Merge pull request #1082 from weaviate/auto-tenant-activation

Add support for auto-tenant-activation

.github/workflows/main.yaml

@@ -16,7 +16,7 @@ on:
 env:
   WEAVIATE_123: 1.23.14
   WEAVIATE_124: 1.24.10
-  WEAVIATE_125: preview--grpc-clean-up-target-vectors-in-hybrid-30c49be
+  WEAVIATE_125: preview--c57fecf
 
 
 jobs:

.pre-commit-config.yaml

@@ -49,4 +49,4 @@ repos:
       language: node
       pass_filenames: false
       types: [python]
-      additional_dependencies: [pyright@1.1.347]
+      additional_dependencies: [pyright@1.1.364]

integration/test_collection_config.py

@@ -238,7 +238,9 @@ def test_collection_config_full(collection_factory: CollectionFactory) -> None:
             stopwords_preset=StopwordsPreset.EN,
             stopwords_removals=["the"],
         ),
-        multi_tenancy_config=Configure.multi_tenancy(enabled=True),
+        multi_tenancy_config=Configure.multi_tenancy(
+            enabled=True, auto_tenant_activation=True, auto_tenant_creation=True
+        ),
         # replication_config=Configure.replication(factor=2), # currently not updateable in RAFT
         vector_index_config=Configure.VectorIndex.hnsw(
             cleanup_interval_seconds=10,
@@ -308,6 +310,11 @@ def test_collection_config_full(collection_factory: CollectionFactory) -> None:
     assert config.inverted_index_config.stopwords.removals == ["the"]
 
     assert config.multi_tenancy_config.enabled is True
+    if collection._connection._weaviate_version.is_at_least(1, 25, 0):
+        assert config.multi_tenancy_config.auto_tenant_activation is True
+    # change to 1.25.2 after it is out
+    if collection._connection._weaviate_version.is_at_least(1, 25, patch=1):
+        assert config.multi_tenancy_config.auto_tenant_creation is True
 
     # assert config.replication_config.factor == 2
 
@@ -342,10 +349,19 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
             Property(name="age", data_type=DataType.INT),
         ],
         ports=(8087, 50058),
+        multi_tenancy_config=Configure.multi_tenancy(
+            enabled=True, auto_tenant_creation=False, auto_tenant_activation=False
+        ),
     )
     config = collection.config.get()
 
     assert config.replication_config.factor == 1
+    assert config.multi_tenancy_config.enabled is True
+    if collection._connection._weaviate_version.is_at_least(1, 25, 0):
+        assert config.multi_tenancy_config.auto_tenant_activation is False
+    # change to 1.25.2 after it is out
+    if collection._connection._weaviate_version.is_at_least(1, 25, patch=1):
+        assert config.multi_tenancy_config.auto_tenant_creation is False
 
     collection.config.update(
         description="Test",
@@ -368,6 +384,9 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
                 training_limit=100001,
             ),
         ),
+        multi_tenancy_config=Reconfigure.multi_tenancy(
+            auto_tenant_creation=True, auto_tenant_activation=True
+        ),
     )
 
     config = collection.config.get()
@@ -404,6 +423,13 @@ def test_collection_config_update(collection_factory: CollectionFactory) -> None
 
     assert config.vector_index_type == VectorIndexType.HNSW
 
+    assert config.multi_tenancy_config.enabled is True
+    if collection._connection._weaviate_version.is_at_least(1, 25, 0):
+        assert config.multi_tenancy_config.auto_tenant_activation is True
+    # change to 1.25.2 after it is out
+    if collection._connection._weaviate_version.is_at_least(1, 25, patch=1):
+        assert config.multi_tenancy_config.auto_tenant_creation is True
+
     collection.config.update(
         vectorizer_config=Reconfigure.VectorIndex.hnsw(
             quantizer=Reconfigure.VectorIndex.Quantizer.pq(enabled=False),

mock_tests/test_collection.py

@@ -165,7 +165,9 @@ def test_missing_multi_tenancy_config(
             index_timestamps=False,
             stopwords=StopwordsConfig(preset=StopwordsPreset.NONE, additions=[], removals=[]),
         ),
-        multi_tenancy_config=MultiTenancyConfig(enabled=True, auto_tenant_creation=False),
+        multi_tenancy_config=MultiTenancyConfig(
+            enabled=True, auto_tenant_creation=False, auto_tenant_activation=False
+        ),
         sharding_config=ShardingConfig(
             virtual_per_physical=0,
             desired_count=0,

weaviate/collections/classes/config.py

@@ -14,26 +14,21 @@
     cast,
 )
 
