qdrant-haystack 4.1.2__py3-none-any.whl → 5.0.0__py3-none-any.whl

haystack_integrations/components/retrievers/qdrant/retriever.py

@@ -4,9 +4,10 @@ from haystack import Document, component, default_from_dict, default_to_dict
 from haystack.dataclasses.sparse_embedding import SparseEmbedding
 from haystack.document_stores.types import FilterPolicy
 from haystack.document_stores.types.filter_policy import apply_filter_policy
-from haystack_integrations.document_stores.qdrant import QdrantDocumentStore
 from qdrant_client.http import models
 
+from haystack_integrations.document_stores.qdrant import QdrantDocumentStore
+
 
 @component
 class QdrantEmbeddingRetriever:

haystack_integrations/document_stores/qdrant/document_store.py

@@ -15,7 +15,6 @@ from haystack.utils.filters import convert as convert_legacy_filters
 from qdrant_client import grpc
 from qdrant_client.http import models as rest
 from qdrant_client.http.exceptions import UnexpectedResponse
-from qdrant_client.hybrid.fusion import reciprocal_rank_fusion
 from tqdm import tqdm
 
 from .converters import (
@@ -335,7 +334,7 @@ class QdrantDocumentStore:
         self,
         documents: List[Document],
         policy: DuplicatePolicy = DuplicatePolicy.FAIL,
-    ):
+    ) -> int:
         """
         Writes documents to Qdrant using the specified policy.
         The QdrantDocumentStore can handle duplicate documents based on the given policy.
@@ -359,7 +358,7 @@ class QdrantDocumentStore:
 
         if len(documents) == 0:
             logger.warning("Calling QdrantDocumentStore.write_documents() with empty list")
-            return
+            return 0
 
         document_objects = self._handle_duplicate_documents(
             documents=documents,
@@ -384,13 +383,13 @@ class QdrantDocumentStore:
                 progress_bar.update(self.write_batch_size)
         return len(document_objects)
 
-    def delete_documents(self, ids: List[str]):
+    def delete_documents(self, document_ids: List[str]) -> None:
         """
         Deletes documents that match the provided `document_ids` from the document store.
 
         :param document_ids: the document ids to delete
         """
-        ids = [convert_id(_id) for _id in ids]
+        ids = [convert_id(_id) for _id in document_ids]
         try:
             self.client.delete(
                 collection_name=self.index,
@@ -537,20 +536,18 @@ class QdrantDocumentStore:
         qdrant_filters = convert_filters_to_qdrant(filters)
         query_indices = query_sparse_embedding.indices
         query_values = query_sparse_embedding.values
-        points = self.client.search(
+        points = self.client.query_points(
             collection_name=self.index,
-            query_vector=rest.NamedSparseVector(
-                name=SPARSE_VECTORS_NAME,
-                vector=rest.SparseVector(
-                    indices=query_indices,
-                    values=query_values,
-                ),
+            query=rest.SparseVector(
+                indices=query_indices,
+                values=query_values,
             ),
+            using=SPARSE_VECTORS_NAME,
             query_filter=qdrant_filters,
             limit=top_k,
             with_vectors=return_embedding,
             score_threshold=score_threshold,
-        )
+        ).points
         results = [
             convert_qdrant_point_to_haystack_document(point, use_sparse_embeddings=self.use_sparse_embeddings)
             for point in points
@@ -588,17 +585,15 @@ class QdrantDocumentStore:
         """
         qdrant_filters = convert_filters_to_qdrant(filters)
 
-        points = self.client.search(
+        points = self.client.query_points(
             collection_name=self.index,
-            query_vector=rest.NamedVector(
-                name=DENSE_VECTORS_NAME if self.use_sparse_embeddings else "",
-                vector=query_embedding,
-            ),
+            query=query_embedding,
+            using=DENSE_VECTORS_NAME if self.use_sparse_embeddings else None,
             query_filter=qdrant_filters,
             limit=top_k,
             with_vectors=return_embedding,
             score_threshold=score_threshold,
-        )
+        ).points
         results = [
             convert_qdrant_point_to_haystack_document(point, use_sparse_embeddings=self.use_sparse_embeddings)
             for point in points
@@ -655,46 +650,34 @@ class QdrantDocumentStore:
 
         qdrant_filters = convert_filters_to_qdrant(filters)
 
-        sparse_request = rest.SearchRequest(
-            vector=rest.NamedSparseVector(
-                name=SPARSE_VECTORS_NAME,
-                vector=rest.SparseVector(
-                    indices=query_sparse_embedding.indices,
-                    values=query_sparse_embedding.values,
-                ),
-            ),
-            filter=qdrant_filters,
-            limit=top_k,
-            with_payload=True,
-            with_vector=return_embedding,
-            score_threshold=score_threshold,
-        )
-
-        dense_request = rest.SearchRequest(
-            vector=rest.NamedVector(
-                name=DENSE_VECTORS_NAME,
-                vector=query_embedding,
-            ),
-            filter=qdrant_filters,
-            limit=top_k,
-            with_payload=True,
-            with_vector=return_embedding,
-        )
-
         try:
-            dense_request_response, sparse_request_response = self.client.search_batch(
-                collection_name=self.index, requests=[dense_request, sparse_request]
-            )
+            points = self.client.query_points(
+                collection_name=self.index,
+                prefetch=[
+                    rest.Prefetch(
+                        query=rest.SparseVector(
+                            indices=query_sparse_embedding.indices,
+                            values=query_sparse_embedding.values,
+                        ),
+                        using=SPARSE_VECTORS_NAME,
+                        filter=qdrant_filters,
+                    ),
+                    rest.Prefetch(
+                        query=query_embedding,
+                        using=DENSE_VECTORS_NAME,
+                        filter=qdrant_filters,
+                    ),
+                ],
+                query=rest.FusionQuery(fusion=rest.Fusion.RRF),
+                limit=top_k,
+                score_threshold=score_threshold,
+                with_payload=True,
+                with_vectors=return_embedding,
+            ).points
         except Exception as e:
             msg = "Error during hybrid search"
             raise QdrantStoreError(msg) from e
 
