pgvector-haystack 1.2.0__tar.gz → 1.3.0__tar.gz

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/CHANGELOG.md

@@ -1,5 +1,19 @@
 # Changelog
 
+## [integrations/pgvector-v1.2.1] - 2025-01-10
+
+### 🐛 Bug Fixes
+
+- PgvectorDocumentStore - use appropriate schema name if dropping index (#1277)
+
+
+## [integrations/pgvector-v1.2.0] - 2024-11-22
+
+### 🚀 Features
+
+- Add `create_extension` parameter to control vector extension creation (#1213)
+
+
 ## [integrations/pgvector-v1.1.0] - 2024-11-21
 
 ### 🚀 Features

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/PKG-INFO

@@ -1,12 +1,13 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: pgvector-haystack
-Version: 1.2.0
+Version: 1.3.0
 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
 Project-URL: Issues, https://github.com/deepset-ai/haystack-core-integrations/issues
 Author-email: deepset GmbH <info@deepset.ai>
-License: Apache-2.0
+License-Expression: Apache-2.0
+License-File: LICENSE.txt
 Classifier: Development Status :: 4 - Beta
 Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Programming Language :: Python
@@ -47,7 +48,7 @@ pip install pgvector-haystack
 
 Ensure that you have a PostgreSQL running with the `pgvector` extension. For a quick setup using Docker, run:
 ```
-docker run -d -p 5432:5432 -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=postgres ankane/pgvector
+docker run -d -p 5432:5432 -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=postgres pgvector/pgvector:pg17
 ```
 
 then run the tests:

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/README.md

@@ -22,7 +22,7 @@ pip install pgvector-haystack
 
 Ensure that you have a PostgreSQL running with the `pgvector` extension. For a quick setup using Docker, run:
 ```
-docker run -d -p 5432:5432 -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=postgres ankane/pgvector
+docker run -d -p 5432:5432 -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=postgres -e POSTGRES_DB=postgres pgvector/pgvector:pg17
 ```
 
 then run the tests:

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/src/haystack_integrations/document_stores/pgvector/document_store.py

@@ -389,7 +389,9 @@ class PgvectorDocumentStore:
             )
             return
 
-        sql_drop_index = SQL("DROP INDEX IF EXISTS {index_name}").format(index_name=Identifier(self.hnsw_index_name))
+        sql_drop_index = SQL("DROP INDEX IF EXISTS {schema_name}.{index_name}").format(
+            schema_name=Identifier(self.schema_name), index_name=Identifier(self.hnsw_index_name)
+        )
         self._execute_sql(sql_drop_index, error_msg="Could not drop HNSW index")
 
         self._create_hnsw_index()

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/src/haystack_integrations/document_stores/pgvector/filters.py

@@ -232,6 +232,20 @@ def _in(field: str, value: Any) -> Tuple[str, List]:
     return f"{field} = ANY(%s)", [value]
 
 
+def _like(field: str, value: Any) -> Tuple[str, Any]:
+    if not isinstance(value, str):
+        msg = f"{field}'s value must be a str when using 'LIKE' "
+        raise FilterError(msg)
+    return f"{field} LIKE %s", value
+
+
+def _not_like(field: str, value: Any) -> Tuple[str, Any]:
+    if not isinstance(value, str):
+        msg = f"{field}'s value must be a str when using 'LIKE' "
+        raise FilterError(msg)
+    return f"{field} NOT LIKE %s", value
+
+
 COMPARISON_OPERATORS = {
     "==": _equal,
     "!=": _not_equal,
@@ -241,4 +255,6 @@ COMPARISON_OPERATORS = {
     "<=": _less_than_equal,
     "in": _in,
     "not in": _not_in,
+    "like": _like,
+    "not like": _not_like,
 }

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/tests/test_document_store.py

