Merge pull request #1669 from weaviate/revert-1635-rob/muvera

Revert "Add muvera config"

Author: dirkkul

integration/test_collection_config.py

@@ -1405,67 +1405,6 @@ def test_config_multi_vector_disabled(
     assert conf.multi_vector is None
 
 
-def test_config_muvera_enabled(
-    collection_factory: CollectionFactory,
-) -> None:
-    dummy = collection_factory("dummy", ports=(8086, 50057))
-    if dummy._connection._weaviate_version.is_lower_than(1, 31, 0):
-        pytest.skip("Muvera is not supported in Weaviate versions lower than 1.31.0")
-
-    collection = collection_factory(
-        ports=(8086, 50057),
-        properties=[Property(name="name", data_type=DataType.TEXT)],
-        vectorizer_config=[
-            Configure.NamedVectors.text2colbert_jinaai(
-                name="vec",
-                vectorize_collection_name=False,
-                vector_index_config=Configure.VectorIndex.hnsw(
-                    multi_vector=Configure.VectorIndex.MultiVector.multi_vector(
-                        encoding=Configure.VectorIndex.MultiVector.Encoding.muvera()
-                    )
-                ),
-            )
-        ],
-    )
-    config = collection.config.get()
-    assert config.vector_config is not None
-    conf = config.vector_config["vec"].vector_index_config
-    assert isinstance(conf, _VectorIndexConfigHNSW)
-    if collection._connection._weaviate_version.is_lower_than(1, 31, 0):
-        assert conf.multi_vector is None
-    else:
-        assert conf.multi_vector is not None
-        assert conf.multi_vector.encoding is not None
-
-
-def test_config_muvera_disabled(
-    collection_factory: CollectionFactory,
-) -> None:
-    dummy = collection_factory("dummy", ports=(8086, 50057))
-    if dummy._connection._weaviate_version.is_lower_than(1, 29, 0):
-        pytest.skip("Multivector is not supported in Weaviate versions lower than 1.29.0")
-
-    collection = collection_factory(
-        ports=(8086, 50057),
-        properties=[Property(name="name", data_type=DataType.TEXT)],
-        vectorizer_config=[
-            Configure.NamedVectors.text2colbert_jinaai(
-                name="vec",
-                vectorize_collection_name=False,
-                vector_index_config=Configure.VectorIndex.hnsw(
-                    multi_vector=Configure.VectorIndex.MultiVector.multi_vector()
-                ),
-            )
-        ],
-    )
-    config = collection.config.get()
-    assert config.vector_config is not None
-    conf = config.vector_config["vec"].vector_index_config
-    assert isinstance(conf, _VectorIndexConfigHNSW)
-    assert conf.multi_vector is not None
-    assert conf.multi_vector.encoding is None
-
-
 @pytest.mark.parametrize(
     "generative_config",
     [

weaviate/collections/classes/config.py

@@ -33,8 +33,6 @@
 )
 from weaviate.collections.classes.config_vector_index import (
     VectorFilterStrategy,
-    _MuveraConfigCreate,
-    _EncodingConfigCreate,
     _MultiVectorConfigCreate,
     _QuantizerConfigCreate,
     _VectorIndexConfigCreate,
@@ -1558,20 +1556,8 @@ class _SQConfig(_ConfigBase):
 SQConfig = _SQConfig
 
 
-@dataclass
-class _MuveraConfig(_ConfigBase):
-    enabled: Optional[bool]
-    ksim: Optional[int]
-    dprojections: Optional[int]
-    repetitions: Optional[int]
-
-
-MuveraConfig = _MuveraConfig
-
-
 @dataclass
 class _MultiVectorConfig(_ConfigBase):
-    encoding: Optional[_MuveraConfig]
     aggregation: str
 
 
@@ -2045,31 +2031,12 @@ def __add_props(
         ret_dict["properties"] = existing_props
 
