connect(): Added support for shared connection; Database.is_closed property

Author: erezsh

data_diff/sqeleton/abcs/database_types.py

@@ -284,6 +284,11 @@ def close(self):
         "Close connection(s) to the database instance. Querying will stop functioning."
         ...
 
+    @property
+    @abstractmethod
+    def is_closed(self) -> bool:
+        "Return whether or not the connection has been closed"
+
     @abstractmethod
     def _normalize_table_path(self, path: DbPath) -> DbPath:
         ...

data_diff/sqeleton/databases/base.py

@@ -251,6 +251,7 @@ class Database(AbstractDatabase):
     CONNECT_URI_KWPARAMS = []
 
     _interactive = False
+    is_closed = False
 
     @property
     def name(self):
@@ -440,6 +441,10 @@ def _query_conn(self, conn, sql_code: Union[str, ThreadLocalInterpreter]) -> lis
         callback = partial(self._query_cursor, c)
         return apply_query(callback, sql_code)
 
+    def close(self):
+        self.is_closed = True
+        return super().close()
+
 
 class ThreadedDatabase(Database):
     """Access the database through singleton threads.
@@ -476,6 +481,7 @@ def create_connection(self):
         ...
 
     def close(self):
+        super().close()
         self._queue.shutdown()
 
     @property

data_diff/sqeleton/databases/bigquery.py

@@ -141,6 +141,7 @@ def _query(self, sql_code: Union[str, ThreadLocalInterpreter]):
         return apply_query(self._query_atom, sql_code)
 
     def close(self):
+        super().close()
         self._client.close()
 
     def select_table_schema(self, path: DbPath) -> str:

data_diff/sqeleton/databases/connect.py

@@ -1,9 +1,11 @@
 from typing import Type, List, Optional, Union, Dict
 from itertools import zip_longest
 import dsnparse
+from contextlib import suppress
 
 from runtype import dataclass
 
+from ..utils import WeakCache
 from .base import Database, ThreadedDatabase
 from .postgresql import PostgreSQL
 from .mysql import MySQL
@@ -19,12 +21,13 @@
 from .duckdb import DuckDB
 
 
+
 @dataclass
 class MatchUriPath:
     database_cls: Type[Database]
     params: List[str]
     kwparams: List[str] = []
-    help_str: str
+    help_str: str = "<unspecified>"
 
     def __post_init__(self):
         assert self.params == self.database_cls.CONNECT_URI_PARAMS, self.params
@@ -101,6 +104,7 @@ def __init__(self, database_by_scheme: Dict[str, Database]):
             name: MatchUriPath(cls, cls.CONNECT_URI_PARAMS, cls.CONNECT_URI_KWPARAMS, help_str=cls.CONNECT_URI_HELP)
             for name, cls in database_by_scheme.items()
         }
+        self.conn_cache = WeakCache()
 
     def connect_to_uri(self, db_uri: str, thread_count: Optional[int] = 1) -> Database:
         """Connect to the given database uri
@@ -200,7 +204,7 @@ def _connection_created(self, db):
         "Nop function to be overridden by subclasses."
         return db
 
-    def __call__(self, db_conf: Union[str, dict], thread_count: Optional[int] = 1) -> Database:
+    def __call__(self, db_conf: Union[str, dict], thread_count: Optional[int] = 1, shared: bool = True) -> Database:
         """Connect to a database using the given database configuration.
 
         Configuration can be given either as a URI string, or as a dict of {option: value}.
@@ -235,8 +239,19 @@ def __call__(self, db_conf: Union[str, dict], thread_count: Optional[int] = 1) -
             >>> connect({"driver": "mysql", "host": "localhost", "database": "db"})
             <data_diff.databases.mysql.MySQL object at 0x0000025DB3F94820>
         """
+        if shared:
+            with suppress(KeyError):
+                conn = self.conn_cache.get(db_conf)
+                if not conn.is_closed:
+                    return conn
+
         if isinstance(db_conf, str):
-            return self.connect_to_uri(db_conf, thread_count)
+            conn = self.connect_to_uri(db_conf, thread_count)
         elif isinstance(db_conf, dict):
-            return self.connect_with_dict(db_conf, thread_count)
-        raise TypeError(f"db configuration must be a URI string or a dictionary. Instead got '{db_conf}'.")
+            conn = self.connect_with_dict(db_conf, thread_count)
+        else:
+            raise TypeError(f"db configuration must be a URI string or a dictionary. Instead got '{db_conf}'.")
+
+        if shared:
+            self.conn_cache.add(db_conf, conn)
+        return conn

