refactor: replace auto_register_python_variables with auto_register_python_objects in SessionContext

Author: kosiew

python/datafusion/context.py

@@ -53,12 +53,6 @@
 
     from datafusion.plan import ExecutionPlan, LogicalPlan
 
-_AUTO_REGISTER_PYTHON_VARIABLES_DEPRECATED = (
-    "SessionContext.auto_register_python_variables is deprecated; use "
-    "SessionContext.set_python_table_lookup() or the "
-    "'auto_register_python_objects' keyword argument instead."
-)
-
 
 class ArrowStreamExportable(Protocol):
     """Type hint for object exporting Arrow C Stream via Arrow PyCapsule Interface.
@@ -500,7 +494,6 @@ def __init__(
         runtime: RuntimeEnvBuilder | None = None,
         *,
         auto_register_python_objects: bool | None = None,
-        auto_register_python_variables: bool | None = None,
     ) -> None:
         """Main interface for executing queries with DataFusion.
 
@@ -517,9 +510,6 @@ def __init__(
                 the value configured via
                 :py:meth:`~datafusion.SessionConfig.with_python_table_lookup`
                 (``False`` unless explicitly enabled).
-            auto_register_python_variables: Deprecated alias for
-                ``auto_register_python_objects``. When provided, it overrides
-                the automatic registration behavior.
 
         Example usage:
 
@@ -536,29 +526,9 @@ def __init__(
             runtime.config_internal if runtime is not None else None,
         )
 
-        if auto_register_python_variables is not None:
-            warnings.warn(
-                _AUTO_REGISTER_PYTHON_VARIABLES_DEPRECATED,
-                DeprecationWarning,
-                stacklevel=2,
-            )
-
-        if (
-            auto_register_python_variables is not None
-            and auto_register_python_objects is not None
-            and auto_register_python_objects != auto_register_python_variables
-        ):
-            conflict_message = (
-                "auto_register_python_objects and auto_register_python_variables "
-                "were provided with conflicting values."
-            )
-            raise ValueError(conflict_message)
-
         # Determine the final value for python table lookup
         if auto_register_python_objects is not None:
             auto_python_table_lookup = auto_register_python_objects
-        elif auto_register_python_variables is not None:
-            auto_python_table_lookup = auto_register_python_variables
         else:
             # Default to session config value or False if not configured
             auto_python_table_lookup = getattr(config, "_python_table_lookup", False)
@@ -596,9 +566,7 @@ def enable_url_table(self) -> SessionContext:
         obj._auto_python_table_lookup = getattr(
             self, "_auto_python_table_lookup", False
         )
-        obj._python_table_bindings = getattr(
-            self, "_python_table_bindings", {}
-        ).copy()
+        obj._python_table_bindings = getattr(self, "_python_table_bindings", {}).copy()
         return obj
 
     def set_python_table_lookup(self, enabled: bool = True) -> SessionContext:
@@ -616,25 +584,6 @@ def set_python_table_lookup(self, enabled: bool = True) -> SessionContext:
         self._auto_python_table_lookup = enabled
         return self
 
-    @property
-    def auto_register_python_variables(self) -> bool:
-        """Deprecated alias for :py:meth:`set_python_table_lookup`."""
-        warnings.warn(
-            _AUTO_REGISTER_PYTHON_VARIABLES_DEPRECATED,
-            DeprecationWarning,
-            stacklevel=2,
-        )
-        return bool(getattr(self, "_auto_python_table_lookup", False))
-
-    @auto_register_python_variables.setter
-    def auto_register_python_variables(self, enabled: bool) -> None:
-        warnings.warn(
-            _AUTO_REGISTER_PYTHON_VARIABLES_DEPRECATED,
-            DeprecationWarning,
-            stacklevel=2,
-        )
-        self.set_python_table_lookup(bool(enabled))
-
     def register_object_store(
         self, schema: str, store: Any, host: str | None = None
     ) -> None:
@@ -792,34 +741,29 @@ def _register_python_tables(self, tables: list[str]) -> bool:
     def _lookup_python_object(name: str) -> Any | None:
         frame = inspect.currentframe()
         try:
-            if frame is not None:
-                frame = frame.f_back
+            frame = frame.f_back if frame is not None else None
             lower_name = name.lower()
 
