Enable extension in Docker

superset-core/src/superset_core/semantic_layers/semantic_view.py

@@ -21,6 +21,7 @@ import enum
 from typing import Protocol, runtime_checkable
 
 from superset_core.semantic_layers.types import (
+    AdhocFilter,
     Dimension,
     Filter,
     GroupLimit,
@@ -68,7 +69,7 @@ class SemanticView(Protocol):
     def get_values(
         self,
         dimension: Dimension,
-        filters: set[Filter] | None = None,
+        filters: set[Filter | AdhocFilter] | None = None,
     ) -> SemanticResult:
         """
         Return distinct values for a dimension.
@@ -78,7 +79,7 @@ class SemanticView(Protocol):
         self,
         metrics: list[Metric],
         dimensions: list[Dimension],
-        filters: set[Filter] | None = None,
+        filters: set[Filter | AdhocFilter] | None = None,
         order: list[OrderTuple] | None = None,
         limit: int | None = None,
         offset: int | None = None,
@@ -93,7 +94,7 @@ class SemanticView(Protocol):
         self,
         metrics: list[Metric],
         dimensions: list[Dimension],
-        filters: set[Filter] | None = None,
+        filters: set[Filter | AdhocFilter] | None = None,
         order: list[OrderTuple] | None = None,
         limit: int | None = None,
         offset: int | None = None,

superset-core/src/superset_core/semantic_layers/types.py

@@ -239,7 +239,6 @@ class Operator(str, enum.Enum):
     NOT_LIKE = "NOT LIKE"
     IS_NULL = "IS NULL"
     IS_NOT_NULL = "IS NOT NULL"
-    ADHOC = "ADHOC"
 
 
 FilterValues = str | int | float | bool | datetime | date | time | timedelta | None
@@ -253,11 +252,19 @@ class PredicateType(enum.Enum):
 @dataclass(frozen=True, order=True)
 class Filter:
     type: PredicateType
-    column: Dimension | Metric | None
+    column: Dimension | Metric
     operator: Operator
     value: FilterValues | frozenset[FilterValues]
 
 
+# TODO (betodealmeida): convert into Operator:
+# Filter(type=..., column=None, operator=Operator.AdHoc, value="some definition")
+@dataclass(frozen=True, order=True)
+class AdhocFilter:
+    type: PredicateType
+    definition: str
+
+
 class OrderDirection(enum.Enum):
     ASC = "ASC"
     DESC = "DESC"
@@ -284,7 +291,7 @@ class GroupLimit:
     metric: Metric | None
     direction: OrderDirection = OrderDirection.DESC
     group_others: bool = False
-    filters: set[Filter] | None = None
+    filters: set[Filter | AdhocFilter] | None = None
 
 
 @dataclass(frozen=True)
@@ -321,7 +328,7 @@ class SemanticQuery:
 
     metrics: list[Metric]
     dimensions: list[Dimension]
-    filters: set[Filter] | None = None
+    filters: set[Filter | AdhocFilter] | None = None
     order: list[OrderTuple] | None = None
     limit: int | None = None
     offset: int | None = None