-        try:
-            points = reciprocal_rank_fusion(responses=[dense_request_response, sparse_request_response], limit=top_k)
-        except Exception as e:
-            msg = "Error while applying Reciprocal Rank Fusion"
-            raise QdrantStoreError(msg) from e
-
         results = [convert_qdrant_point_to_haystack_document(point, use_sparse_embeddings=True) for point in points]
 
         return results

haystack_integrations/document_stores/qdrant/migrate_to_sparse.py

@@ -1,9 +1,10 @@
 import logging
 import time
 
-from haystack_integrations.document_stores.qdrant import QdrantDocumentStore
 from qdrant_client.http import models
 
+from haystack_integrations.document_stores.qdrant import QdrantDocumentStore
+
 logger = logging.getLogger(__name__)
 logger.addHandler(logging.StreamHandler())
 logger.setLevel(logging.INFO)

{qdrant_haystack-4.1.2.dist-info → qdrant_haystack-5.0.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: qdrant-haystack
-Version: 4.1.2
+Version: 5.0.0
 Summary: An integration of Qdrant ANN vector database backend with Haystack
 Project-URL: Source, https://github.com/deepset-ai/haystack-core-integrations
 Project-URL: Documentation, https://github.com/deepset-ai/haystack-core-integrations/blob/main/integrations/qdrant/README.md

{qdrant_haystack-4.1.2.dist-info → qdrant_haystack-5.0.0.dist-info}/RECORD

@@ -1,11 +1,11 @@
 haystack_integrations/components/retrievers/qdrant/__init__.py,sha256=IRjcM4f8b5eKFEMn8tn6h6RrfslEGP3WafU7mrzNzQM,313
-haystack_integrations/components/retrievers/qdrant/retriever.py,sha256=qM3zopQmbPo_mejnFGZ24TMr1HR_jL7tc90M5fnYPOw,18448
+haystack_integrations/components/retrievers/qdrant/retriever.py,sha256=DIqwa2JMVP7z52wmKdCfFk1ZdV0j50ZYCuRluQu5llk,18449
 haystack_integrations/document_stores/qdrant/__init__.py,sha256=kUGc5uewqArhmVR-JqB_NmJ4kNkTIQIvYDNSoO2ELn0,302
 haystack_integrations/document_stores/qdrant/converters.py,sha256=2hcuI3kty1dVHzX1WGXxEtlrnZ9E8TAG56XATCFa6Pw,2491
-haystack_integrations/document_stores/qdrant/document_store.py,sha256=eLw4P1h8GCj40R-BIlQOvJG9MpDzvtmQ7Hpb3AZhMSo,38117
+haystack_integrations/document_stores/qdrant/document_store.py,sha256=Rr95CcyWImFwYuy8cKHdtH6784w4wKbuypSJ-wijgSs,37622
 haystack_integrations/document_stores/qdrant/filters.py,sha256=Nv_eKIYKwUWvldJfa0omfFQ0kgqi6L3DUFeMuIWziOY,11751
-haystack_integrations/document_stores/qdrant/migrate_to_sparse.py,sha256=i6wBC_9_JVzYZtqKm3dhHKTxhwNdcAdpgki8GABDp1c,4909
-qdrant_haystack-4.1.2.dist-info/METADATA,sha256=Yr9ah4TZp8JjwYndH0kbQt4moQe52EhR3WWKoSni89k,1863
-qdrant_haystack-4.1.2.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
-qdrant_haystack-4.1.2.dist-info/licenses/LICENSE.txt,sha256=B05uMshqTA74s-0ltyHKI6yoPfJ3zYgQbvcXfDVGFf8,10280
-qdrant_haystack-4.1.2.dist-info/RECORD,,
+haystack_integrations/document_stores/qdrant/migrate_to_sparse.py,sha256=yhZr4GB6N1S-Ikzl52hpuZt2aHNIb4leqFDhVMU3Uho,4910
+qdrant_haystack-5.0.0.dist-info/METADATA,sha256=1N-UiU98RJX8yY6Ky7UMcXE3aKSjThqB9pBCqE6J_zY,1863
+qdrant_haystack-5.0.0.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+qdrant_haystack-5.0.0.dist-info/licenses/LICENSE.txt,sha256=B05uMshqTA74s-0ltyHKI6yoPfJ3zYgQbvcXfDVGFf8,10280
+qdrant_haystack-5.0.0.dist-info/RECORD,,