data_diff/sqeleton/databases/duckdb.py

@@ -132,6 +132,7 @@ def _query(self, sql_code: Union[str, ThreadLocalInterpreter]):
         return self._query_conn(self._conn, sql_code)
 
     def close(self):
+        super().close()
         self._conn.close()
 
     def create_connection(self):

data_diff/sqeleton/databases/presto.py

@@ -171,6 +171,7 @@ def _query(self, sql_code: str) -> list:
         return query_cursor(c, sql_code)
 
     def close(self):
+        super().close()
         self._conn.close()
 
     def select_table_schema(self, path: DbPath) -> str:

data_diff/sqeleton/databases/snowflake.py

@@ -134,6 +134,7 @@ def __init__(self, *, schema: str, **kw):
         self.default_schema = schema
 
     def close(self):
+        super().close()
         self._conn.close()
 
     def _query(self, sql_code: Union[str, ThreadLocalInterpreter]):

data_diff/sqeleton/utils.py

@@ -1,3 +1,5 @@
+from typing import Union, Dict, Any, Hashable
+from weakref import ref
 from typing import TypeVar
 from typing import Iterable, Iterator, MutableMapping, Union, Any, Sequence, Dict
 from abc import abstractmethod
@@ -9,6 +11,30 @@
 # -- Common --
 
 
+class WeakCache:
+    def __init__(self):
+        self._cache = {}
+
+    def _hashable_key(self, k: Union[dict, Hashable]) -> Hashable:
+        if isinstance(k, dict):
+            return tuple(k.items())
+        return k
+
+    def add(self, key: Union[dict, Hashable], value: Any):
+        key = self._hashable_key(key)
+        self._cache[key] = ref(value)
+
+    def get(self, key: Union[dict, Hashable]) -> Any:
+        key = self._hashable_key(key)
+
+        value = self._cache[key]()
+        if value is None:
+            del self._cache[key]
+            raise KeyError(f"Key {key} not found, or no longer a valid reference")
+
+        return value
+
+
 def join_iter(joiner: Any, iterable: Iterable) -> Iterable:
     it = iter(iterable)
     try:

tests/test_api.py

@@ -1,7 +1,7 @@
 import arrow
 from datetime import datetime
 
-from data_diff import diff_tables, connect_to_table
+from data_diff import diff_tables, connect_to_table, Algorithm
 from data_diff.databases import MySQL
 from data_diff.sqeleton.queries import table, commit
 
@@ -36,13 +36,17 @@ def setUp(self) -> None:
         )
 
     def test_api(self):
+        # test basic
         t1 = connect_to_table(TEST_MYSQL_CONN_STRING, self.table_src_name)
         t2 = connect_to_table(TEST_MYSQL_CONN_STRING, (self.table_dst_name,))
-        diff = list(diff_tables(t1, t2))
+        diff = list(diff_tables(t1, t2, algorithm=Algorithm.JOINDIFF))
         assert len(diff) == 1
 
-        t1.database.close()
-        t2.database.close()
+        # test algorithm
+        # (also tests shared connection on connect_to_table)
+        for algo in (Algorithm.HASHDIFF, Algorithm.JOINDIFF):
+            diff = list(diff_tables(t1, t2, algorithm=algo))
+            assert len(diff) == 1
 
         # test where
         diff_id = diff[0][1][0]
@@ -53,9 +57,6 @@ def test_api(self):
         diff = list(diff_tables(t1, t2))
         assert len(diff) == 0
 
-        t1.database.close()
-        t2.database.close()
-
     def test_api_get_stats_dict(self):
         # XXX Likely to change in the future
         expected_dict = {
@@ -76,6 +77,3 @@ def test_api_get_stats_dict(self):
         self.assertEqual(expected_dict, output)
         self.assertIsNotNone(diff)
         assert len(list(diff)) == 1
-
-        t1.database.close()
-        t2.database.close()