elasticsearch-haystack 0.6.0__py3-none-any.whl → 0.7.1__py3-none-any.whl

{elasticsearch_haystack-0.6.0.dist-info → elasticsearch_haystack-0.7.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: elasticsearch-haystack
-Version: 0.6.0
+Version: 0.7.1
 Summary: Haystack 2.x Document Store for ElasticSearch
 Project-URL: Documentation, https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/elasticsearch#readme
 Project-URL: Issues, https://github.com/deepset-ai/haystack-core-integrations/issues

{elasticsearch_haystack-0.6.0.dist-info → elasticsearch_haystack-0.7.1.dist-info}/RECORD

@@ -1,10 +1,10 @@
 haystack_integrations/components/retrievers/elasticsearch/__init__.py,sha256=cSJBsYjz_T4kK-M-auAHVUnYIcgUqqwwQe_hsF0_IG4,307
-haystack_integrations/components/retrievers/elasticsearch/bm25_retriever.py,sha256=fFx554MTcUHnQZa2SgC0PzIR85YVbqAdMNOiXKkVSu8,4849
-haystack_integrations/components/retrievers/elasticsearch/embedding_retriever.py,sha256=RcIbSMELiKIJsD-8F_u76J33YRt5bLr6lHnoX-hVQ1M,4990
+haystack_integrations/components/retrievers/elasticsearch/bm25_retriever.py,sha256=4kCb10S073IHXnBxP4gVU2bxeYlqQbK6FhV8aTg3wVs,5850
+haystack_integrations/components/retrievers/elasticsearch/embedding_retriever.py,sha256=9at7bGJt4ZLSwD4kmTKYGLc7bHAn8zWshx3-sZgDU0g,5991
 haystack_integrations/document_stores/elasticsearch/__init__.py,sha256=YTfu94dtVUBogbJFr1aJrKuaI6-Bw9VuHfPoyU7M8os,207
 haystack_integrations/document_stores/elasticsearch/document_store.py,sha256=H5aqriF7rFYYpqALqAhvBSL41jzGtOxa-vSIPcLgXGw,18719
 haystack_integrations/document_stores/elasticsearch/filters.py,sha256=L1tN7YCIDuNdhGrBQdPoqXFk37x__2-K038xZ6PRdNQ,9923
-elasticsearch_haystack-0.6.0.dist-info/METADATA,sha256=KwdNHi8bOY4M646WWQrkGJRVHzQWQDi16sWFX4idbCQ,2168
-elasticsearch_haystack-0.6.0.dist-info/WHEEL,sha256=zEMcRr9Kr03x1ozGwg5v9NQBKn3kndp6LSoSlVg-jhU,87
-elasticsearch_haystack-0.6.0.dist-info/licenses/LICENSE,sha256=_M2kulivnaiTHiW-5CRlZrPmH47tt04pBgAgeDvfYi4,11342
-elasticsearch_haystack-0.6.0.dist-info/RECORD,,
+elasticsearch_haystack-0.7.1.dist-info/METADATA,sha256=G6FdW5Hm8aNTFKon36o6RoigWXUNY2IrozKQUXksTg8,2168
+elasticsearch_haystack-0.7.1.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+elasticsearch_haystack-0.7.1.dist-info/licenses/LICENSE,sha256=_M2kulivnaiTHiW-5CRlZrPmH47tt04pBgAgeDvfYi4,11342
+elasticsearch_haystack-0.7.1.dist-info/RECORD,,

{elasticsearch_haystack-0.6.0.dist-info → elasticsearch_haystack-0.7.1.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: hatchling 1.24.2
+Generator: hatchling 1.25.0
 Root-Is-Purelib: true
 Tag: py3-none-any

haystack_integrations/components/retrievers/elasticsearch/bm25_retriever.py

@@ -1,10 +1,12 @@
 # SPDX-FileCopyrightText: 2023-present deepset GmbH <info@deepset.ai>
 #
 # SPDX-License-Identifier: Apache-2.0
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, List, Optional, Union
 
 from haystack import component, default_from_dict, default_to_dict
 from haystack.dataclasses import Document
+from haystack.document_stores.types import FilterPolicy
+from haystack.document_stores.types.filter_policy import apply_filter_policy
 from haystack_integrations.document_stores.elasticsearch.document_store import ElasticsearchDocumentStore
 
 
@@ -48,6 +50,7 @@ class ElasticsearchBM25Retriever:
         fuzziness: str = "AUTO",
         top_k: int = 10,
         scale_score: bool = False,
+        filter_policy: Union[str, FilterPolicy] = FilterPolicy.REPLACE,
     ):
         """
         Initialize ElasticsearchBM25Retriever with an instance ElasticsearchDocumentStore.
@@ -60,6 +63,7 @@ class ElasticsearchBM25Retriever:
             for more details.
         :param top_k: Maximum number of Documents to return.
         :param scale_score: If `True` scales the Document`s scores between 0 and 1.
+        :param filter_policy: Policy to determine how filters are applied.
         :raises ValueError: If `document_store` is not an instance of `ElasticsearchDocumentStore`.
         """
 
@@ -72,6 +76,7 @@ class ElasticsearchBM25Retriever:
         self._fuzziness = fuzziness
         self._top_k = top_k
         self._scale_score = scale_score
+        self._filter_policy = FilterPolicy.from_str(filter_policy) if isinstance(filter_policy, str) else filter_policy
 
     def to_dict(self) -> Dict[str, Any]:
         """
@@ -86,6 +91,7 @@ class ElasticsearchBM25Retriever:
             fuzziness=self._fuzziness,
             top_k=self._top_k,
             scale_score=self._scale_score,
+            filter_policy=self._filter_policy.value,
             document_store=self._document_store.to_dict(),
         )
 