superset/create_pandas_semantic_layer.py

@@ -1,99 +0,0 @@
-"""
-Script to create a Pandas semantic layer and Sales semantic view in Superset.
-
-Run this inside the superset_app container:
-    python /app/superset/create_pandas_semantic_layer.py
-"""
-
-from __future__ import annotations
-
-import logging
-import sys
-from typing import TYPE_CHECKING
-
-# Add the Superset application directory to the Python path
-sys.path.insert(0, "/app")
-
-from superset.app import create_app
-from superset.extensions import db
-from superset.utils import json
-
-if TYPE_CHECKING:
-    from superset.semantic_layers.models import SemanticLayer, SemanticView
-
-app = create_app()
-app.app_context().push()
-
-# Configure logging
-logging.basicConfig(
-    level=logging.INFO,
-    format="%(asctime)s - %(levelname)s - %(message)s",
-)
-logger = logging.getLogger(__name__)
-
-
-def create_pandas_semantic_layer() -> SemanticLayer:
-    """Create a Pandas semantic layer with minimal configuration."""
-    from superset.semantic_layers.models import SemanticLayer
-
-    logger.info("Creating Pandas semantic layer...")
-
-    configuration = {
-        "dataset": "sales",
-    }
-
-    semantic_layer = SemanticLayer(
-        name="Pandas Semantic Layer",
-        description="In-memory semantic layer backed by a Pandas DataFrame",
-        type="pandas",
-        configuration=json.dumps(configuration),
-        cache_timeout=3600,
-    )
-
-    db.session.add(semantic_layer)
-    db.session.commit()
-
-    logger.info("Created semantic layer:")
-    logger.info("  Name: %s", semantic_layer.name)
-    logger.info("  UUID: %s", semantic_layer.uuid)
-    logger.info("  Type: %s", semantic_layer.type)
-
-    return semantic_layer
-
-
-def create_sales_semantic_view(semantic_layer: SemanticLayer) -> SemanticView:
-    """Create the Sales semantic view."""
-    from superset.semantic_layers.models import SemanticView
-
-    logger.info("Creating Sales semantic view...")
-
-    semantic_view = SemanticView(
-        name="sales",
-        configuration="{}",
-        cache_timeout=1800,
-        semantic_layer_uuid=semantic_layer.uuid,
-    )
-
-    db.session.add(semantic_view)
-    db.session.commit()
-
-    logger.info("Created semantic view:")
-    logger.info("  Name: %s", semantic_view.name)
-    logger.info("  UUID: %s", semantic_view.uuid)
-    logger.info("  Semantic Layer UUID: %s", semantic_view.semantic_layer_uuid)
-
-    return semantic_view
-
-
-def main() -> None:
-    """Main script execution."""
-    logger.info("=" * 60)
-    logger.info("Creating Pandas Semantic Layer and Sales Semantic View")
-    logger.info("=" * 60)
-
-    semantic_layer = create_pandas_semantic_layer()
-    create_sales_semantic_view(semantic_layer)
-
-
-if __name__ == "__main__":
-    main()

superset/semantic_layers/mapper.py

@@ -32,6 +32,7 @@ import numpy as np
 from superset_core.semantic_layers.semantic_view import SemanticViewFeature
 from superset_core.semantic_layers.types import (
     AdhocExpression,
+    AdhocFilter,
     Day,
     Dimension,
     Filter,
@@ -369,14 +370,14 @@ def _get_filters_from_query_object(
     query_object: ValidatedQueryObject,
     time_offset: str | None,
     all_dimensions: dict[str, Dimension],
-) -> set[Filter]:
+) -> set[Filter | AdhocFilter]:
     """
     Extract all filters from the query object, including time range filters.
 
     This simplifies the complexity of from_dttm/to_dttm/inner_from_dttm/inner_to_dttm
     by converting all time constraints into filters.
     """
-    filters: set[Filter] = set()
+    filters: set[Filter | AdhocFilter] = set()
 
     # 1. Add fetch values predicate if present
     if (
@@ -384,11 +385,9 @@ def _get_filters_from_query_object(
         and query_object.datasource.fetch_values_predicate
     ):
         filters.add(
-            Filter(
+            AdhocFilter(
                 type=PredicateType.WHERE,
-                column=None,
-                operator=Operator.ADHOC,
-                value=query_object.datasource.fetch_values_predicate,
+                definition=query_object.datasource.fetch_values_predicate,
             )
         )
 
@@ -416,7 +415,7 @@ def _get_filters_from_query_object(
     return filters
 
 
-def _get_filters_from_extras(extras: dict[str, Any]) -> set[Filter]:
+def _get_filters_from_extras(extras: dict[str, Any]) -> set[AdhocFilter]:
     """
     Extract filters from the extras dict.
 