             def _match(mapping: dict[str, Any]) -> Any | None:
-                if not mapping:
-                    return None
-
                 value = mapping.get(name)
                 if value is not None:
                     return value
 
                 for key, candidate in mapping.items():
-                    if isinstance(key, str) and key.lower() == lower_name:
-                        if candidate is not None:
-                            return candidate
+                    if (
+                        isinstance(key, str)
+                        and key.lower() == lower_name
+                        and candidate is not None
+                    ):
+                        return candidate
 
                 return None
 
             while frame is not None:
-                locals_dict = frame.f_locals
-                match = _match(locals_dict)
-                if match is not None:
-                    return match
-                globals_dict = frame.f_globals
-                match = _match(globals_dict)
-                if match is not None:
-                    return match
+                for scope in (frame.f_locals, frame.f_globals):
+                    match = _match(scope)
+                    if match is not None:
+                        return match
                 frame = frame.f_back
         finally:
             del frame

python/tests/test_context.py

@@ -295,7 +295,7 @@ def __init__(self) -> None:
 
 
 def test_sql_auto_register_arrow_table():
-    ctx = SessionContext(auto_register_python_variables=True)
+    ctx = SessionContext(auto_register_python_objects=True)
     arrow_table = pa.Table.from_pydict({"value": [1, 2, 3]})  # noqa: F841
 
     result = ctx.sql(
@@ -326,9 +326,7 @@ def test_sql_auto_register_multiple_tables_single_query():
     ).collect()
 
     actual = pa.Table.from_batches(result)
-    expected = pa.Table.from_pydict(
-        {"customer_id": [1, 2], "order_id": [100, 200]}
-    )
+    expected = pa.Table.from_pydict({"customer_id": [1, 2], "order_id": [100, 200]})
 
     assert actual.equals(expected)
     assert ctx.table_exist("customers")
@@ -337,7 +335,7 @@ def test_sql_auto_register_multiple_tables_single_query():
 
 def test_sql_auto_register_arrow_outer_scope():
     ctx = SessionContext()
-    ctx.auto_register_python_variables = True
+    ctx.set_python_table_lookup(True)
     arrow_table = pa.Table.from_pydict({"value": [1, 2, 3, 4]})  # noqa: F841
 
     def run_query():
@@ -365,7 +363,7 @@ def run_query():
 
 def test_sql_auto_register_case_insensitive_lookup():
     ctx = SessionContext(auto_register_python_objects=True)
-    MyTable = pa.Table.from_pydict({"value": [2, 3]})  # noqa: F841
+    MyTable = pa.Table.from_pydict({"value": [2, 3]})  # noqa: N806,F841
 
     batches = ctx.sql(
         "SELECT SUM(value) AS total FROM mytable",
@@ -377,7 +375,7 @@ def test_sql_auto_register_case_insensitive_lookup():
 def test_sql_auto_register_pandas_dataframe():
     pd = pytest.importorskip("pandas")
 
-    ctx = SessionContext(auto_register_python_variables=True)
+    ctx = SessionContext(auto_register_python_objects=True)
     pandas_df = pd.DataFrame({"value": [1, 2, 3, 4]})  # noqa: F841
 
     result = ctx.sql(
@@ -411,7 +409,7 @@ def test_sql_auto_register_refreshes_reassigned_dataframe():
 def test_sql_auto_register_polars_dataframe():
     pl = pytest.importorskip("polars")
 
-    ctx = SessionContext(auto_register_python_variables=True)
+    ctx = SessionContext(auto_register_python_objects=True)
     polars_df = pl.DataFrame({"value": [2, 4, 6]})  # noqa: F841
 
     result = ctx.sql(
@@ -421,39 +419,6 @@ def test_sql_auto_register_polars_dataframe():
     assert result[0].column(0).to_pylist()[0] == 2
 
