refactor: weaviate, chroma, elasticsearch, opensearch, azure_ai_search use _normalize_metadata_field_name from haystack.utils (#2953)

* initial import

* formatting

* increasing lowest haystack dependency + formatting

* increasing lowest haystack dependency for Azure AI

Author: davidsbatista

integrations/azure_ai_search/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
   "Programming Language :: Python :: Implementation :: PyPy",
 ]
 dependencies = [
-  "haystack-ai>=2.24.0",
+  "haystack-ai>=2.26.0",
   "azure-search-documents>=11.5",
   "azure-identity"
 ]

integrations/azure_ai_search/src/haystack_integrations/document_stores/azure_ai_search/document_store.py

@@ -42,6 +42,7 @@
 from haystack.dataclasses import Document
 from haystack.document_stores.types import DuplicatePolicy
 from haystack.utils import Secret, deserialize_secrets_inplace
+from haystack.utils.misc import _normalize_metadata_field_name
 
 from .errors import AzureAISearchDocumentStoreConfigError, AzureAISearchDocumentStoreError
 from .filters import _normalize_filters
@@ -364,13 +365,6 @@ def count_documents(self) -> int:
         """
         return self.client.get_document_count()
 
-    @staticmethod
-    def _normalize_metadata_field_name(metadata_field: str) -> str:
-        """
-        Normalizes a metadata field name by removing the `meta.` prefix if present.
-        """
-        return metadata_field[5:] if metadata_field.startswith("meta.") else metadata_field
-
     def _get_index_schema_fields(self) -> dict[str, Any]:
         """
         Returns the index schema fields keyed by field name.
@@ -486,7 +480,7 @@ def count_unique_metadata_by_filter(self, filters: dict[str, Any], metadata_fiel
         :param metadata_fields: List of field names to count unique values for.
         :returns: Dictionary mapping field names to counts of unique values.
         """
-        normalized_metadata_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_metadata_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
         self._validate_index_fields(normalized_metadata_fields)
 
         documents = self._fetch_raw_documents(filters=filters, select=normalized_metadata_fields)
@@ -515,7 +509,7 @@ def get_metadata_field_min_max(self, metadata_field: str) -> dict[str, Any]:
         :param metadata_field: The metadata field to get the minimum and maximum values for.
         :returns: A dictionary with the keys "min" and "max".
         """
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         self._validate_index_fields([field_name])
 
         documents = self._fetch_raw_documents(select=[field_name])
@@ -533,7 +527,7 @@ def get_metadata_field_unique_values(
         :param size: Number of values to return.
         :returns: Tuple of (list of unique values, total count of matching values).
         """
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         self._validate_index_fields([field_name])
 
         documents = self._fetch_raw_documents(select=[field_name])

integrations/chroma/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
   "Programming Language :: Python :: Implementation :: PyPy",
 ]
 dependencies = [
-  "haystack-ai>=2.24.0",
+  "haystack-ai>=2.26.0",
   "chromadb>=1.5.0"
 ]
 

integrations/chroma/src/haystack_integrations/document_stores/chroma/document_store.py

@@ -13,6 +13,7 @@
 from haystack.dataclasses import Document
 from haystack.document_stores.errors import DocumentStoreError
 from haystack.document_stores.types import DuplicatePolicy
+from haystack.utils.misc import _normalize_metadata_field_name
 
 from .filters import _convert_filters
 from .utils import get_embedding_function
@@ -242,16 +243,6 @@ def _prepare_query_kwargs(filters: dict[str, Any] | None = None) -> dict[str, An
             "include": ["embeddings", "documents", "metadatas", "distances"],
         }
 
-    @staticmethod
-    def _normalize_metadata_field_name(metadata_field: str) -> str:
-        """
-        Normalizes a metadata field name by removing the "meta." prefix if present.
-
-        :param metadata_field: The metadata field name to normalize.
-        :returns: The normalized field name without "meta." prefix.
-        """
-        return metadata_field[5:] if metadata_field.startswith("meta.") else metadata_field
-
     @staticmethod
     def _infer_type_from_value(value: Any) -> str:
         """
@@ -1087,7 +1078,7 @@ def count_unique_metadata_by_filter(self, filters: dict[str, Any], metadata_fiel
         self._ensure_initialized()
         assert self._collection is not None
 
-        normalized_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
 
         kwargs = ChromaDocumentStore._prepare_get_kwargs(filters)
         kwargs["include"] = ["metadatas"]
@@ -1114,7 +1105,7 @@ async def count_unique_metadata_by_filter_async(
         await self._ensure_initialized_async()
         assert self._async_collection is not None
 
