pgvector-haystack 6.3.1__tar.gz → 6.4.0__tar.gz

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/CHANGELOG.md

@@ -1,5 +1,29 @@
 # Changelog
 
+## [integrations/pgvector-v6.3.1] - 2026-06-01
+
+### 🐛 Bug Fixes
+
+- Replace in place dataclass mutations for pgvector #11087 (#3142)
+- *(pgvector)* Order retrieval by distance operator (#3370)
+
+### 🚜 Refactor
+
+- *(pgvector)* Use async DocumentStore mixin tests (#3094)
+
+### 🧪 Testing
+
+- `PGVectorDocumentStore`e use Mixin tests (#3003)
+- Test compatible integrations with python 3.14; update pyproject (#3001)
+- Track test coverage for all integrations (#3065)
+- Add tests for pgvector document store (#3156)
+
+### 🧹 Chores
+
+- Add ANN ruff ruleset to optimum, paddleocr, pgvector, pinecone, pyversity, qdrant, ragas, snowflake (#2992)
+- Enforce ruff docstring rules in integrations 31-40 (openrouter, opensearch, optimum, paddleocr, pgvector, pinecone, pyversity, qdrant, ragas, snowflake) (#3011)
+
+
 ## [integrations/pgvector-v6.3.0] - 2026-03-17
 
 ### 🐛 Bug Fixes

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pgvector-haystack
-Version: 6.3.1
+Version: 6.4.0
 Summary: An integration of pgvector (vector search extension for Postgres) 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/pgvector/README.md

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/examples/embedding_retrieval.py

@@ -14,10 +14,13 @@ import glob
 
 from haystack import Pipeline
 from haystack.components.converters import MarkdownToDocument
-from haystack.components.embedders import SentenceTransformersDocumentEmbedder, SentenceTransformersTextEmbedder
 from haystack.components.preprocessors import DocumentSplitter
 from haystack.components.writers import DocumentWriter
 
+from haystack_integrations.components.embedders.sentence_transformers import (
+    SentenceTransformersDocumentEmbedder,
+    SentenceTransformersTextEmbedder,
+)
 from haystack_integrations.components.retrievers.pgvector import PgvectorEmbeddingRetriever
 from haystack_integrations.document_stores.pgvector import PgvectorDocumentStore
 

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/examples/hybrid_retrieval.py

@@ -14,11 +14,14 @@ import glob
 
 from haystack import Pipeline
 from haystack.components.converters import MarkdownToDocument
-from haystack.components.embedders import SentenceTransformersDocumentEmbedder, SentenceTransformersTextEmbedder
 from haystack.components.joiners import DocumentJoiner
 from haystack.components.preprocessors import DocumentSplitter
 from haystack.components.writers import DocumentWriter
 
+from haystack_integrations.components.embedders.sentence_transformers import (
+    SentenceTransformersDocumentEmbedder,
+    SentenceTransformersTextEmbedder,
+)
 from haystack_integrations.components.retrievers.pgvector import PgvectorEmbeddingRetriever, PgvectorKeywordRetriever
 from haystack_integrations.document_stores.pgvector import PgvectorDocumentStore
 

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/src/haystack_integrations/components/retrievers/pgvector/embedding_retriever.py

@@ -21,7 +21,9 @@ class PgvectorEmbeddingRetriever:
     ```python
     from haystack.document_stores import DuplicatePolicy
     from haystack import Document, Pipeline
-    from haystack.components.embedders import SentenceTransformersTextEmbedder, SentenceTransformersDocumentEmbedder
+    # Requires: pip install sentence-transformers-haystack
+    from haystack_integrations.components.embedders.sentence_transformers import SentenceTransformersTextEmbedder
+    from haystack_integrations.components.embedders.sentence_transformers import SentenceTransformersDocumentEmbedder
 
     from haystack_integrations.document_stores.pgvector import PgvectorDocumentStore
     from haystack_integrations.components.retrievers.pgvector import PgvectorEmbeddingRetriever
@@ -40,7 +42,6 @@ class PgvectorEmbeddingRetriever:
                  Document(content="In certain places, you can witness the phenomenon of bioluminescent waves.")]
 
     document_embedder = SentenceTransformersDocumentEmbedder()
-    document_embedder.warm_up()
     documents_with_embeddings = document_embedder.run(documents)
 
     document_store.write_documents(documents_with_embeddings.get("documents"), policy=DuplicatePolicy.OVERWRITE)
@@ -135,6 +136,18 @@ class PgvectorEmbeddingRetriever:
             data["init_parameters"]["filter_policy"] = FilterPolicy.from_str(filter_policy)
         return default_from_dict(cls, data)
 
+    def close(self) -> None:
+        """
+        Release the synchronous resources of the underlying Document Store.
+        """
+        self.document_store.close()
+
+    async def close_async(self) -> None:
+        """
+        Release the asynchronous resources of the underlying Document Store.
+        """
+        await self.document_store.close_async()
+
     @component.output_types(documents=list[Document])
     def run(
         self,

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/src/haystack_integrations/components/retrievers/pgvector/keyword_retriever.py

