remove epsilon parameter from AggregateHybridQuery

Author: nkanu17

docs/api/query.rst

@@ -28,11 +28,9 @@ VectorQuery
    **HNSW Parameters:**
 
    - ``ef_runtime``: Controls search accuracy (higher = better recall, slower search)
-   - ``epsilon``: Range search approximation factor (for range queries)
 
    **SVS-VAMANA Parameters:**
 
-   - ``epsilon``: Range search approximation factor
    - ``search_window_size``: Size of search window for KNN searches
    - ``use_search_history``: Whether to use search buffer (OFF/ON/AUTO)
    - ``search_buffer_capacity``: Tuning parameter for 2-level compression
@@ -47,8 +45,7 @@ VectorQuery
           vector=[0.1, 0.2, 0.3],
           vector_field_name="embedding",
           num_results=10,
-          ef_runtime=150,  # Higher for better recall
-          epsilon=0.05     # For range search approximation
+          ef_runtime=150  # Higher for better recall
       )
 
    Example with SVS-VAMANA runtime parameters:
@@ -61,8 +58,7 @@ VectorQuery
           num_results=10,
           search_window_size=20,
           use_search_history='ON',
-          search_buffer_capacity=30,
-          epsilon=0.01
+          search_buffer_capacity=30
       )
 
 
@@ -130,21 +126,21 @@ HybridQuery
 .. note::
    **Runtime Parameters for Hybrid Queries**
 
-   AggregateHybridQuery (and the deprecated HybridQuery) support runtime parameters for the vector search component:
+   AggregateHybridQuery (and the deprecated HybridQuery) support runtime parameters for the vector search component.
+
+   **Note:** The ``epsilon`` parameter is only supported for range queries (VectorRangeQuery), not for KNN queries used in hybrid search.
 
    **HNSW Parameters:**
 
    - ``ef_runtime``: Controls search accuracy for the vector component
-   - ``epsilon``: Range search approximation factor
 
    **SVS-VAMANA Parameters:**
 
-   - ``epsilon``: Range search approximation factor
    - ``search_window_size``: Size of search window for KNN searches
    - ``use_search_history``: Whether to use search buffer (OFF/ON/AUTO)
    - ``search_buffer_capacity``: Tuning parameter for 2-level compression
 
-   Example:
+   Example with HNSW:
 
    .. code-block:: python
 
@@ -156,10 +152,22 @@ HybridQuery
           vector=[0.1, 0.2, 0.3],
           vector_field_name="embedding",
           alpha=0.7,
-          ef_runtime=150,           # HNSW parameter
-          epsilon=0.05,             # HNSW & SVS-VAMANA
-          search_window_size=20,    # SVS-VAMANA only
-          use_search_history='ON'   # SVS-VAMANA only
+          ef_runtime=150  # HNSW parameter
+      )
+
+   Example with SVS-VAMANA:
+
+   .. code-block:: python
+
+      query = AggregateHybridQuery(
+          text="search query",
+          text_field_name="description",
+          vector=[0.1, 0.2, 0.3],
+          vector_field_name="embedding",
+          alpha=0.7,
+          search_window_size=20,    # SVS-VAMANA
+          use_search_history='ON',  # SVS-VAMANA
+          search_buffer_capacity=30 # SVS-VAMANA
       )
 
 

docs/user_guide/11_advanced_queries.ipynb

@@ -876,7 +876,6 @@
     "    vector_field_name=\"text_embedding\",\n",
     "    alpha=0.7,\n",
     "    ef_runtime=150,  # Higher for better recall on HNSW indexes\n",
-    "    epsilon=0.05,    # For range search approximation\n",
     "    return_fields=[\"product_id\", \"brief_description\", \"category\"],\n",
     "    num_results=5\n",
     ")\n",
@@ -907,7 +906,6 @@
     "    vector_field_name=\"text_embedding\",\n",
     "    alpha=0.8,\n",
     "    search_window_size=40,      # Larger window for better recall\n",
-    "    epsilon=0.01,               # Approximation factor\n",
     "    use_search_history='ON',    # Use search history\n",
     "    search_buffer_capacity=50,  # Buffer capacity\n",
     "    return_fields=[\"product_id\", \"brief_description\", \"price\"],\n",

redisvl/query/aggregate.py

@@ -92,7 +92,6 @@ class AggregateHybridQuery(AggregationQuery):
     # HNSW runtime parameters
     EF_RUNTIME: str = "EF_RUNTIME"
     EF_RUNTIME_PARAM: str = "EF"
-    EPSILON_PARAM: str = "EPSILON"
 