@@ -431,29 +430,25 @@ def _get_filters_from_extras(extras: dict[str, Any]) -> set[Filter]:
       Handled in _convert_time_grain() and used for dimension grain matching
 
     Note: The WHERE and HAVING clauses from extras are SQL expressions that
-    are passed through as-is to the semantic layer as adhoc Filter objects.
+    are passed through as-is to the semantic layer as AdhocFilter objects.
     """
-    filters: set[Filter] = set()
+    filters: set[AdhocFilter] = set()
 
     # Add WHERE clause from extras
     if where_clause := extras.get("where"):
         filters.add(
-            Filter(
+            AdhocFilter(
                 type=PredicateType.WHERE,
-                column=None,
-                operator=Operator.ADHOC,
-                value=where_clause,
+                definition=where_clause,
             )
         )
 
     # Add HAVING clause from extras
     if having_clause := extras.get("having"):
         filters.add(
-            Filter(
+            AdhocFilter(
                 type=PredicateType.HAVING,
-                column=None,
-                operator=Operator.ADHOC,
-                value=having_clause,
+                definition=having_clause,
             )
         )
 
@@ -545,7 +540,7 @@ def _convert_query_object_filter(
     all_dimensions: dict[str, Dimension],
 ) -> set[Filter] | None:
     """
-    Convert a QueryObject filter dict to a semantic layer Filter.
+    Convert a QueryObject filter dict to a semantic layer Filter or AdhocFilter.
     """
     operator_str = filter_["op"]
 
@@ -569,6 +564,7 @@ def _convert_query_object_filter(
     if operator_str == FilterOperator.TEMPORAL_RANGE.value:
         if not isinstance(value, str) or value == NO_TIME_RANGE:
             return None
+
         start, end = value.split(" : ")
         return {
             Filter(
@@ -681,7 +677,7 @@ def _get_group_limit_from_query_object(
 def _get_group_limit_filters(
     query_object: ValidatedQueryObject,
     all_dimensions: dict[str, Dimension],
-) -> set[Filter] | None:
+) -> set[Filter | AdhocFilter] | None:
     """
     Get separate filters for the group limit subquery if needed.
 
@@ -704,7 +700,7 @@ def _get_group_limit_filters(
         return None
 
     # Create separate filters for the group limit subquery
-    filters: set[Filter] = set()
+    filters: set[Filter | AdhocFilter] = set()
 
     # Add time range filter using inner bounds
     if query_object.granularity:
@@ -737,11 +733,9 @@ def _get_group_limit_filters(
         and query_object.datasource.fetch_values_predicate
     ):
         filters.add(
-            Filter(
+            AdhocFilter(
                 type=PredicateType.WHERE,
-                column=None,
-                operator=Operator.ADHOC,
-                value=query_object.datasource.fetch_values_predicate,
+                definition=query_object.datasource.fetch_values_predicate,
             )
         )
 

tests/unit_tests/semantic_layers/mapper_test.py

@@ -24,6 +24,7 @@ from pytest_mock import MockerFixture
 from superset_core.semantic_layers.semantic_view import SemanticViewFeature
 from superset_core.semantic_layers.types import (
     AdhocExpression,
+    AdhocFilter,
     Day,
     Dimension,
     Filter,
@@ -201,11 +202,9 @@ def test_get_filters_from_extras_where() -> None:
 
     assert len(result) == 1
     filter_ = next(iter(result))
-    assert isinstance(filter_, Filter)
+    assert isinstance(filter_, AdhocFilter)
     assert filter_.type == PredicateType.WHERE
-    assert filter_.column is None
-    assert filter_.operator == Operator.ADHOC
-    assert filter_.value == "customer_id > 100"
+    assert filter_.definition == "customer_id > 100"
 
 
 def test_get_filters_from_extras_having() -> None:
@@ -216,12 +215,7 @@ def test_get_filters_from_extras_having() -> None:
     result = _get_filters_from_extras(extras)
 
     assert result == {
-        Filter(
-            type=PredicateType.HAVING,
-            column=None,
-            operator=Operator.ADHOC,
-            value="SUM(sales) > 1000",
-        ),
+        AdhocFilter(type=PredicateType.HAVING, definition="SUM(sales) > 1000"),
     }
 
