add function for parallel execution of SQL scripts

[nominatim.git] / nominatim / db / connection.py

diff --git a/nominatim/db/connection.py b/nominatim/db/connection.py
index 90194e63432070d434c47c6600bdbdfdaf531286..77d463d81d8e4dc0522bf1b2702aff0db66f21df 100644 (file)
--- a/nominatim/db/connection.py
+++ b/nominatim/db/connection.py
@@ -7,6 +7,7 @@
 """
 Specialised connection and cursor functions.
 """
 """
 Specialised connection and cursor functions.
 """

+from typing import Optional, Any, Callable, ContextManager, Dict, cast, overload, Tuple, Iterable

 import contextlib
 import logging
 import os
 import contextlib
 import logging
 import os


@@ -16,24 +17,27 @@ import psycopg2.extensions
 import psycopg2.extras
 from psycopg2 import sql as pysql
 
 import psycopg2.extras
 from psycopg2 import sql as pysql
 

+from nominatim.typing import SysEnv, Query, T_cursor

 from nominatim.errors import UsageError
 
 LOG = logging.getLogger()
 
 from nominatim.errors import UsageError
 
 LOG = logging.getLogger()
 

-class _Cursor(psycopg2.extras.DictCursor):
+class Cursor(psycopg2.extras.DictCursor):

     """ A cursor returning dict-like objects and providing specialised
         execution functions.
     """
     """ A cursor returning dict-like objects and providing specialised
         execution functions.
     """

-
-    def execute(self, query, args=None): # pylint: disable=W0221
+    # pylint: disable=arguments-renamed,arguments-differ
+    def execute(self, query: Query, args: Any = None) -> None:

         """ Query execution that logs the SQL query when debugging is enabled.
         """
         """ Query execution that logs the SQL query when debugging is enabled.
         """

-        LOG.debug(self.mogrify(query, args).decode('utf-8'))
+        if LOG.isEnabledFor(logging.DEBUG):
+            LOG.debug(self.mogrify(query, args).decode('utf-8')) # type: ignore[no-untyped-call]

 
         super().execute(query, args)
 
 
 
         super().execute(query, args)
 
 

-    def execute_values(self, sql, argslist, template=None):
+    def execute_values(self, sql: Query, argslist: Iterable[Tuple[Any, ...]],
+                       template: Optional[Query] = None) -> None:

         """ Wrapper for the psycopg2 convenience function to execute
             SQL for a list of values.
         """
         """ Wrapper for the psycopg2 convenience function to execute
             SQL for a list of values.
         """


@@ -42,7 +46,7 @@ class _Cursor(psycopg2.extras.DictCursor):
         psycopg2.extras.execute_values(self, sql, argslist, template=template)
 
 
         psycopg2.extras.execute_values(self, sql, argslist, template=template)
 
 

-    def scalar(self, sql, args=None):
+    def scalar(self, sql: Query, args: Any = None) -> Any:

         """ Execute query that returns a single value. The value is returned.
             If the query yields more than one row, a ValueError is raised.
         """
         """ Execute query that returns a single value. The value is returned.
             If the query yields more than one row, a ValueError is raised.
         """


@@ -51,12 +55,15 @@ class _Cursor(psycopg2.extras.DictCursor):
         if self.rowcount != 1:
             raise RuntimeError("Query did not return a single row.")
 
         if self.rowcount != 1:
             raise RuntimeError("Query did not return a single row.")
 

-        return self.fetchone()[0]
+        result = self.fetchone()
+        assert result is not None
+
+        return result[0]

 
 
 
 

-    def drop_table(self, name, if_exists=True, cascade=False):
+    def drop_table(self, name: str, if_exists: bool = True, cascade: bool = False) -> None:

         """ Drop the table with the given name.
         """ Drop the table with the given name.

-            Set `if_exists` to False if a non-existant table should raise
+            Set `if_exists` to False if a non-existent table should raise

             an exception instead of just being ignored. If 'cascade' is set
             to True then all dependent tables are deleted as well.
         """
             an exception instead of just being ignored. If 'cascade' is set
             to True then all dependent tables are deleted as well.
         """


@@ -70,27 +77,38 @@ class _Cursor(psycopg2.extras.DictCursor):
         self.execute(pysql.SQL(sql).format(pysql.Identifier(name)))
 
 
         self.execute(pysql.SQL(sql).format(pysql.Identifier(name)))
 
 

-class _Connection(psycopg2.extensions.connection):
+class Connection(psycopg2.extensions.connection):

     """ A connection that provides the specialised cursor by default and
         adds convenience functions for administrating the database.
     """
     """ A connection that provides the specialised cursor by default and
         adds convenience functions for administrating the database.
     """

+    @overload # type: ignore[override]
+    def cursor(self) -> Cursor:
+        ...
+
+    @overload
+    def cursor(self, name: str) -> Cursor:
+        ...

 
 

-    def cursor(self, cursor_factory=_Cursor, **kwargs):
+    @overload
+    def cursor(self, cursor_factory: Callable[..., T_cursor]) -> T_cursor:
+        ...
+
+    def cursor(self, cursor_factory  = Cursor, **kwargs): # type: ignore

         """ Return a new cursor. By default the specialised cursor is returned.
         """
         return super().cursor(cursor_factory=cursor_factory, **kwargs)
 
 
         """ Return a new cursor. By default the specialised cursor is returned.
         """
         return super().cursor(cursor_factory=cursor_factory, **kwargs)
 
 

-    def table_exists(self, table):
+    def table_exists(self, table: str) -> bool:

