower + slower_b = self.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + if self._compare_edges(slower, slower_b, olower, olower_b) < 0: + rlower = slower + rlower_b = slower_b + else: + rlower = olower + rlower_b = olower_b + + if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0: + rupper = supper + rupper_b = supper_b + else: + rupper = oupper + rupper_b = oupper_b + + return Range( + rlower, rupper, bounds=cast(_BoundsType, rlower_b + rupper_b) + ) + + def __add__(self, other: Range[_T]) -> Range[_T]: + return self.union(other) + + def difference(self, other: Range[_T]) -> Range[_T]: + """Compute the difference between this range and the `other`. + + This raises a ``ValueError`` exception if the two ranges are + "disjunct", that is neither adjacent nor overlapping. + """ + + # Subtracting an empty range is a no-op + if self.empty or other.empty: + return self + + slower = self.lower + slower_b = self.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + sl_vs_ol = self._compare_edges(slower, slower_b, olower, olower_b) + su_vs_ou = self._compare_edges(supper, supper_b, oupper, oupper_b) + if sl_vs_ol < 0 and su_vs_ou > 0: + raise ValueError( + "Subtracting a strictly inner range is not implemented" + ) + + sl_vs_ou = self._compare_edges(slower, slower_b, oupper, oupper_b) + su_vs_ol = self._compare_edges(supper, supper_b, olower, olower_b) + + # If the ranges do not overlap, result is simply the first + if sl_vs_ou > 0 or su_vs_ol < 0: + return self + + # If this range is completely contained by the other, result is empty + if sl_vs_ol >= 0 and su_vs_ou <= 0: + return Range(None, None, empty=True) + + # If this range extends to the left of the other and ends in its + # middle + if sl_vs_ol <= 0 and su_vs_ol >= 0 and su_vs_ou <= 0: + rupper_b = ")" if olower_b == "[" else "]" + if ( + slower_b != "[" + and rupper_b != "]" + and self._compare_edges(slower, slower_b, olower, rupper_b) + == 0 + ): + return Range(None, None, empty=True) + else: + return Range( + slower, + olower, + bounds=cast(_BoundsType, slower_b + rupper_b), + ) + + # If this range starts in the middle of the other and extends to its + # right + if sl_vs_ol >= 0 and su_vs_ou >= 0 and sl_vs_ou <= 0: + rlower_b = "(" if oupper_b == "]" else "[" + if ( + rlower_b != "[" + and supper_b != "]" + and self._compare_edges(oupper, rlower_b, supper, supper_b) + == 0 + ): + return Range(None, None, empty=True) + else: + return Range( + oupper, + supper, + bounds=cast(_BoundsType, rlower_b + supper_b), + ) + + assert False, f"Unhandled case computing {self} - {other}" + + def __sub__(self, other: Range[_T]) -> Range[_T]: + return self.difference(other) + + def intersection(self, other: Range[_T]) -> Range[_T]: + """Compute the intersection of this range with the `other`. + + .. versionadded:: 2.0.10 + + """ + if self.empty or other.empty or not self.overlaps(other): + return Range(None, None, empty=True) + + slower = self.lower + slower_b = self.bounds[0] + supper = self.upper + supper_b = self.bounds[1] + olower = other.lower + olower_b = other.bounds[0] + oupper = other.upper + oupper_b = other.bounds[1] + + if self._compare_edges(slower, slower_b, olower, olower_b) < 0: + rlower = olower + rlower_b = olower_b + else: + rlower = slower + rlower_b = slower_b + + if self._compare_edges(supper, supper_b, oupper, oupper_b) > 0: + rupper = oupper + rupper_b = oupper_b + else: + rupper = supper + rupper_b = supper_b + + return Range( + rlower, + rupper, + bounds=cast(_BoundsType, rlower_b + rupper_b), + ) + + def __mul__(self, other: Range[_T]) -> Range[_T]: + return self.intersection(other) + + def __str__(self) -> str: + return self._stringify() + + def _stringify(self) -> str: + if self.empty: + return "empty" + + l, r = self.lower, self.upper + l = "" if l is None else l # type: ignore + r = "" if r is None else r # type: ignore + + b0, b1 = cast("Tuple[str, str]", self.bounds) + + return f"{b0}{l},{r}{b1}" + + +class MultiRange(List[Range[_T]]): + """Represents a multirange sequence. + + This