 
@@ -236,18 +230,8 @@ def test_get_filters_from_extras_both() -> None:
     result = _get_filters_from_extras(extras)
 
     assert result == {
-        Filter(
-            type=PredicateType.WHERE,
-            column=None,
-            operator=Operator.ADHOC,
-            value="region = 'US'",
-        ),
-        Filter(
-            type=PredicateType.HAVING,
-            column=None,
-            operator=Operator.ADHOC,
-            value="COUNT(*) > 10",
-        ),
+        AdhocFilter(type=PredicateType.WHERE, definition="region = 'US'"),
+        AdhocFilter(type=PredicateType.HAVING, definition="COUNT(*) > 10"),
     }
 
 
@@ -466,11 +450,9 @@ def test_get_filters_from_query_object_with_extras(mock_datasource: MagicMock) -
             operator=Operator.LESS_THAN,
             value=datetime(2025, 10, 22),
         ),
-        Filter(
+        AdhocFilter(
             type=PredicateType.WHERE,
-            column=None,
-            operator=Operator.ADHOC,
-            value="customer_id > 100",
+            definition="customer_id > 100",
         ),
     }
 
@@ -512,11 +494,9 @@ def test_get_filters_from_query_object_with_fetch_values(
             operator=Operator.LESS_THAN,
             value=datetime(2025, 10, 22),
         ),
-        Filter(
+        AdhocFilter(
             type=PredicateType.WHERE,
-            column=None,
-            operator=Operator.ADHOC,
-            value="tenant_id = 123",
+            definition="tenant_id = 123",
         ),
     }
 
@@ -816,11 +796,9 @@ def test_get_group_limit_filters_with_extras(mock_datasource: MagicMock) -> None
             operator=Operator.LESS_THAN,
             value=datetime(2025, 10, 22),
         ),
-        Filter(
+        AdhocFilter(
             type=PredicateType.WHERE,
-            column=None,
-            operator=Operator.ADHOC,
-            value="customer_id > 100",
+            definition="customer_id > 100",
         ),
     }
 
@@ -2041,11 +2019,9 @@ def test_get_group_limit_filters_with_fetch_values_predicate(
 
     assert result is not None
     assert (
-        Filter(
+        AdhocFilter(
             type=PredicateType.WHERE,
-            column=None,
-            operator=Operator.ADHOC,
-            value="tenant_id = 123",
+            definition="tenant_id = 123",
         )
         in result
     )
@@ -2396,7 +2372,6 @@ def test_get_filters_from_query_object_with_filter_loop(
         f
         for f in result
         if isinstance(f, Filter)
-        and f.column
         and f.column.name == "category"
         and f.operator == Operator.EQUALS
     ]
@@ -2469,7 +2444,6 @@ def test_get_group_limit_filters_with_filter_loop(
         f
         for f in result
         if isinstance(f, Filter)
-        and f.column
         and f.column.name == "category"
         and f.operator == Operator.EQUALS
     ]
@@ -2581,7 +2555,6 @@ def test_get_filters_from_query_object_filter_returns_none(
         f
         for f in result
         if isinstance(f, Filter)
-        and f.column
         and f.column.name == "category"
         and f.operator == Operator.EQUALS
     ]
@@ -2634,7 +2607,6 @@ def test_get_group_limit_filters_filter_returns_none(
         f
         for f in result
         if isinstance(f, Filter)
-        and f.column
         and f.column.name == "category"
         and f.operator == Operator.EQUALS
     ]