Merge pull request #1857 from weaviate/multi2multivec-weaviate

add multi2multivec-weaviate vectorizer

Author: databyjp

test/collection/test_config.py

@@ -19,7 +19,6 @@
     Multi2VecField,
     VectorDistances,
 )
-
 from weaviate.collections.classes.config_vectors import _VectorConfigCreate
 
 DEFAULTS = {
@@ -2173,6 +2172,24 @@ def test_config_with_named_vectors(
             }
         },
     ),
+    (
+        [Configure.MultiVectors.multi2vec_weaviate(name="test", image_field="prop")],
+        {
+            "test": {
+                "vectorizer": {
+                    "multi2multivec-weaviate": {
+                        "imageFields": ["prop"],
+                    }
+                },
+                "vectorIndexConfig": {
+                    "multivector": {
+                        "enabled": True,
+                    },
+                },
+                "vectorIndexType": "hnsw",
+            }
+        },
+    ),
     (
         [Configure.Vectors.text2vec_gpt4all(name="test", source_properties=["prop"])],
         {

weaviate/collections/classes/config_vectorizers.py

@@ -78,6 +78,7 @@
 WeaviateModel: TypeAlias = Literal[
     "Snowflake/snowflake-arctic-embed-l-v2.0", "Snowflake/snowflake-arctic-embed-m-v1.5"
 ]
+WeaviateMultimodalModel: TypeAlias = Literal["ModernVBERT/colmodernvbert"]
 
 
 class Vectorizers(str, Enum):
@@ -134,6 +135,7 @@ class Vectorizers(str, Enum):
     MULTI2VEC_COHERE = "multi2vec-cohere"
     MULTI2VEC_JINAAI = "multi2vec-jinaai"
     MULTI2MULTI_JINAAI = "multi2multivec-jinaai"
+    MULTI2MULTI_WEAVIATE = "multi2multivec-weaviate"
     MULTI2VEC_BIND = "multi2vec-bind"
     MULTI2VEC_PALM = "multi2vec-palm"  # change to google once 1.27 is the lowest supported version
     MULTI2VEC_VOYAGEAI = "multi2vec-voyageai"
@@ -516,6 +518,20 @@ def _to_dict(self) -> Dict[str, Any]:
         return ret_dict
 
 
+class _Multi2MultiVecWeaviateConfig(_Multi2VecBase):
+    vectorizer: Union[Vectorizers, _EnumLikeStr] = Field(
+        default=Vectorizers.MULTI2MULTI_WEAVIATE, frozen=True, exclude=True
+    )
+    baseURL: Optional[AnyHttpUrl]
+    model: Optional[str]
+
+    def _to_dict(self) -> Dict[str, Any]:
+        ret_dict = super()._to_dict()
+        if self.baseURL is not None:
+            ret_dict["baseURL"] = self.baseURL.unicode_string()
+        return ret_dict
+
+
 class _Multi2VecClipConfig(_Multi2VecBase):
     vectorizer: Union[Vectorizers, _EnumLikeStr] = Field(
         default=Vectorizers.MULTI2VEC_CLIP, frozen=True, exclude=True

weaviate/collections/classes/config_vectors.py

@@ -37,9 +37,11 @@
     VoyageModel,
     VoyageMultimodalModel,
     WeaviateModel,
+    WeaviateMultimodalModel,
     _Img2VecNeuralConfig,
     _map_multi2vec_fields,
     _Multi2MultiVecJinaConfig,
+    _Multi2MultiVecWeaviateConfig,
     _Multi2VecAWSConfig,
     _Multi2VecBindConfig,
     _Multi2VecClipConfig,
@@ -288,6 +290,43 @@ def multi2vec_jinaai(
             ),
         )
 
+    @staticmethod
+    def multi2vec_weaviate(
+        *,
+        image_field: str,
+        name: Optional[str] = None,
+        encoding: Optional[_MultiVectorEncodingConfigCreate] = None,
+        quantizer: Optional[_QuantizerConfigCreate] = None,
+        base_url: Optional[AnyHttpUrl] = None,
+        model: Optional[Union[WeaviateMultimodalModel, str]] = None,
+        multi_vector_config: Optional[_MultiVectorConfigCreate] = None,
+        vector_index_config: Optional[_VectorIndexConfigCreate] = None,
+    ) -> _VectorConfigCreate:
+        """Create a vector using the `multi2multivec-weaviate` module.
+
+        Args:
+            image_field: The image field to use in vectorization.
+            name: The name of the vector.
+            encoding: The type of multi-vector encoding to use in the vector index. Defaults to `None`, which uses the server-defined default.
+            quantizer: The quantizer to use for the vector index. If not provided, no quantization will be applied.
+            base_url: The base URL to use where API requests should go. Defaults to `None`, which uses the server-defined default.
+            model: The model to use. Defaults to `None`, which uses the server-defined default.
+            multi_vector_config: The configuration for the multi-vector index. Use `wvc.config.Configure.VectorIndex.MultiVector` to create a multi-vector configuration. None by default
+            vector_index_config: The configuration for Weaviate's vector index. Use `wvc.config.Configure.VectorIndex` to create a vector index configuration. None by default
+        """
+        return _VectorConfigCreate(
+            name=name,
+            vectorizer=_Multi2MultiVecWeaviateConfig(
+                baseURL=base_url,
+                model=model,
+                imageFields=_map_multi2vec_fields([image_field]),
+                textFields=None,
+            ),
+            vector_index_config=_IndexWrappers.multi(
+                vector_index_config, quantizer, multi_vector_config, encoding
+            ),
+        )
+
 
 class _Vectors:
     @staticmethod