@@ -5,6 +5,7 @@
 from unittest.mock import patch
 
 import numpy as np
+import psycopg
 import pytest
 from haystack.dataclasses.document import ByteStream, Document
 from haystack.document_stores.errors import DuplicateDocumentError
@@ -259,3 +260,47 @@ def test_from_pg_to_haystack_documents():
     assert haystack_docs[2].meta == {"meta_key": "meta_value"}
     assert haystack_docs[2].embedding == [0.7, 0.8, 0.9]
     assert haystack_docs[2].score is None
+
+
+@pytest.mark.integration
+def test_hnsw_index_recreation():
+    def get_index_oid(document_store, schema_name, index_name):
+        sql_get_index_oid = """
+            SELECT c.oid
+            FROM pg_class c
+            JOIN pg_namespace n ON n.oid = c.relnamespace
+            WHERE c.relkind = 'i'
+            AND n.nspname = %s
+            AND c.relname = %s;
+        """
+        return document_store.cursor.execute(sql_get_index_oid, (schema_name, index_name)).fetchone()[0]
+
+    # create a new schema
+    connection_string = "postgresql://postgres:postgres@localhost:5432/postgres"
+    schema_name = "test_schema"
+    with psycopg.connect(connection_string, autocommit=True) as conn:
+        conn.execute(f"CREATE SCHEMA IF NOT EXISTS {schema_name}")
+
+    # create a first document store and trigger the creation of the hnsw index
+    params = {
+        "connection_string": Secret.from_token(connection_string),
+        "schema_name": schema_name,
+        "table_name": "haystack_test_hnsw_index_recreation",
+        "search_strategy": "hnsw",
+    }
+    ds1 = PgvectorDocumentStore(**params)
+    ds1._initialize_table()
+
+    # get the hnsw index oid
+    hnws_index_name = "haystack_hnsw_index"
+    first_oid = get_index_oid(ds1, ds1.schema_name, hnws_index_name)
+
+    # create second document store with recreation enabled
+    ds2 = PgvectorDocumentStore(**params, hnsw_recreate_index_if_exists=True)
+    ds2._initialize_table()
+
+    # get the index oid
+    second_oid = get_index_oid(ds2, ds2.schema_name, hnws_index_name)
+
+    # verify that oids differ
+    assert second_oid != first_oid, "Index was not recreated (OID remained the same)"

{pgvector_haystack-1.2.0 → pgvector_haystack-1.3.0}/tests/test_filters.py

@@ -2,7 +2,7 @@ from typing import List
 
 import pytest
 from haystack.dataclasses.document import Document
-from haystack.testing.document_store import FilterDocumentsTest
+from haystack.testing.document_store import FilterDocumentsTest, FilterDocumentsTestWithDataframe
 from pandas import DataFrame
 from psycopg.sql import SQL
 from psycopg.types.json import Jsonb
@@ -17,7 +17,7 @@ from haystack_integrations.document_stores.pgvector.filters import (
 
 
 @pytest.mark.integration
-class TestFilters(FilterDocumentsTest):
+class TestFilters(FilterDocumentsTest, FilterDocumentsTestWithDataframe):
     def assert_documents_are_equal(self, received: List[Document], expected: List[Document]):
         """
         This overrides the default assert_documents_are_equal from FilterDocumentsTest.
@@ -40,6 +40,62 @@ class TestFilters(FilterDocumentsTest):
     @pytest.mark.skip(reason="NOT operator is not supported in PgvectorDocumentStore")
     def test_not_operator(self, document_store, filterable_docs): ...
 
+    def test_like_operator(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "like", "value": "%Foo%"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(result, [d for d in filterable_docs if "Foo" in d.content])
+
+    def test_like_operator_startswith(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "like", "value": "Foo%"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(result, [d for d in filterable_docs if d.content.startswith("Foo")])
+
+    def test_like_operator_nb_chars(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "like", "value": "A Foobar Document__"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(
+            result,
+            [
+                d
+                for d in filterable_docs
+                if (d.content.startswith("A Foobar Document") and len(d.content) == (len("A Foobar Document") + 2))
+            ],
+        )
+
+    def test_not_like_operator(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "not like", "value": "%Foo%"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(result, [d for d in filterable_docs if "Foo" not in d.content])
+
+    def test_not_like_operator_startswith(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "not like", "value": "Foo%"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(result, [d for d in filterable_docs if not d.content.startswith("Foo")])
+
+    def test_not_like_operator_nb_chars(self, document_store, filterable_docs):
+        document_store.write_documents(filterable_docs)
+        filters = {"field": "content", "operator": "not like", "value": "A Foobar Document__"}
+        result = document_store.filter_documents(filters=filters)
+
+        self.assert_documents_are_equal(
+            result,
+            [
+                d
+                for d in filterable_docs
+                if not (d.content.startswith("A Foobar Document") and len(d.content) == (len("A Foobar Document") + 2))
+            ],
+        )
+
     def test_complex_filter(self, document_store, filterable_docs):
         document_store.write_documents(filterable_docs)
         filters = {