@@ -109,6 +109,18 @@ class PgvectorKeywordRetriever:
             data["init_parameters"]["filter_policy"] = FilterPolicy.from_str(filter_policy)
         return default_from_dict(cls, data)
 
+    def close(self) -> None:
+        """
+        Release the synchronous resources of the underlying Document Store.
+        """
+        self.document_store.close()
+
+    async def close_async(self) -> None:
+        """
+        Release the asynchronous resources of the underlying Document Store.
+        """
+        await self.document_store.close_async()
+
     @component.output_types(documents=list[Document])
     def run(
         self,

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/src/haystack_integrations/document_stores/pgvector/document_store.py

@@ -2,6 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
+from contextlib import suppress
 from typing import Any, Literal, overload
 
 from haystack import default_from_dict, default_to_dict, logging
@@ -223,6 +224,28 @@ class PgvectorDocumentStore:
         deserialize_secrets_inplace(data["init_parameters"], ["connection_string"])
         return default_from_dict(cls, data)
 
+    def close(self) -> None:
+        """
+        Release the associated synchronous resources.
+        """
+        if self._connection is not None:
+            with suppress(Exception):
+                self._connection.close()
+            self._connection = None
+            self._cursor = None
+            self._dict_cursor = None
+
+    async def close_async(self) -> None:
+        """
+        Release the associated asynchronous resources.
+        """
+        if self._async_connection is not None:
+            with suppress(Exception):
+                await self._async_connection.close()
+            self._async_connection = None
+            self._async_cursor = None
+            self._async_dict_cursor = None
+
     @staticmethod
     def _connection_is_valid(connection: Connection) -> bool:
         """

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/src/haystack_integrations/document_stores/pgvector/filters.py

@@ -20,7 +20,7 @@ PYTHON_TYPES_TO_PG_TYPES = {
     bool: "boolean",
 }
 
-NO_VALUE = "no_value"
+NO_VALUE = object()  # unique sentinel; can never equal a user-supplied filter value
 
 
 def _validate_filters(filters: dict[str, Any] | None = None) -> None:
@@ -48,7 +48,7 @@ def _convert_filters_to_where_clause_and_params(
         query, values = _parse_logical_condition(filters)
 
     where_clause = SQL(f" {operator} ") + query
-    params = tuple(value for value in values if value != NO_VALUE)
+    params = tuple(value for value in values if value is not NO_VALUE)
 
     return where_clause, params
 

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/tests/test_document_store.py

@@ -103,6 +103,18 @@ class TestDocumentStore(
             document_store._ensure_db_setup()
         assert "Failed to connect to PostgreSQL database" in str(e)
 
+    def test_close_and_reopen(self, document_store: PgvectorDocumentStore):
+        assert document_store.count_documents() == 0
+        assert document_store._connection is not None
+
+        document_store.close()
+        assert document_store._connection is None
+        assert document_store._cursor is None
+        assert document_store._dict_cursor is None
+
+        assert document_store.count_documents() == 0
+        assert document_store._connection is not None
+
 
 @pytest.mark.usefixtures("patches_for_unit_tests")
 def test_init(monkeypatch):
@@ -353,6 +365,33 @@ def test_process_count_unique_metadata_result_uses_zero_for_missing_keys():
     assert counts == {"category": 5, "language": 0}
 
 
+def test_close(mock_store):
+    mock_connection = Mock(spec=Connection)
+    mock_store._connection = mock_connection
+    mock_store._cursor = Mock(spec=Cursor)
+    mock_store._dict_cursor = Mock(spec=Cursor)
+
+    mock_store.close()
+
+    mock_connection.close.assert_called_once()
+    assert mock_store._connection is None
+    assert mock_store._cursor is None
+    assert mock_store._dict_cursor is None
+
+    mock_store.close()
+    mock_connection.close.assert_called_once()
+
+
+def test_close_is_exception_safe(mock_store):
+    mock_connection = Mock(spec=Connection)
+    mock_connection.close.side_effect = RuntimeError("boom")
+    mock_store._connection = mock_connection
+
+    mock_store.close()
+
+    assert mock_store._connection is None
+
+
 @pytest.mark.integration
 def test_halfvec_hnsw_write_documents(document_store_w_halfvec_hnsw_index: PgvectorDocumentStore):
     documents = [

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/tests/test_document_store_async.py

@@ -2,7 +2,7 @@
 #
 # SPDX-License-Identifier: Apache-2.0
 
-from unittest.mock import Mock, patch
+from unittest.mock import AsyncMock, Mock, patch
 
 import psycopg
 import pytest
@@ -131,6 +131,18 @@ class TestDocumentStoreAsync(
         """
         await document_store.delete_table_async()  # if throw error, test fails
 
+    async def test_close_async_and_reopen(self, document_store: PgvectorDocumentStore):
+        assert await document_store.count_documents_async() == 0
+        assert document_store._async_connection is not None
+
+        await document_store.close_async()
+        assert document_store._async_connection is None
+        assert document_store._async_cursor is None
+        assert document_store._async_dict_cursor is None
+
+        assert await document_store.count_documents_async() == 0
+        assert document_store._async_connection is not None
+
 