list subclass is an utility to allow automatic type inference of + the proper multi-range SQL type depending on the single range values. + This is useful when operating on literal multi-ranges:: + + import sqlalchemy as sa + from sqlalchemy.dialects.postgresql import MultiRange, Range + + value = literal(MultiRange([Range(2, 4)])) + + select(tbl).where(tbl.c.value.op("@")(MultiRange([Range(-3, 7)]))) + + .. versionadded:: 2.0.26 + + .. seealso:: + + - :ref:`postgresql_multirange_list_use`. + """ + + @property + def __sa_type_engine__(self) -> AbstractMultiRange[_T]: + return AbstractMultiRange() + + +class AbstractRange(sqltypes.TypeEngine[_T]): + """Base class for single and multi Range SQL types.""" + + render_bind_cast = True + + __abstract__ = True + + @overload + def adapt(self, cls: Type[_TE], **kw: Any) -> _TE: ... + + @overload + def adapt( + self, cls: Type[TypeEngineMixin], **kw: Any + ) -> TypeEngine[Any]: ... + + def adapt( + self, + cls: Type[Union[TypeEngine[Any], TypeEngineMixin]], + **kw: Any, + ) -> TypeEngine[Any]: + """Dynamically adapt a range type to an abstract impl. + + For example ``INT4RANGE().adapt(_Psycopg2NumericRange)`` should + produce a type that will have ``_Psycopg2NumericRange`` behaviors + and also render as ``INT4RANGE`` in SQL and DDL. + + """ + if ( + issubclass(cls, (AbstractSingleRangeImpl, AbstractMultiRangeImpl)) + and cls is not self.__class__ + ): + # two ways to do this are: 1. create a new type on the fly + # or 2. have AbstractRangeImpl(visit_name) constructor and a + # visit_abstract_range_impl() method in the PG compiler. + # I'm choosing #1 as the resulting type object + # will then make use of the same mechanics + # as if we had made all these sub-types explicitly, and will + # also look more obvious under pdb etc. + # The adapt() operation here is cached per type-class-per-dialect, + # so is not much of a performance concern + visit_name = self.__visit_name__ + return type( # type: ignore + f"{visit_name}RangeImpl", + (cls, self.__class__), + {"__visit_name__": visit_name}, + )() + else: + return super().adapt(cls) + + class comparator_factory(TypeEngine.Comparator[Range[Any]]): + """Define comparison operations for range types.""" + + def contains(self, other: Any, **kw: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the right hand operand, + which can be an element or a range, is contained within the + column. + + kwargs may be ignored by this operator but are required for API + conformance. + """ + return self.expr.operate(CONTAINS, other) + + def contained_by(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the column is contained + within the right hand operand. + """ + return self.expr.operate(CONTAINED_BY, other) + + def overlaps(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the column overlaps + (has points in common with) the right hand operand. + """ + return self.expr.operate(OVERLAP, other) + + def strictly_left_of(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the column is strictly + left of the right hand operand. + """ + return self.expr.operate(STRICTLY_LEFT_OF, other) + + __lshift__ = strictly_left_of + + def strictly_right_of(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the column is strictly + right of the right hand operand. + """ + return self.expr.operate(STRICTLY_RIGHT_OF, other) + + __rshift__ = strictly_right_of + + def not_extend_right_of(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the range in the column + does not extend right of the range in the operand. + """ + return self.expr.operate(NOT_EXTEND_RIGHT_OF, other) + + def not_extend_left_of(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the range in the column + does not extend left of the range in the operand. + """ + return self.expr.operate(NOT_EXTEND_LEFT_OF, other) + + def adjacent_to(self, other: Any) -> ColumnElement[bool]: + """Boolean expression. Returns true if the range in the column + is adjacent to the range in the operand. + """ + return