-from typing_extensions import TypeAlias
-
 from pydantic import AnyHttpUrl, Field, field_validator
-
-from weaviate.util import _capitalize_first_letter
-from weaviate.collections.classes.config_vectorizers import (
-    _Vectorizer,
-    _VectorizerConfigCreate,
-    CohereModel,
-    Vectorizers as VectorizersAlias,
-    VectorDistances as VectorDistancesAlias,
-)
+from typing_extensions import TypeAlias
 
 from weaviate.collections.classes.config_base import (
     _ConfigBase,
     _ConfigCreateModel,
     _ConfigUpdateModel,
     _QuantizerConfigUpdate,
 )
-
+from weaviate.collections.classes.config_named_vectors import (
+    _NamedVectorConfigCreate,
+    _NamedVectorConfigUpdate,
+    _NamedVectors,
+    _NamedVectorsUpdate,
+)
 from weaviate.collections.classes.config_vector_index import (
     _QuantizerConfigCreate,
     _VectorIndexConfigCreate,
@@ -47,14 +42,15 @@
     _VectorIndexConfigUpdate,
     VectorIndexType as VectorIndexTypeAlias,
 )
-
-from weaviate.collections.classes.config_named_vectors import (
-    _NamedVectorConfigCreate,
-    _NamedVectorConfigUpdate,
-    _NamedVectors,
-    _NamedVectorsUpdate,
+from weaviate.collections.classes.config_vectorizers import (
+    _Vectorizer,
+    _VectorizerConfigCreate,
+    CohereModel,
+    Vectorizers as VectorizersAlias,
+    VectorDistances as VectorDistancesAlias,
 )
 from weaviate.exceptions import WeaviateInvalidInputError
+from weaviate.util import _capitalize_first_letter
 from weaviate.warnings import _Warnings
 
 # BC for direct imports
@@ -360,10 +356,12 @@ class _InvertedIndexConfigUpdate(_ConfigUpdateModel):
 class _MultiTenancyConfigCreate(_ConfigCreateModel):
     enabled: bool
     autoTenantCreation: Optional[bool]
+    autoTenantActivation: Optional[bool]
 
 
 class _MultiTenancyConfigUpdate(_ConfigUpdateModel):
-    autoTenantCreation: Optional[bool] = None
+    autoTenantCreation: Optional[bool]
+    autoTenantActivation: Optional[bool]
 
 
 class _GenerativeConfigCreate(_ConfigCreateModel):
@@ -952,6 +950,7 @@ class _InvertedIndexConfig(_ConfigBase):
 class _MultiTenancyConfig(_ConfigBase):
     enabled: bool
     auto_tenant_creation: bool
+    auto_tenant_activation: bool
 
 
 MultiTenancyConfig = _MultiTenancyConfig
@@ -1745,17 +1744,25 @@ def inverted_index(
 
     @staticmethod
     def multi_tenancy(
-        enabled: bool = True, auto_tenant_creation: bool = False
+        enabled: bool = True,
+        auto_tenant_creation: Optional[bool] = None,
+        auto_tenant_activation: Optional[bool] = None,
     ) -> _MultiTenancyConfigCreate:
         """Create a `MultiTenancyConfigCreate` object to be used when defining the multi-tenancy configuration of Weaviate.
 
         Arguments:
             `enabled`
                 Whether multi-tenancy is enabled. Defaults to `True`.
             `auto_tenant_creation`
-                Automatically create nonexistent tenants during batch import. Defaults to `False`
+                Automatically create nonexistent tenants during batch import. Defaults to `None`, which uses the server-defined default.
+            `auto_tenant_activation`
+                Automatically turn tenants implicitly HOT when they are accessed. Defaults to `None`, which uses the server-defined default.
         """
-        return _MultiTenancyConfigCreate(enabled=enabled, autoTenantCreation=auto_tenant_creation)
+        return _MultiTenancyConfigCreate(
+            enabled=enabled,
+            autoTenantCreation=auto_tenant_creation,
+            autoTenantActivation=auto_tenant_activation,
+        )
 
     @staticmethod
     def replication(factor: Optional[int] = None) -> _ReplicationConfigCreate:
@@ -1972,13 +1979,19 @@ def replication(factor: Optional[int] = None) -> _ReplicationConfigUpdate:
         return _ReplicationConfigUpdate(factor=factor)
 
     @staticmethod
-    def multi_tenancy(auto_tenant_creation: Optional[bool] = None) -> _MultiTenancyConfigUpdate:
+    def multi_tenancy(
+        auto_tenant_creation: Optional[bool] = None, auto_tenant_activation: Optional[bool] = None
+    ) -> _MultiTenancyConfigUpdate:
         """Create a `MultiTenancyConfigUpdate` object.
 
         Use this method when defining the `multi_tenancy` argument in `collection.update()`.
 
         Arguments:
             `auto_tenant_creation`
                 When set, implicitly creates nonexisting tenants during batch imports
+            `auto_tenant_activation`
+                Automatically turn tenants implicitly HOT when they are accessed. Defaults to `None`, which uses the server-defined default.
         """
-        return _MultiTenancyConfigUpdate(autoTenantCreation=auto_tenant_creation)
+        return _MultiTenancyConfigUpdate(
+            autoTenantCreation=auto_tenant_creation, autoTenantActivation=auto_tenant_activation
+        )

weaviate/collections/classes/config_methods.py

@@ -245,6 +245,9 @@ def _collection_config_from_json(schema: Dict[str, Any]) -> _CollectionConfig:
             auto_tenant_creation=schema.get("multiTenancyConfig", {}).get(
                 "autoTenantCreation", False
             ),
+            auto_tenant_activation=schema.get("multiTenancyConfig", {}).get(
+                "autoTenantActivation", False
+            ),
         ),
         properties=_properties_from_config(schema) if schema.get("properties") is not None else [],
         references=_references_from_config(schema) if schema.get("properties") is not None else [],