Filter iterator

.github/workflows/main.yaml

@@ -24,7 +24,7 @@ env:
   WEAVIATE_130: 1.30.12
   WEAVIATE_131: 1.31.5
   WEAVIATE_132: 1.32.0-rc.1-098c594
-  WEAVIATE_133: preview-remove-support-for-old-clients-a52a482.amd64
+  WEAVIATE_133: 1.33.0-dev-cb09e79.amd64
 
 jobs:
   lint-and-format:

integration/test_client.py

@@ -226,7 +226,7 @@ def test_create_export_and_recreate(client: weaviate.WeaviateClient, request: Su
         vectorizer_config=Configure.Vectorizer.text2vec_contextionary(
             vectorize_collection_name=False
         ),
-        generative_config=Configure.Generative.cohere(model="something", k=10),
+        generative_config=Configure.Generative.cohere(model="command-medium-nightly", k=10),
         properties=[
             Property(
                 name="name",
@@ -270,7 +270,7 @@ def test_create_export_and_recreate(client: weaviate.WeaviateClient, request: Su
 
     assert export.generative_config is not None
     assert export.generative_config.generative == GenerativeSearches.COHERE
-    assert export.generative_config.model["model"] == "something"
+    assert export.generative_config.model["model"] == "command-medium-nightly"
     assert export.generative_config.model["kProperty"] == 10
 
     client.collections.delete([name1, name2])

integration/test_iterator.py

@@ -20,6 +20,7 @@
 )
 from weaviate.collections.iterator import ITERATOR_CACHE_SIZE
 from weaviate.exceptions import WeaviateInvalidInputError
+import weaviate.classes as wvc
 
 
 @pytest.fixture(scope="module")
@@ -220,8 +221,33 @@ def test_iterator_with_after(collection_factory: CollectionFactory) -> None:
     collection.data.insert_many([DataObject(properties={"data": i}) for i in range(10)])
 
     uuids = [obj.uuid for obj in collection.iterator()]
+    obj6 = collection.query.fetch_object_by_id(uuids[6])
     iterator = collection.iterator(after=uuids[5])
-    assert (
-        next(iterator).properties["data"]
-        == collection.query.fetch_object_by_id(uuids[6]).properties["data"]
+    next_object = next(iterator)
+    assert next_object.properties["data"] == obj6.properties["data"]
+
+
+def test_iterator_with_filter(collection_factory: CollectionFactory) -> None:
+    collection = collection_factory(
+        properties=[
+            Property(name="bool", data_type=DataType.BOOL),
+            Property(name="count", data_type=DataType.INT),
+        ],
+        vectorizer_config=Configure.Vectorizer.none(),
+        data_model_properties=Dict[str, int],
+    )
+
+    if collection._connection._weaviate_version.is_lower_than(1, 33, 0):
+        pytest.skip("Iterator with filters requires Weaviate version 1.33 or higher")
+
+    num_objects = 1000
+    res = collection.data.insert_many(
+        [DataObject(properties={"bool": i % 2 == 0, "count": i}) for i in range(num_objects)]
     )
+    assert not res.has_errors
+
+    count = 0
+    for obj in collection.iterator(filters=wvc.query.Filter.by_property("bool").equal(True)):
+        assert obj.properties["bool"] is True
+        count += 1
+    assert count == num_objects / 2

weaviate/collections/collection/async_.py

@@ -7,6 +7,7 @@
 from weaviate.collections.backups import _CollectionBackupAsync
 from weaviate.collections.classes.cluster import Shard
 from weaviate.collections.classes.config import ConsistencyLevel
+from weaviate.collections.classes.filters import _Filters
 from weaviate.collections.classes.grpc import METADATA, PROPERTIES, REFERENCES
 from weaviate.collections.classes.internal import (
     CrossReferences,
@@ -24,6 +25,7 @@
 from weaviate.collections.query import _QueryCollectionAsync
 from weaviate.collections.tenants import _TenantsAsync
 from weaviate.connect.v4 import ConnectionAsync
+from weaviate.exceptions import WeaviateUnsupportedFeatureError
 from weaviate.types import UUID
 
 from .base import _CollectionBase
@@ -197,6 +199,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: Literal[None] = None,
         after: Optional[UUID] = None,
@@ -209,6 +212,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: REFERENCES,
         after: Optional[UUID] = None,
@@ -221,6 +225,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: Type[TReferences],
         after: Optional[UUID] = None,
@@ -233,6 +238,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: Literal[None] = None,
         after: Optional[UUID] = None,
@@ -245,6 +251,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: REFERENCES,
         after: Optional[UUID] = None,
@@ -257,6 +264,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: Type[TReferences],
         after: Optional[UUID] = None,
@@ -268,6 +276,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[ReturnProperties[TProperties]] = None,
         return_references: Optional[ReturnReferences[TReferences]] = None,
         after: Optional[UUID] = None,
@@ -301,6 +310,11 @@ def iterator(
         Raises:
             weaviate.exceptions.WeaviateGRPCQueryError: If the request to the Weaviate server fails.
         """
+        if filters is not None and self.query._connection._weaviate_version.is_lower_than(1, 33, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Iterator with filters", self._connection.server_version, "1.33.0"
+            )
+
         return _ObjectAIterator(
             self.query,
             _IteratorInputs(
@@ -309,6 +323,7 @@ def iterator(
                 return_properties=return_properties,
                 return_references=return_references,
                 after=after,
+                filters=filters,
             ),
             cache_size=cache_size,
         )