@@ -102,6 +108,10 @@ class ElasticsearchBM25Retriever:
         data["init_parameters"]["document_store"] = ElasticsearchDocumentStore.from_dict(
             data["init_parameters"]["document_store"]
         )
+        # Pipelines serialized with old versions of the component might not
+        # have the filter_policy field.
+        if filter_policy := data["init_parameters"].get("filter_policy"):
+            data["init_parameters"]["filter_policy"] = FilterPolicy.from_str(filter_policy)
         return default_from_dict(cls, data)
 
     @component.output_types(documents=List[Document])
@@ -110,14 +120,17 @@ class ElasticsearchBM25Retriever:
         Retrieve documents using the BM25 keyword-based algorithm.
 
         :param query: String to search in `Document`s' text.
-        :param filters: Filters applied to the retrieved `Document`s.
+        :param filters: Filters applied to the retrieved Documents. The way runtime filters are applied depends on
+                        the `filter_policy` chosen at retriever initialization. See init method docstring for more
+                        details.
         :param top_k: Maximum number of `Document` to return.
         :returns: A dictionary with the following keys:
             - `documents`: List of `Document`s that match the query.
         """
+        filters = apply_filter_policy(self._filter_policy, self._filters, filters)
         docs = self._document_store._bm25_retrieval(
             query=query,
-            filters=filters or self._filters,
+            filters=filters,
             fuzziness=self._fuzziness,
             top_k=top_k or self._top_k,
             scale_score=self._scale_score,

haystack_integrations/components/retrievers/elasticsearch/embedding_retriever.py

@@ -1,10 +1,12 @@
 # SPDX-FileCopyrightText: 2023-present deepset GmbH <info@deepset.ai>
 #
 # SPDX-License-Identifier: Apache-2.0
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, List, Optional, Union
 
 from haystack import component, default_from_dict, default_to_dict
 from haystack.dataclasses import Document
+from haystack.document_stores.types import FilterPolicy
+from haystack.document_stores.types.filter_policy import apply_filter_policy
 from haystack_integrations.document_stores.elasticsearch.document_store import ElasticsearchDocumentStore
 
 
@@ -49,6 +51,7 @@ class ElasticsearchEmbeddingRetriever:
         filters: Optional[Dict[str, Any]] = None,
         top_k: int = 10,
         num_candidates: Optional[int] = None,
+        filter_policy: Union[str, FilterPolicy] = FilterPolicy.REPLACE,
     ):
         """
         Create the ElasticsearchEmbeddingRetriever component.
@@ -61,6 +64,7 @@ class ElasticsearchEmbeddingRetriever:
             Increasing this value will improve search accuracy at the cost of slower search speeds.
             You can read more about it in the Elasticsearch
             [documentation](https://www.elastic.co/guide/en/elasticsearch/reference/current/knn-search.html#tune-approximate-knn-for-speed-accuracy)
+        :param filter_policy: Policy to determine how filters are applied.
         :raises ValueError: If `document_store` is not an instance of ElasticsearchDocumentStore.
         """
         if not isinstance(document_store, ElasticsearchDocumentStore):
@@ -71,6 +75,7 @@ class ElasticsearchEmbeddingRetriever:
         self._filters = filters or {}
         self._top_k = top_k
         self._num_candidates = num_candidates
+        self._filter_policy = FilterPolicy.from_str(filter_policy) if isinstance(filter_policy, str) else filter_policy
 
     def to_dict(self) -> Dict[str, Any]:
         """
@@ -84,6 +89,7 @@ class ElasticsearchEmbeddingRetriever:
             filters=self._filters,
             top_k=self._top_k,
             num_candidates=self._num_candidates,
+            filter_policy=self._filter_policy.value,
             document_store=self._document_store.to_dict(),
         )
 
@@ -100,6 +106,10 @@ class ElasticsearchEmbeddingRetriever:
         data["init_parameters"]["document_store"] = ElasticsearchDocumentStore.from_dict(
             data["init_parameters"]["document_store"]
         )
+        # Pipelines serialized with old versions of the component might not
+        # have the filter_policy field.
+        if filter_policy := data["init_parameters"].get("filter_policy"):
+            data["init_parameters"]["filter_policy"] = FilterPolicy.from_str(filter_policy)
         return default_from_dict(cls, data)
 
     @component.output_types(documents=List[Document])
@@ -108,14 +118,17 @@ class ElasticsearchEmbeddingRetriever:
         Retrieve documents using a vector similarity metric.
 
         :param query_embedding: Embedding of the query.
-        :param filters: Filters applied to the retrieved `Document`s.
+        :param filters: Filters applied to the retrieved Documents. The way runtime filters are applied depends on
+                        the `filter_policy` chosen at retriever initialization. See init method docstring for more
+                        details.
         :param top_k: Maximum number of `Document`s to return.
         :returns: A dictionary with the following keys:
             - `documents`: List of `Document`s most similar to the given `query_embedding`
         """
+        filters = apply_filter_policy(self._filter_policy, self._filters, filters)
         docs = self._document_store._embedding_retrieval(
             query_embedding=query_embedding,
-            filters=filters or self._filters,
+            filters=filters,
             top_k=top_k or self._top_k,
             num_candidates=self._num_candidates,
         )