 @pytest.mark.integration
 @pytest.mark.asyncio
@@ -302,3 +314,33 @@ async def test_write_documents_async_rejects_non_document_items(mock_store):
 async def test_count_unique_metadata_by_filter_async_rejects_empty_fields(mock_store):
     with pytest.raises(ValueError, match="metadata_fields must be a non-empty list"):
         await mock_store.count_unique_metadata_by_filter_async(filters={}, metadata_fields=[])
+
+
+@pytest.mark.asyncio
+async def test_close_async(mock_store):
+    mock_connection = Mock(spec=AsyncConnection)
+    mock_connection.close = AsyncMock()
+    mock_store._async_connection = mock_connection
+    mock_store._async_cursor = Mock(spec=AsyncCursor)
+    mock_store._async_dict_cursor = Mock(spec=AsyncCursor)
+
+    await mock_store.close_async()
+
+    mock_connection.close.assert_awaited_once()
+    assert mock_store._async_connection is None
+    assert mock_store._async_cursor is None
+    assert mock_store._async_dict_cursor is None
+
+    await mock_store.close_async()
+    mock_connection.close.assert_awaited_once()
+
+
+@pytest.mark.asyncio
+async def test_close_async_is_exception_safe(mock_store):
+    mock_connection = Mock(spec=AsyncConnection)
+    mock_connection.close = AsyncMock(side_effect=RuntimeError("boom"))
+    mock_store._async_connection = mock_connection
+
+    await mock_store.close_async()
+
+    assert mock_store._async_connection is None

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/tests/test_filters.py

@@ -265,6 +265,18 @@ def test_convert_filters_to_where_clause_and_params_handle_null():
     assert params == ("intro",)
 
 
+def test_convert_filters_to_where_clause_and_params_value_equal_to_no_value_sentinel():
+    """A legitimate equality filter whose value is the literal string the internal
+    NO_VALUE sentinel used to be ("no_value") must not be dropped from params --
+    the query still has a placeholder for it."""
+    filters = {"field": "meta.tag", "operator": "==", "value": "no_value"}
+    where_clause, params = _convert_filters_to_where_clause_and_params(filters)
+
+    expected_sql = " WHERE meta->>'tag' = %s"
+    assert _render(where_clause) == expected_sql
+    assert params == ("no_value",)
+
+
 def test_validate_filters():
     filters = {"field": "meta.number", "operator": "==", "value": 100}
     _validate_filters(filters)

{pgvector_haystack-6.3.1 → pgvector_haystack-6.4.0}/tests/test_retrievers.py

@@ -1,7 +1,7 @@
 # SPDX-FileCopyrightText: 2023-present deepset GmbH <info@deepset.ai>
 #
 # SPDX-License-Identifier: Apache-2.0
-from unittest.mock import Mock
+from unittest.mock import AsyncMock, Mock
 
 import pytest
 from haystack.dataclasses import Document
@@ -126,6 +126,26 @@ class TestEmbeddingRetriever:
         assert retriever.filter_policy == FilterPolicy.REPLACE
         assert retriever.vector_function == "l2_distance"
 
+    def test_close(self):
+        mock_store = Mock(spec=PgvectorDocumentStore)
+        retriever = PgvectorEmbeddingRetriever(document_store=mock_store, vector_function="l2_distance")
+
+        retriever.close()
+
+        mock_store.close.assert_called_once()
+        assert retriever.document_store is mock_store
+
+    @pytest.mark.asyncio
+    async def test_close_async(self):
+        mock_store = Mock(spec=PgvectorDocumentStore)
+        mock_store.close_async = AsyncMock()
+        retriever = PgvectorEmbeddingRetriever(document_store=mock_store, vector_function="l2_distance")
+
+        await retriever.close_async()
+
+        mock_store.close_async.assert_awaited_once()
+        assert retriever.document_store is mock_store
+
     def test_run(self):
         mock_store = Mock(spec=PgvectorDocumentStore)
         doc = Document(content="Test doc", embedding=[0.1, 0.2])
@@ -315,6 +335,26 @@ class TestKeywordRetriever:
         assert retriever.filter_policy == FilterPolicy.REPLACE  # defaults to REPLACE
         assert retriever.top_k == 5
 
+    def test_close(self):
+        mock_store = Mock(spec=PgvectorDocumentStore)
+        retriever = PgvectorKeywordRetriever(document_store=mock_store)
+
+        retriever.close()
+
+        mock_store.close.assert_called_once()
+        assert retriever.document_store is mock_store
+
+    @pytest.mark.asyncio
+    async def test_close_async(self):
+        mock_store = Mock(spec=PgvectorDocumentStore)
+        mock_store.close_async = AsyncMock()
+        retriever = PgvectorKeywordRetriever(document_store=mock_store)
+
+        await retriever.close_async()
+
+        mock_store.close_async.assert_awaited_once()
+        assert retriever.document_store is mock_store
+
     def test_run(self):
         mock_store = Mock(spec=PgvectorDocumentStore)
         doc = Document(content="Test doc", embedding=[0.1, 0.2])