-        normalized_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
 
         kwargs = ChromaDocumentStore._prepare_get_kwargs(filters)
         kwargs["include"] = ["metadatas"]
@@ -1205,7 +1196,7 @@ def get_metadata_field_min_max(self, metadata_field: str) -> dict[str, Any]:
         self._ensure_initialized()
         assert self._collection is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         result = self._collection.get(include=["metadatas"])
         return self._compute_field_min_max(result.get("metadatas", []), field_name)
@@ -1229,7 +1220,7 @@ async def get_metadata_field_min_max_async(self, metadata_field: str) -> dict[st
         await self._ensure_initialized_async()
         assert self._async_collection is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         result = await self._async_collection.get(include=["metadatas"])
         return self._compute_field_min_max(result.get("metadatas", []), field_name)
@@ -1256,7 +1247,7 @@ def get_metadata_field_unique_values(
         self._ensure_initialized()
         assert self._collection is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         kwargs: dict[str, Any] = {"include": ["metadatas"]}
         if search_term:
@@ -1289,7 +1280,7 @@ async def get_metadata_field_unique_values_async(
         await self._ensure_initialized_async()
         assert self._async_collection is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         kwargs: dict[str, Any] = {"include": ["metadatas"]}
         if search_term:

integrations/elasticsearch/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
   "Programming Language :: Python :: Implementation :: PyPy",
 ]
 dependencies = [
-  "haystack-ai>=2.24.0",
+  "haystack-ai>=2.26.0",
   "elasticsearch>=8,<9",
   "aiohttp>=3.9.0" # for async support https://elasticsearch-py.readthedocs.io/en/latest/async.html#valueerror-when-initializing-asyncelasticsearch
 ]

integrations/elasticsearch/src/haystack_integrations/document_stores/elasticsearch/document_store.py

@@ -17,6 +17,7 @@
 from haystack.document_stores.errors import DocumentStoreError, DuplicateDocumentError
 from haystack.document_stores.types import DuplicatePolicy
 from haystack.utils import Secret
+from haystack.utils.misc import _normalize_metadata_field_name
 from haystack.version import __version__ as haystack_version
 
 from elasticsearch import AsyncElasticsearch, Elasticsearch, helpers
@@ -1077,13 +1078,6 @@ async def count_documents_by_filter_async(self, filters: dict[str, Any]) -> int:
         result = await self.async_client.count(index=self._index, body=body)
         return result["count"]
 
-    @staticmethod
-    def _normalize_metadata_field_name(metadata_field: str) -> str:
-        """
-        Normalizes a metadata field name by removing the "meta." prefix if present.
-        """
-        return metadata_field[5:] if metadata_field.startswith("meta.") else metadata_field
-
     @staticmethod
     def _build_cardinality_aggregations(index_mapping: dict[str, Any], fields: list[str]) -> dict[str, Any]:
         """
@@ -1160,7 +1154,7 @@ def count_unique_metadata_by_filter(self, filters: dict[str, Any], metadata_fiel
         index_mapping = mapping[self._index]["mappings"]["properties"]
 
         # normalize field names, e.g: remove "meta." prefix if present
-        normalized_metadata_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_metadata_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
 
         # validate that all requested fields exist in the index mapping
         missing_fields = [f for f in normalized_metadata_fields if f not in index_mapping]
@@ -1204,7 +1198,7 @@ async def count_unique_metadata_by_filter_async(
         index_mapping = mapping[self._index]["mappings"]["properties"]
 
         # normalize field names
-        normalized_metadata_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_metadata_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
         # validate that all requested fields exist in the index mapping
         missing_fields = [f for f in normalized_metadata_fields if f not in index_mapping]
         if missing_fields:
@@ -1325,7 +1319,7 @@ def get_metadata_field_min_max(self, metadata_field: str) -> dict[str, int | Non
         """
         self._ensure_initialized()
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         body = self._build_min_max_query_body(field_name)
         result = self.client.search(index=self._index, body=body)
         stats = result.get("aggregations", {}).get("field_stats", {})
@@ -1342,7 +1336,7 @@ async def get_metadata_field_min_max_async(self, metadata_field: str) -> dict[st
         """
         self._ensure_initialized()
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         body = self._build_min_max_query_body(field_name)
         result = await self.async_client.search(index=self._index, body=body)
         stats = result.get("aggregations", {}).get("field_stats", {})
@@ -1373,7 +1367,7 @@ def get_metadata_field_unique_values(
         """
         self._ensure_initialized()
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         # filter by search_term if provided
         query: dict[str, Any] = {"match_all": {}}
@@ -1439,7 +1433,7 @@ async def get_metadata_field_unique_values_async(
         """
         self._ensure_initialized()
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         # filter by search_term if provided
         query: dict[str, Any] = {"match_all": {}}