 
-class _VectorIndexMultivectorEncoding:
-    @staticmethod
-    def muvera(
-        ksim: Optional[int] = None,
-        dprojections: Optional[int] = None,
-        repetitions: Optional[int] = None,
-    ) -> _EncodingConfigCreate:
-        return _MuveraConfigCreate(
-            enabled=True,
-            ksim=ksim,
-            dprojections=dprojections,
-            repetitions=repetitions,
-        )
-
-
 class _VectorIndexMultiVector:
-    Encoding = _VectorIndexMultivectorEncoding
-
     @staticmethod
     def multi_vector(
-        encoding: Optional[_EncodingConfigCreate] = None,
         aggregation: Optional[MultiVectorAggregation] = None,
     ) -> _MultiVectorConfigCreate:
         return _MultiVectorConfigCreate(
-            encoding=encoding if encoding is not None else None,
             aggregation=aggregation.value if aggregation is not None else None,
         )
 

weaviate/collections/classes/config_methods.py

@@ -21,7 +21,6 @@
     _InvertedIndexConfig,
     _MultiTenancyConfig,
     _MultiVectorConfig,
-    _MuveraConfig,
     _NamedVectorConfig,
     _NamedVectorizerConfig,
     _NestedProperty,
@@ -150,31 +149,12 @@ def __get_quantizer_config(
     return quantizer
 
 
-def __get_multivector_encoding(config: Dict[str, Any]) -> Optional[_MuveraConfig]:
-    return (
-        None
-        if config.get("muvera") is None
-        or not config.get("muvera", {"enabled": False}).get("enabled")
-        else _MuveraConfig(
-            enabled=config["muvera"]["enabled"],
-            ksim=config["muvera"]["ksim"],
-            dprojections=config["muvera"]["dprojections"],
-            repetitions=config["muvera"]["repetitions"],
-        )
-    )
-
-
 def __get_multivector(config: Dict[str, Any]) -> Optional[_MultiVectorConfig]:
     return (
         None
         if config.get("multivector") is None
         or not config.get("multivector", {"enabled": False}).get("enabled")
         else _MultiVectorConfig(
-            encoding=(
-                None
-                if config["multivector"].get("muvera") is None
-                else __get_multivector_encoding(config["multivector"])
-            ),
             aggregation=config["multivector"]["aggregation"],
         )
     )
@@ -264,7 +244,6 @@ def __get_vector_config(
                 ),
                 vector_index_config=vector_index_config,
             )
-
         return named_vectors
     else:
         return None

weaviate/collections/classes/config_vector_index.py

@@ -42,26 +42,7 @@ class _MultiVectorConfigCreateBase(_ConfigCreateModel):
     enabled: bool = Field(default=True)
 
 
-class _EncodingConfigCreate(_MultiVectorConfigCreateBase):
-    enabled: bool = Field(default=True)
-
-    @staticmethod
-    @abstractmethod
-    def encoding_name() -> str: ...
-
-
-class _MuveraConfigCreate(_EncodingConfigCreate):
-    ksim: Optional[int]
-    dprojections: Optional[int]
-    repetitions: Optional[int]
-
-    @staticmethod
-    def encoding_name() -> str:
-        return "muvera"
-
-
 class _MultiVectorConfigCreate(_MultiVectorConfigCreateBase):
-    encoding: Optional[_EncodingConfigCreate] = Field(exclude=True)
     aggregation: Optional[str]
 
 
@@ -80,10 +61,6 @@ def _to_dict(self) -> Dict[str, Any]:
             ret_dict[self.quantizer.quantizer_name()] = self.quantizer._to_dict()
         if self.distance is not None:
             ret_dict["distance"] = str(self.distance.value)
-        if self.multivector is not None and self.multivector.encoding is not None:
-            ret_dict["multivector"][self.multivector.encoding.encoding_name()] = (
-                self.multivector.encoding._to_dict()
-            )
 
         return ret_dict