self.expr.operate(ADJACENT_TO, other) + + def union(self, other: Any) -> ColumnElement[bool]: + """Range expression. Returns the union of the two ranges. + Will raise an exception if the resulting range is not + contiguous. + """ + return self.expr.operate(operators.add, other) + + def difference(self, other: Any) -> ColumnElement[bool]: + """Range expression. Returns the union of the two ranges. + Will raise an exception if the resulting range is not + contiguous. + """ + return self.expr.operate(operators.sub, other) + + def intersection(self, other: Any) -> ColumnElement[Range[_T]]: + """Range expression. Returns the intersection of the two ranges. + Will raise an exception if the resulting range is not + contiguous. + """ + return self.expr.operate(operators.mul, other) + + +class AbstractSingleRange(AbstractRange[Range[_T]]): + """Base for PostgreSQL RANGE types. + + These are types that return a single :class:`_postgresql.Range` object. + + .. seealso:: + + `PostgreSQL range functions <https://www.postgresql.org/docs/current/static/functions-range.html>`_ + + """ # noqa: E501 + + __abstract__ = True + + def _resolve_for_literal(self, value: Range[Any]) -> Any: + spec = value.lower if value.lower is not None else value.upper + + if isinstance(spec, int): + # pg is unreasonably picky here: the query + # "select 1::INTEGER <@ '[1, 4)'::INT8RANGE" raises + # "operator does not exist: integer <@ int8range" as of pg 16 + if _is_int32(value): + return INT4RANGE() + else: + return INT8RANGE() + elif isinstance(spec, (Decimal, float)): + return NUMRANGE() + elif isinstance(spec, datetime): + return TSRANGE() if not spec.tzinfo else TSTZRANGE() + elif isinstance(spec, date): + return DATERANGE() + else: + # empty Range, SQL datatype can't be determined here + return sqltypes.NULLTYPE + + +class AbstractSingleRangeImpl(AbstractSingleRange[_T]): + """Marker for AbstractSingleRange that will apply a subclass-specific + adaptation""" + + +class AbstractMultiRange(AbstractRange[Sequence[Range[_T]]]): + """Base for PostgreSQL MULTIRANGE types. + + these are types that return a sequence of :class:`_postgresql.Range` + objects. + + """ + + __abstract__ = True + + def _resolve_for_literal(self, value: Sequence[Range[Any]]) -> Any: + if not value: + # empty MultiRange, SQL datatype can't be determined here + return sqltypes.NULLTYPE + first = value[0] + spec = first.lower if first.lower is not None else first.upper + + if isinstance(spec, int): + # pg is unreasonably picky here: the query + # "select 1::INTEGER <@ '{[1, 4),[6,19)}'::INT8MULTIRANGE" raises + # "operator does not exist: integer <@ int8multirange" as of pg 16 + if all(_is_int32(r) for r in value): + return INT4MULTIRANGE() + else: + return INT8MULTIRANGE() + elif isinstance(spec, (Decimal, float)): + return NUMMULTIRANGE() + elif isinstance(spec, datetime): + return TSMULTIRANGE() if not spec.tzinfo else TSTZMULTIRANGE() + elif isinstance(spec, date): + return DATEMULTIRANGE() + else: + # empty Range, SQL datatype can't be determined here + return sqltypes.NULLTYPE + + +class AbstractMultiRangeImpl(AbstractMultiRange[_T]): + """Marker for AbstractMultiRange that will apply a subclass-specific + adaptation""" + + +class INT4RANGE(AbstractSingleRange[int]): + """Represent the PostgreSQL INT4RANGE type.""" + + __visit_name__ = "INT4RANGE" + + +class INT8RANGE(AbstractSingleRange[int]): + """Represent the PostgreSQL INT8RANGE type.""" + + __visit_name__ = "INT8RANGE" + + +class NUMRANGE(AbstractSingleRange[Decimal]): + """Represent the PostgreSQL NUMRANGE type.""" + + __visit_name__ = "NUMRANGE" + + +class DATERANGE(AbstractSingleRange[date]): + """Represent the PostgreSQL DATERANGE type.""" + + __visit_name__ = "DATERANGE" + + +class TSRANGE(AbstractSingleRange[datetime]): + """Represent the PostgreSQL TSRANGE type.""" + + __visit_name__ = "TSRANGE" + + +class TSTZRANGE(AbstractSingleRange[datetime]): + """Represent the PostgreSQL TSTZRANGE type.""" + + __visit_name__ = "TSTZRANGE" + + +class