     # SVS-VAMANA runtime parameters
     SEARCH_WINDOW_SIZE: str = "SEARCH_WINDOW_SIZE"
@@ -118,7 +117,6 @@ def __init__(
         dialect: int = 2,
         text_weights: Optional[Dict[str, float]] = None,
         ef_runtime: Optional[int] = None,
-        epsilon: Optional[float] = None,
         search_window_size: Optional[int] = None,
         use_search_history: Optional[str] = None,
         search_buffer_capacity: Optional[int] = None,
@@ -157,10 +155,6 @@ def __init__(
             ef_runtime (Optional[int]): The size of the dynamic candidate list for HNSW indexes.
                 Increasing this value generally yields more accurate but slower search results.
                 Defaults to None, which uses the index-defined value (typically 10).
-            epsilon (Optional[float]): The relative factor for vector range queries (HNSW and SVS-VAMANA),
-                setting boundaries for candidates within radius * (1 + epsilon).
-                Higher values increase recall at the expense of performance.
-                Defaults to None, which uses the index-defined epsilon (typically 0.01).
             search_window_size (Optional[int]): The size of the search window for SVS-VAMANA KNN searches.
                 Increasing this value generally yields more accurate but slower search results.
                 Defaults to None, which uses the index-defined value (typically 10).
@@ -192,7 +186,6 @@ def __init__(
         self._dtype = dtype
         self._num_results = num_results
         self._ef_runtime = ef_runtime
-        self._epsilon = epsilon
         self._search_window_size = search_window_size
         self._use_search_history = use_search_history
         self._search_buffer_capacity = search_buffer_capacity
@@ -231,10 +224,6 @@ def params(self) -> Dict[str, Any]:
         if self._ef_runtime is not None:
             params[self.EF_RUNTIME_PARAM] = self._ef_runtime
 
-        # Add EPSILON parameter if specified (HNSW and SVS-VAMANA)
-        if self._epsilon is not None:
-            params[self.EPSILON_PARAM] = self._epsilon
-
         # Add SEARCH_WINDOW_SIZE parameter if specified (SVS-VAMANA)
         if self._search_window_size is not None:
             params[self.SEARCH_WINDOW_SIZE_PARAM] = self._search_window_size
@@ -374,10 +363,6 @@ def _build_query_string(self) -> str:
         if self._ef_runtime is not None:
             knn_query += f" {self.EF_RUNTIME} ${self.EF_RUNTIME_PARAM}"
 
-        # Add EPSILON parameter if specified (HNSW and SVS-VAMANA)
-        if self._epsilon is not None:
-            knn_query += f" EPSILON ${self.EPSILON_PARAM}"
-
         # Add SEARCH_WINDOW_SIZE parameter if specified (SVS-VAMANA)
         if self._search_window_size is not None:
             knn_query += f" {self.SEARCH_WINDOW_SIZE} ${self.SEARCH_WINDOW_SIZE_PARAM}"

tests/unit/test_aggregation_types.py

@@ -267,27 +267,6 @@ def test_aggregate_hybrid_query_ef_runtime():
     assert query.params.get("EF") == 100
 
 
-def test_aggregate_hybrid_query_epsilon():
-    """Test that AggregateHybridQuery correctly handles epsilon parameter."""
-    query = AggregateHybridQuery(
-        text=sample_text,
-        text_field_name="description",
-        vector=sample_vector,
-        vector_field_name="embedding",
-        epsilon=0.05,
-    )
-
-    # Check properties
-    assert query._epsilon == 0.05
-
-    # Check query string
-    query_string = str(query)
-    assert "EPSILON $EPSILON" in query_string
-
-    # Check params dictionary
-    assert query.params.get("EPSILON") == 0.05
-
-
 def test_aggregate_hybrid_query_search_window_size():
     """Test that AggregateHybridQuery correctly handles search_window_size parameter (SVS-VAMANA)."""
     query = AggregateHybridQuery(
@@ -360,31 +339,27 @@ def test_aggregate_hybrid_query_all_runtime_params():
         vector=sample_vector,
         vector_field_name="embedding",
         ef_runtime=100,
-        epsilon=0.05,
         search_window_size=40,
         use_search_history="ON",
         search_buffer_capacity=50,
     )
 
     # Check all properties
     assert query._ef_runtime == 100
-    assert query._epsilon == 0.05
     assert query._search_window_size == 40
     assert query._use_search_history == "ON"
     assert query._search_buffer_capacity == 50
 
     # Check query string contains all parameters
     query_string = str(query)
     assert "EF_RUNTIME $EF" in query_string
-    assert "EPSILON $EPSILON" in query_string
     assert "SEARCH_WINDOW_SIZE $SEARCH_WINDOW_SIZE" in query_string
     assert "USE_SEARCH_HISTORY $USE_SEARCH_HISTORY" in query_string
     assert "SEARCH_BUFFER_CAPACITY $SEARCH_BUFFER_CAPACITY" in query_string
 
     # Check params dictionary contains all parameters
     params = query.params
     assert params["EF"] == 100
-    assert params["EPSILON"] == 0.05
     assert params["SEARCH_WINDOW_SIZE"] == 40
     assert params["USE_SEARCH_HISTORY"] == "ON"
     assert params["SEARCH_BUFFER_CAPACITY"] == 50