         """ Check that a table with the given name exists in the database.
         """
         with self.cursor() as cur:
             num = cur.scalar("""SELECT count(*) FROM pg_tables
                                 WHERE tablename = %s and schemaname = 'public'""", (table, ))
         """ Check that a table with the given name exists in the database.
         """
         with self.cursor() as cur:
             num = cur.scalar("""SELECT count(*) FROM pg_tables
                                 WHERE tablename = %s and schemaname = 'public'""", (table, ))

-            return num == 1
+            return num == 1 if isinstance(num, int) else False

 
 
 
 

-    def table_has_column(self, table, column):
+    def table_has_column(self, table: str, column: str) -> bool:

         """ Check if the table 'table' exists and has a column with name 'column'.
         """
         with self.cursor() as cur:
         """ Check if the table 'table' exists and has a column with name 'column'.
         """
         with self.cursor() as cur:


@@ -98,10 +116,10 @@ class _Connection(psycopg2.extensions.connection):
                                        WHERE table_name = %s
                                              and column_name = %s""",
                                     (table, column))
                                        WHERE table_name = %s
                                              and column_name = %s""",
                                     (table, column))

-            return has_column > 0
+            return has_column > 0 if isinstance(has_column, int) else False

 
 
 
 

-    def index_exists(self, index, table=None):
+    def index_exists(self, index: str, table: Optional[str] = None) -> bool:

         """ Check that an index with the given name exists in the database.
             If table is not None then the index must relate to the given
             table.
         """ Check that an index with the given name exists in the database.
             If table is not None then the index must relate to the given
             table.


@@ -114,14 +132,16 @@ class _Connection(psycopg2.extensions.connection):
 
             if table is not None:
                 row = cur.fetchone()
 
             if table is not None:
                 row = cur.fetchone()

+                if row is None or not isinstance(row[0], str):
+                    return False

                 return row[0] == table
 
         return True
 
 
                 return row[0] == table
 
         return True
 
 

-    def drop_table(self, name, if_exists=True, cascade=False):
+    def drop_table(self, name: str, if_exists: bool = True, cascade: bool = False) -> None:

         """ Drop the table with the given name.
         """ Drop the table with the given name.

-            Set `if_exists` to False if a non-existant table should raise
+            Set `if_exists` to False if a non-existent table should raise

             an exception instead of just being ignored.
         """
         with self.cursor() as cur:
             an exception instead of just being ignored.
         """
         with self.cursor() as cur:


@@ -129,18 +149,18 @@ class _Connection(psycopg2.extensions.connection):
         self.commit()
 
 
         self.commit()
 
 

-    def server_version_tuple(self):
+    def server_version_tuple(self) -> Tuple[int, int]:

         """ Return the server version as a tuple of (major, minor).
             Converts correctly for pre-10 and post-10 PostgreSQL versions.
         """
         version = self.server_version
         if version < 100000:
         """ Return the server version as a tuple of (major, minor).
             Converts correctly for pre-10 and post-10 PostgreSQL versions.
         """
         version = self.server_version
         if version < 100000:

-            return (int(version / 10000), (version % 10000) / 100)
+            return (int(version / 10000), int((version % 10000) / 100))

 
         return (int(version / 10000), version % 10000)
 
 
 
         return (int(version / 10000), version % 10000)
 
 

-    def postgis_version_tuple(self):
+    def postgis_version_tuple(self) -> Tuple[int, int]:

         """ Return the postgis version installed in the database as a
             tuple of (major, minor). Assumes that the PostGIS extension
             has been installed already.
         """ Return the postgis version installed in the database as a
             tuple of (major, minor). Assumes that the PostGIS extension
             has been installed already.


@@ -148,18 +168,27 @@ class _Connection(psycopg2.extensions.connection):
         with self.cursor() as cur:
             version = cur.scalar('SELECT postgis_lib_version()')
 
         with self.cursor() as cur:
             version = cur.scalar('SELECT postgis_lib_version()')
 

-        return tuple((int(x) for x in version.split('.')[:2]))
+        version_parts = version.split('.')
+        if len(version_parts) < 2:
+            raise UsageError(f"Error fetching Postgis version. Bad format: {version}")
+
+        return (int(version_parts[0]), int(version_parts[1]))

 
 

+class ConnectionContext(ContextManager[Connection]):
+    """ Context manager of the connection that also provides direct access
+        to the underlying connection.
+    """
+    connection: Connection

 
 

-def connect(dsn):
+def connect(dsn: str) -> ConnectionContext:

     """ Open a connection to the database using the specialised connection
         factory. The returned object may be used in conjunction with 'with'.
         When used outside a context manager, use the `connection` attribute
         to get the connection.
     """
     try:
     """ Open a connection to the database using the specialised connection
         factory. The returned object may be used in conjunction with 'with'.
         When used outside a context manager, use the `connection` attribute
         to get the connection.
     """
     try:

-        conn = psycopg2.connect(dsn, connection_factory=_Connection)
-        ctxmgr = contextlib.closing(conn)
+        conn = psycopg2.connect(dsn, connection_factory=Connection)
+        ctxmgr = cast(ConnectionContext, contextlib.closing(conn))

         ctxmgr.connection = conn
         return ctxmgr
     except psycopg2.OperationalError as err:
         ctxmgr.connection = conn
         return ctxmgr
     except psycopg2.OperationalError as err:


@@ -198,7 +227,8 @@ _PG_CONNECTION_STRINGS = {
 }
 
 
 }
 
 

-def get_pg_env(dsn, base_env=None):
+def get_pg_env(dsn: str,
+               base_env: Optional[SysEnv] = None) -> Dict[str, str]:

     """ Return a copy of `base_env` with the environment variables for
         PostgresSQL set up from the given database connection string.
         If `base_env` is None, then the OS environment is used as a base
     """ Return a copy of `base_env` with the environment variables for
         PostgresSQL set up from the given database connection string.
         If `base_env` is None, then the OS environment is used as a base