INT4MULTIRANGE(AbstractMultiRange[int]): + """Represent the PostgreSQL INT4MULTIRANGE type.""" + + __visit_name__ = "INT4MULTIRANGE" + + +class INT8MULTIRANGE(AbstractMultiRange[int]): + """Represent the PostgreSQL INT8MULTIRANGE type.""" + + __visit_name__ = "INT8MULTIRANGE" + + +class NUMMULTIRANGE(AbstractMultiRange[Decimal]): + """Represent the PostgreSQL NUMMULTIRANGE type.""" + + __visit_name__ = "NUMMULTIRANGE" + + +class DATEMULTIRANGE(AbstractMultiRange[date]): + """Represent the PostgreSQL DATEMULTIRANGE type.""" + + __visit_name__ = "DATEMULTIRANGE" + + +class TSMULTIRANGE(AbstractMultiRange[datetime]): + """Represent the PostgreSQL TSRANGE type.""" + + __visit_name__ = "TSMULTIRANGE" + + +class TSTZMULTIRANGE(AbstractMultiRange[datetime]): + """Represent the PostgreSQL TSTZRANGE type.""" + + __visit_name__ = "TSTZMULTIRANGE" + + +_max_int_32 = 2**31 - 1 +_min_int_32 = -(2**31) + + +def _is_int32(r: Range[int]) -> bool: + return (r.lower is None or _min_int_32 <= r.lower <= _max_int_32) and ( + r.upper is None or _min_int_32 <= r.upper <= _max_int_32 + ) diff --git a/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/types.py b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/types.py new file mode 100644 index 0000000..2acf63b --- /dev/null +++ b/venv/lib/python3.11/site-packages/sqlalchemy/dialects/postgresql/types.py @@ -0,0 +1,303 @@ +# dialects/postgresql/types.py +# Copyright (C) 2013-2024 the SQLAlchemy authors and contributors +# <see AUTHORS file> +# +# This module is part of SQLAlchemy and is released under +# the MIT License: https://www.opensource.org/licenses/mit-license.php +from __future__ import annotations + +import datetime as dt +from typing import Any +from typing import Optional +from typing import overload +from typing import Type +from typing import TYPE_CHECKING +from uuid import UUID as _python_UUID + +from ...sql import sqltypes +from ...sql import type_api +from ...util.typing import Literal + +if TYPE_CHECKING: + from ...engine.interfaces import Dialect + from ...sql.operators import OperatorType + from ...sql.type_api import _LiteralProcessorType + from ...sql.type_api import TypeEngine + +_DECIMAL_TYPES = (1231, 1700) +_FLOAT_TYPES = (700, 701, 1021, 1022) +_INT_TYPES = (20, 21, 23, 26, 1005, 1007, 1016) + + +class PGUuid(sqltypes.UUID[sqltypes._UUID_RETURN]): + render_bind_cast = True + render_literal_cast = True + + if TYPE_CHECKING: + + @overload + def __init__( + self: PGUuid[_python_UUID], as_uuid: Literal[True] = ... + ) -> None: ... + + @overload + def __init__( + self: PGUuid[str], as_uuid: Literal[False] = ... + ) -> None: ... + + def __init__(self, as_uuid: bool = True) -> None: ... + + +class BYTEA(sqltypes.LargeBinary): + __visit_name__ = "BYTEA" + + +class INET(sqltypes.TypeEngine[str]): + __visit_name__ = "INET" + + +PGInet = INET + + +class CIDR(sqltypes.TypeEngine[str]): + __visit_name__ = "CIDR" + + +PGCidr = CIDR + + +class MACADDR(sqltypes.TypeEngine[str]): + __visit_name__ = "MACADDR" + + +PGMacAddr = MACADDR + + +class MACADDR8(sqltypes.TypeEngine[str]): + __visit_name__ = "MACADDR8" + + +PGMacAddr8 = MACADDR8 + + +class MONEY(sqltypes.TypeEngine[str]): + r"""Provide the PostgreSQL MONEY type. + + Depending on driver, result rows using this type may return a + string value which includes currency symbols. + + For this reason, it may be preferable to provide conversion to a + numerically-based currency datatype using :class:`_types.TypeDecorator`:: + + import re + import decimal + from sqlalchemy import Dialect + from sqlalchemy import TypeDecorator + + class NumericMoney(TypeDecorator): + impl = MONEY + + def process_result_value( + self, value: Any, dialect: Dialect + ) -> None: + if value is not None: + # adjust this for the currency and numeric + m = re.match(r"\$([\d.]