weaviate/collections/collection/sync.py

@@ -8,6 +8,7 @@
 from weaviate.collections.batch.collection import _BatchCollectionWrapper
 from weaviate.collections.classes.cluster import Shard
 from weaviate.collections.classes.config import ConsistencyLevel
+from weaviate.collections.classes.filters import _Filters
 from weaviate.collections.classes.grpc import METADATA, PROPERTIES, REFERENCES
 from weaviate.collections.classes.internal import (
     CrossReferences,
@@ -25,6 +26,7 @@
 from weaviate.collections.query import _QueryCollection
 from weaviate.collections.tenants import _Tenants
 from weaviate.connect.v4 import ConnectionSync
+from weaviate.exceptions import WeaviateUnsupportedFeatureError
 from weaviate.types import UUID
 
 from .base import _CollectionBase
@@ -221,6 +223,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: Literal[None] = None,
         after: Optional[UUID] = None,
@@ -233,6 +236,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: REFERENCES,
         after: Optional[UUID] = None,
@@ -245,6 +249,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[PROPERTIES] = None,
         return_references: Type[TReferences],
         after: Optional[UUID] = None,
@@ -257,6 +262,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: Literal[None] = None,
         after: Optional[UUID] = None,
@@ -269,6 +275,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: REFERENCES,
         after: Optional[UUID] = None,
@@ -281,6 +288,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Type[TProperties],
         return_references: Type[TReferences],
         after: Optional[UUID] = None,
@@ -292,6 +300,7 @@ def iterator(
         include_vector: bool = False,
         return_metadata: Optional[METADATA] = None,
         *,
+        filters: Optional[_Filters] = None,
         return_properties: Optional[ReturnProperties[TProperties]] = None,
         return_references: Optional[ReturnReferences[TReferences]] = None,
         after: Optional[UUID] = None,
@@ -315,6 +324,7 @@ def iterator(
         are returned. Use `wvc.QueryReference` to specify which references to return.
 
         Args:
+            filters:            The filters to apply to the query.
             include_vector:     Whether to include the vector in the metadata of the returned objects.
             return_metadata:     The metadata to return with each object.
             return_properties:     The properties to return with each object.
@@ -325,9 +335,15 @@ def iterator(
         Raises:
             weaviate.exceptions.WeaviateGRPCQueryError: If the request to the Weaviate server fails.
         """
+        if filters is not None and self.query._connection._weaviate_version.is_lower_than(1, 33, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Iterator with filters", self._connection.server_version, "1.33.0"
+            )
+
         return _ObjectIterator(
             self.query,
             _IteratorInputs(
+                filters=filters,
                 include_vector=include_vector,
                 return_metadata=return_metadata,
                 return_properties=return_properties,

weaviate/collections/iterator.py

@@ -11,6 +11,7 @@
 )
 from uuid import UUID
 
+from weaviate.collections.classes.filters import _Filters
 from weaviate.collections.classes.grpc import METADATA
 from weaviate.collections.classes.internal import (
     Object,
@@ -30,15 +31,19 @@
 
 @dataclass
 class _IteratorInputs(Generic[TProperties, TReferences]):
+    filters: Optional[_Filters]
     include_vector: bool
     return_metadata: Optional[METADATA]
     return_properties: Optional[ReturnProperties[TProperties]]
     return_references: Optional[ReturnReferences[TReferences]]
     after: Optional[UUIDorStr]
 
 
-def _parse_after(after: Optional[UUIDorStr]) -> Optional[UUID]:
-    return after if after is None or isinstance(after, UUID) else UUID(after)
+def _parse_after(after: Optional[UUIDorStr]) -> UUIDorStr:
+    if after is None:
+        return ""
+
+    return UUID(after) if not isinstance(after, UUID) else after
 
 
 class _ObjectIterator(
@@ -55,7 +60,7 @@ def __init__(
         self.__inputs = inputs
 
         self.__iter_object_cache: List[Object[TProperties, TReferences]] = []
-        self.__iter_object_last_uuid: Optional[UUID] = _parse_after(self.__inputs.after)
+        self.__iter_object_last_uuid: Optional[UUIDorStr] = _parse_after(self.__inputs.after)
         self.__iter_cache_size = cache_size or ITERATOR_CACHE_SIZE
 
     def __iter__(
@@ -74,6 +79,7 @@ def __next__(self) -> Object[TProperties, TReferences]:
                 return_metadata=self.__inputs.return_metadata,
                 return_properties=self.__inputs.return_properties,
                 return_references=self.__inputs.return_references,
+                filters=self.__inputs.filters,
             )
             self.__iter_object_cache = res.objects  # type: ignore
             if len(self.__iter_object_cache) == 0:
@@ -101,7 +107,7 @@ def __init__(
         self.__inputs = inputs
 
         self.__iter_object_cache: List[Object[TProperties, TReferences]] = []
-        self.__iter_object_last_uuid: Optional[UUID] = _parse_after(self.__inputs.after)
+        self.__iter_object_last_uuid: UUIDorStr = _parse_after(self.__inputs.after)
         self.__iter_cache_size = cache_size or ITERATOR_CACHE_SIZE
 
     def __aiter__(
@@ -122,6 +128,7 @@ async def __anext__(
                 return_metadata=self.__inputs.return_metadata,
                 return_properties=self.__inputs.return_properties,
                 return_references=self.__inputs.return_references,
+                filters=self.__inputs.filters,
             )
             self.__iter_object_cache = res.objects  # type: ignore
             if len(self.__iter_object_cache) == 0: