Add support for acorn

.github/workflows/main.yaml

@@ -18,7 +18,7 @@ env:
   WEAVIATE_124: preview-rename-object-deletion-to-deletion-strategy-727ecf7
   WEAVIATE_125: stable-v1.25-d9f6cbb
   WEAVIATE_126: stable-v1.26-6a411a4
-  WEAVIATE_127: preview-merge-stable-v1-27-into-main-33adc57
+  WEAVIATE_127: stable-v1.27-0694720
 
 jobs:
   lint-and-format:

integration/test_collection_config.py

@@ -269,6 +269,7 @@ def test_collection_config_full(collection_factory: CollectionFactory) -> None:
             dynamic_ef_min=10,
             ef=-2,
             ef_construction=100,
+            filter_strategy=wvc.config.FilterStrategyHNSW.ACORN,
             flat_search_cutoff=41000,
             max_connections=72,
             quantizer=Configure.VectorIndex.Quantizer.pq(
@@ -375,6 +376,11 @@ def test_collection_config_full(collection_factory: CollectionFactory) -> None:
     assert config.vector_index_config.quantizer.training_limit == 1000001
     assert config.vector_index_config.skip is False
     assert config.vector_index_config.vector_cache_max_objects == 100000
+    if collection._connection._weaviate_version.is_at_least(1, 27, 0):
+        assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.ACORN
+    else:
+        # default value if not present in schema
+        assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.SWEEPING
 
     assert config.vector_index_type == VectorIndexType.HNSW
 
@@ -399,6 +405,8 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
     assert config.multi_tenancy_config.auto_tenant_activation is False
     assert config.multi_tenancy_config.auto_tenant_creation is False
 
+    assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.SWEEPING
+
     collection.config.update(
         description="Test",
         inverted_index_config=Reconfigure.inverted_index(
@@ -414,6 +422,7 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
         ),  # currently not updateable in RAFT
         vectorizer_config=Reconfigure.VectorIndex.hnsw(
             vector_cache_max_objects=2000000,
+            filter_strategy=wvc.config.FilterStrategyHNSW.ACORN,
             quantizer=Reconfigure.VectorIndex.Quantizer.pq(
                 centroids=128,
                 encoder_type=PQEncoderType.TILE,
@@ -472,6 +481,12 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
     assert config.vector_index_config.skip is False
     assert config.vector_index_config.vector_cache_max_objects == 2000000
 
+    if collection._connection._weaviate_version.is_at_least(1, 27, 0):
+        assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.ACORN
+    else:
+        # default value if not present in schema
+        assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.SWEEPING
+
     assert config.vector_index_type == VectorIndexType.HNSW
 
     assert config.multi_tenancy_config.enabled is True
@@ -488,6 +503,7 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
 
     collection.config.update(
         vectorizer_config=Reconfigure.VectorIndex.hnsw(
+            filter_strategy=wvc.config.FilterStrategyHNSW.SWEEPING,
             quantizer=Reconfigure.VectorIndex.Quantizer.pq(enabled=False),
         )
     )
@@ -529,6 +545,7 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
     assert config.vector_index_config.quantizer is None
     assert config.vector_index_config.skip is False
     assert config.vector_index_config.vector_cache_max_objects == 2000000
+    assert config.vector_index_config.filter_strategy == wvc.config.FilterStrategyHNSW.SWEEPING
 
     assert config.vector_index_type == VectorIndexType.HNSW
 

mock_tests/test_collection.py

@@ -221,7 +221,7 @@ def test_return_from_bind_module(
     prop_modconf: Dict[str, Any] = {"multi2vec-bind": {}}
 
     hnsw_config = config.VectorIndex.hnsw(
-        1, VectorDistances.COSINE, 1, 1, 1, 1, 1, 1, 1, 1
+        1, VectorDistances.COSINE, 1, 1, 1, 1, 1, None, 1, 1, 1
     )._to_dict()
     hnsw_config["skip"] = True
     ii_config = config.inverted_index(

weaviate/classes/config.py

@@ -12,17 +12,17 @@
     Tokenization,
     VectorDistances,
 )
-
+from weaviate.collections.classes.config_vector_index import FilterStrategyHNSW
 from weaviate.collections.classes.config_vectorizers import Multi2VecField, Vectorizers
 from weaviate.connect.integrations import Integrations
 
-
 __all__ = [
     "Configure",
     "ConsistencyLevel",
     "Reconfigure",
     "DataType",
     "GenerativeSearches",
+    "FilterStrategyHNSW",
     "Integrations",
     "Multi2VecField",
     "DeletionStrategy",

weaviate/collections/classes/config.py

@@ -30,7 +30,10 @@
     _NamedVectors,
     _NamedVectorsUpdate,
 )
-from weaviate.collections.classes.config_vector_index import VectorIndexType as VectorIndexTypeAlias
+from weaviate.collections.classes.config_vector_index import (
+    VectorIndexType as VectorIndexTypeAlias,
+    FilterStrategyHNSW,
+)
 from weaviate.collections.classes.config_vector_index import (
     _QuantizerConfigCreate,
     _VectorIndexConfigCreate,
@@ -1533,6 +1536,7 @@ class _VectorIndexConfigHNSW(_VectorIndexConfig):
     dynamic_ef_factor: int
     ef: int
     ef_construction: int
+    filter_strategy: FilterStrategyHNSW
     flat_search_cutoff: int
     max_connections: int
     skip: bool
@@ -2064,6 +2068,7 @@ def hnsw(
         dynamic_ef_min: Optional[int] = None,
         ef: Optional[int] = None,
         ef_construction: Optional[int] = None,
+        filter_strategy: Optional[FilterStrategyHNSW] = None,
         flat_search_cutoff: Optional[int] = None,
         max_connections: Optional[int] = None,
         vector_cache_max_objects: Optional[int] = None,
@@ -2084,6 +2089,7 @@ def hnsw(
             dynamicEfFactor=dynamic_ef_factor,
             efConstruction=ef_construction,
             ef=ef,
+            filterStrategy=filter_strategy,
             flatSearchCutoff=flat_search_cutoff,
             maxConnections=max_connections,
             vectorCacheMaxObjects=vector_cache_max_objects,
@@ -2335,6 +2341,7 @@ def hnsw(
         dynamic_ef_max: Optional[int] = None,
         ef: Optional[int] = None,
         flat_search_cutoff: Optional[int] = None,
+        filter_strategy: Optional[FilterStrategyHNSW] = None,
         vector_cache_max_objects: Optional[int] = None,
         quantizer: Optional[Union[_PQConfigUpdate, _BQConfigUpdate, _SQConfigUpdate]] = None,
     ) -> _VectorIndexConfigHNSWUpdate:
@@ -2350,6 +2357,7 @@ def hnsw(
             dynamicEfMax=dynamic_ef_max,
             dynamicEfFactor=dynamic_ef_factor,
             ef=ef,
+            filterStrategy=filter_strategy,
             flatSearchCutoff=flat_search_cutoff,
             vectorCacheMaxObjects=vector_cache_max_objects,
             quantizer=quantizer,

weaviate/collections/classes/config_methods.py

@@ -36,6 +36,7 @@
     Rerankers,
     _NestedProperty,
     DeletionStrategy,
+    FilterStrategyHNSW,
 )
 
 
@@ -157,6 +158,11 @@ def __get_hnsw_config(config: Dict[str, Any]) -> _VectorIndexConfigHNSW:
         dynamic_ef_factor=config["dynamicEfFactor"],
         ef=config["ef"],
         ef_construction=config["efConstruction"],
+        filter_strategy=(
+            FilterStrategyHNSW(config["filterStrategy"])
+            if "filterStrategy" in config
+            else FilterStrategyHNSW.SWEEPING
+        ),
         flat_search_cutoff=config["flatSearchCutoff"],
         max_connections=config["maxConnections"],
         quantizer=quantizer,

weaviate/collections/classes/config_vector_index.py

@@ -1,6 +1,7 @@
 from abc import abstractmethod
 from enum import Enum
 from typing import Any, Dict, Optional
+
 from pydantic import Field
 
 from weaviate.collections.classes.config_base import (
@@ -9,10 +10,21 @@
     _QuantizerConfigCreate,
     _QuantizerConfigUpdate,
 )
-
 from weaviate.collections.classes.config_vectorizers import VectorDistances
 
 
+class FilterStrategyHNSW(str, Enum):
+    """Set the strategy when doing a filtered HNSW search.
+
+    Attributes:
+        SWEEPING: Do normal ANN search and skip nodes.
+        ACORN: Multi-hop search to find new candidates matching the filter.
+    """
+
+    SWEEPING = "sweeping"
+    ACORN = "acorn"
+
+
 class VectorIndexType(str, Enum):
     """The available vector index types in Weaviate.
 
@@ -67,6 +79,7 @@ class _VectorIndexConfigHNSWCreate(_VectorIndexConfigCreate):
     dynamicEfFactor: Optional[int]
     efConstruction: Optional[int]
     ef: Optional[int]
+    filterStrategy: Optional[FilterStrategyHNSW]
     flatSearchCutoff: Optional[int]
     maxConnections: Optional[int]
     vectorCacheMaxObjects: Optional[int]
@@ -89,6 +102,7 @@ class _VectorIndexConfigHNSWUpdate(_VectorIndexConfigUpdate):
     dynamicEfMax: Optional[int]
     dynamicEfFactor: Optional[int]
     ef: Optional[int]
+    filterStrategy: Optional[FilterStrategyHNSW]
     flatSearchCutoff: Optional[int]
     vectorCacheMaxObjects: Optional[int]