+)", value) + if m: + value = decimal.Decimal(m.group(1)) + return value + + Alternatively, the conversion may be applied as a CAST using + the :meth:`_types.TypeDecorator.column_expression` method as follows:: + + import decimal + from sqlalchemy import cast + from sqlalchemy import TypeDecorator + + class NumericMoney(TypeDecorator): + impl = MONEY + + def column_expression(self, column: Any): + return cast(column, Numeric()) + + .. versionadded:: 1.2 + + """ + + __visit_name__ = "MONEY" + + +class OID(sqltypes.TypeEngine[int]): + """Provide the PostgreSQL OID type.""" + + __visit_name__ = "OID" + + +class REGCONFIG(sqltypes.TypeEngine[str]): + """Provide the PostgreSQL REGCONFIG type. + + .. versionadded:: 2.0.0rc1 + + """ + + __visit_name__ = "REGCONFIG" + + +class TSQUERY(sqltypes.TypeEngine[str]): + """Provide the PostgreSQL TSQUERY type. + + .. versionadded:: 2.0.0rc1 + + """ + + __visit_name__ = "TSQUERY" + + +class REGCLASS(sqltypes.TypeEngine[str]): + """Provide the PostgreSQL REGCLASS type. + + .. versionadded:: 1.2.7 + + """ + + __visit_name__ = "REGCLASS" + + +class TIMESTAMP(sqltypes.TIMESTAMP): + """Provide the PostgreSQL TIMESTAMP type.""" + + __visit_name__ = "TIMESTAMP" + + def __init__( + self, timezone: bool = False, precision: Optional[int] = None + ) -> None: + """Construct a TIMESTAMP. + + :param timezone: boolean value if timezone present, default False + :param precision: optional integer precision value + + .. versionadded:: 1.4 + + """ + super().__init__(timezone=timezone) + self.precision = precision + + +class TIME(sqltypes.TIME): + """PostgreSQL TIME type.""" + + __visit_name__ = "TIME" + + def __init__( + self, timezone: bool = False, precision: Optional[int] = None + ) -> None: + """Construct a TIME. + + :param timezone: boolean value if timezone present, default False + :param precision: optional integer precision value + + .. versionadded:: 1.4 + + """ + super().__init__(timezone=timezone) + self.precision = precision + + +class INTERVAL(type_api.NativeForEmulated, sqltypes._AbstractInterval): + """PostgreSQL INTERVAL type.""" + + __visit_name__ = "INTERVAL" + native = True + + def __init__( + self, precision: Optional[int] = None, fields: Optional[str] = None + ) -> None: + """Construct an INTERVAL. + + :param precision: optional integer precision value + :param fields: string fields specifier. allows storage of fields + to be limited, such as ``"YEAR"``, ``"MONTH"``, ``"DAY TO HOUR"``, + etc. + + .. versionadded:: 1.2 + + """ + self.precision = precision + self.fields = fields + + @classmethod + def adapt_emulated_to_native( + cls, interval: sqltypes.Interval, **kw: Any # type: ignore[override] + ) -> INTERVAL: + return INTERVAL(precision=interval.second_precision) + + @property + def _type_affinity(self) -> Type[sqltypes.Interval]: + return sqltypes.Interval + + def as_generic(self, allow_nulltype: bool = False) -> sqltypes.Interval: + return sqltypes.Interval(native=True, second_precision=self.precision) + + @property + def python_type(self) -> Type[dt.timedelta]: + return dt.timedelta + + def literal_processor( + self, dialect: Dialect + ) -> Optional[_LiteralProcessorType[dt.timedelta]]: + def process(value: dt.timedelta) -> str: + return f"make_interval(secs=>{value.total_seconds()})" + + return process + + +PGInterval = INTERVAL + + +class BIT(sqltypes.TypeEngine[int]): + __visit_name__ = "BIT" + + def __init__( + self, length: Optional[int] = None, varying: bool = False + ) -> None: + if varying: + # BIT VARYING can be unlimited-length, so no default + self.length = length + else: + # BIT without VARYING defaults to length 1 + self.length = length or 1 + self.varying = varying + + +PGBit = BIT + + +class TSVECTOR(sqltypes.TypeEngine[str]): + """The :class:`_postgresql.TSVECTOR` type implements the PostgreSQL + text search type TSVECTOR. + + It can be used to do full text queries on natural language + documents. + + .. seealso:: + + :ref:`postgresql_match` + + """ + + __visit_name__ = "TSVECTOR" + + +class CITEXT(sqltypes.TEXT): + """Provide the PostgreSQL CITEXT type. + + .. versionadded:: 2.0.7 + + """ + + __visit_name__ = "CITEXT" + + def coerce_compared_value( + self, op: Optional[OperatorType], value: Any + ) -> TypeEngine[Any]: + return self |