 
-def test_session_context_constructor_alias_disables_lookup():
-    with pytest.deprecated_call():
-        ctx = SessionContext(auto_register_python_variables=False)
-
-    arrow_table = pa.Table.from_pydict({"value": [1, 2, 3]})  # noqa: F841
-
-    with pytest.raises(Exception, match="not found|No table named"):
-        ctx.sql("SELECT * FROM arrow_table").collect()
-
-    with pytest.deprecated_call():
-        assert ctx.auto_register_python_variables is False
-
-
-def test_session_context_property_alias_setter_enables_lookup():
-    ctx = SessionContext(auto_register_python_objects=False)
-    arrow_table = pa.Table.from_pydict({"value": [1, 2, 3]})  # noqa: F841
-
-    with pytest.raises(Exception, match="not found|No table named"):
-        ctx.sql("SELECT COUNT(*) FROM arrow_table").collect()
-
-    with pytest.deprecated_call():
-        ctx.auto_register_python_variables = True
-
-    result = ctx.sql(
-        "SELECT SUM(value) AS total FROM arrow_table",
-    ).collect()
-
-    assert result[0].column(0).to_pylist()[0] == 6
-
-    with pytest.deprecated_call():
-        assert ctx.auto_register_python_variables is True
-
-
 def test_from_pydict(ctx):
     # create a dataframe from Python dictionary
     data = {"a": [1, 2, 3], "b": [4, 5, 6]}
@@ -486,7 +451,7 @@ def test_from_pandas(ctx):
 
 def test_sql_from_local_arrow_table(ctx):
     ctx.set_python_table_lookup(True)  # Enable implicit table lookup
-    arrow_table = pa.Table.from_pydict({"a": [1, 2], "b": ["x", "y"]})
+    arrow_table = pa.Table.from_pydict({"a": [1, 2], "b": ["x", "y"]})  # noqa: F841
 
     result = ctx.sql("SELECT * FROM arrow_table ORDER BY a").collect()
     actual = pa.Table.from_batches(result)
@@ -498,7 +463,7 @@ def test_sql_from_local_arrow_table(ctx):
 def test_sql_from_local_pandas_dataframe(ctx):
     ctx.set_python_table_lookup(True)  # Enable implicit table lookup
     pd = pytest.importorskip("pandas")
-    pandas_df = pd.DataFrame({"a": [3, 1], "b": ["z", "y"]})
+    pandas_df = pd.DataFrame({"a": [3, 1], "b": ["z", "y"]})  # noqa: F841
 
     result = ctx.sql("SELECT * FROM pandas_df ORDER BY a").collect()
     actual = pa.Table.from_batches(result)
@@ -510,7 +475,7 @@ def test_sql_from_local_pandas_dataframe(ctx):
 def test_sql_from_local_polars_dataframe(ctx):
     ctx.set_python_table_lookup(True)  # Enable implicit table lookup
     pl = pytest.importorskip("polars")
-    polars_df = pl.DataFrame({"a": [2, 1], "b": ["beta", "alpha"]})
+    polars_df = pl.DataFrame({"a": [2, 1], "b": ["beta", "alpha"]})  # noqa: F841
 
     result = ctx.sql("SELECT * FROM polars_df ORDER BY a").collect()
     actual = pa.Table.from_batches(result)
@@ -520,7 +485,7 @@ def test_sql_from_local_polars_dataframe(ctx):
 
 
 def test_sql_from_local_unsupported_object(ctx):
-    unsupported = object()
+    unsupported = object()  # noqa: F841
 
     with pytest.raises(Exception, match="table 'unsupported' not found"):
         ctx.sql("SELECT * FROM unsupported").collect()
@@ -876,7 +841,7 @@ def test_sql_with_options_no_statements(ctx):
 def test_session_config_python_table_lookup_enables_auto_registration():
     pd = pytest.importorskip("pandas")
 
-    ctx = SessionContext(config=SessionConfig().with_python_table_lookup(True))
+    ctx = SessionContext(config=SessionConfig().with_python_table_lookup(enabled=True))
     pdf = pd.DataFrame({"value": [1, 2, 3]})
     assert len(pdf) == 3