integrations/opensearch/pyproject.toml

@@ -24,7 +24,7 @@ classifiers = [
 ]
 
 dependencies = [
-  "haystack-ai>=2.24.0",
+  "haystack-ai>=2.26.0",
   "opensearch-py[async]>=3.0.0"
 ]
 

integrations/opensearch/src/haystack_integrations/document_stores/opensearch/document_store.py

@@ -13,6 +13,7 @@
 from haystack.document_stores.errors import DocumentStoreError, DuplicateDocumentError
 from haystack.document_stores.types import DuplicatePolicy
 from haystack.utils.auth import Secret
+from haystack.utils.misc import _normalize_metadata_field_name
 from opensearchpy import AsyncHttpConnection, AsyncOpenSearch, OpenSearch
 from opensearchpy.helpers import async_bulk, bulk
 
@@ -1659,7 +1660,7 @@ def count_unique_metadata_by_filter(self, filters: dict[str, Any], metadata_fiel
         index_mapping = mapping[self._index]["mappings"]["properties"]
 
         # normalize field names
-        normalized_metadata_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_metadata_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
         # validate that all requested fields exist in the index mapping
         missing_fields = [f for f in normalized_metadata_fields if f not in index_mapping]
         if missing_fields:
@@ -1703,7 +1704,7 @@ async def count_unique_metadata_by_filter_async(
         index_mapping = mapping[self._index]["mappings"]["properties"]
 
         # normalize field names
-        normalized_metadata_fields = [self._normalize_metadata_field_name(field) for field in metadata_fields]
+        normalized_metadata_fields = [_normalize_metadata_field_name(field) for field in metadata_fields]
         # validate that all requested fields exist in the index mapping
         missing_fields = [f for f in normalized_metadata_fields if f not in index_mapping]
         if missing_fields:
@@ -1790,13 +1791,6 @@ async def get_metadata_fields_info_async(self) -> dict[str, dict[str, str]]:
         index_mapping = {k: v for k, v in index_mapping.items() if k not in SPECIAL_FIELDS}
         return index_mapping
 
-    @staticmethod
-    def _normalize_metadata_field_name(metadata_field: str) -> str:
-        """
-        Normalizes a metadata field name by removing the "meta." prefix if present.
-        """
-        return metadata_field[5:] if metadata_field.startswith("meta.") else metadata_field
-
     @staticmethod
     def _build_min_max_query_body(field_name: str) -> dict[str, Any]:
         """
@@ -1834,7 +1828,7 @@ def get_metadata_field_min_max(self, metadata_field: str) -> dict[str, int | Non
         self._ensure_initialized()
         assert self._client is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         body = self._build_min_max_query_body(field_name)
         result = self._client.search(index=self._index, body=body)
         stats = result.get("aggregations", {}).get("field_stats", {})
@@ -1852,7 +1846,7 @@ async def get_metadata_field_min_max_async(self, metadata_field: str) -> dict[st
         await self._ensure_initialized_async()
         assert self._async_client is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
         body = self._build_min_max_query_body(field_name)
         result = await self._async_client.search(index=self._index, body=body)
         stats = result.get("aggregations", {}).get("field_stats", {})
@@ -1882,7 +1876,7 @@ def get_metadata_field_unique_values(
         self._ensure_initialized()
         assert self._client is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         # filter by search_term if provided
         query: dict[str, Any] = {"match_all": {}}
@@ -1947,7 +1941,7 @@ async def get_metadata_field_unique_values_async(
         await self._ensure_initialized_async()
         assert self._async_client is not None
 
-        field_name = self._normalize_metadata_field_name(metadata_field)
+        field_name = _normalize_metadata_field_name(metadata_field)
 
         # filter by search_term if provided
         query: dict[str, Any] = {"match_all": {}}

integrations/weaviate/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
   "Programming Language :: Python :: Implementation :: PyPy",
 ]
 dependencies = [
-  "haystack-ai>=2.24.0",
+  "haystack-ai>=2.26.0",
   "weaviate-client>=4.20",
   "python-dateutil",
 ]