chroma-haystack 2.0.0__tar.gz → 2.0.1__tar.gz

{chroma_haystack-2.0.0 → chroma_haystack-2.0.1}/CHANGELOG.md

@@ -1,5 +1,27 @@
 # Changelog
 
+## [integrations/chroma-v1.0.1] - 2025-02-14
+
+### 🧹 Chores
+
+- Inherit from `FilterDocumentsTestWithDataframe` in Document Stores (#1290)
+
+### 🌀 Miscellaneous
+
+- Fix: Chroma - make `filter_documents` method return embeddings (#1361)
+
+## [integrations/chroma-v2.0.0] - 2025-01-02
+
+### 🧹 Chores
+
+- Fix linting/isort (#1215)
+- Chroma - pin `tokenizers` (#1223)
+
+### 🌀 Miscellaneous
+
+- Unpin tokenizers (#1233)
+- Fix: updates for Chroma 0.6.0 (#1270)
+
 ## [integrations/chroma-v1.0.0] - 2024-11-06
 
 ### 🐛 Bug Fixes
@@ -7,13 +29,14 @@
 - Fixing Chroma tests due `chromadb` update behaviour change (#1148)
 - Adapt our implementation to breaking changes in Chroma 0.5.17  (#1165)
 
-### ⚙️ Miscellaneous Tasks
+### ⚙️ CI
 
 - Adopt uv as installer (#1142)
 
+
 ## [integrations/chroma-v0.22.1] - 2024-09-30
 
-### Chroma
+### 🌀 Miscellaneous
 
 - Empty filters should behave as no filters (#1117)
 
@@ -26,32 +49,36 @@
 
 ### 🐛 Bug Fixes
 
-- Fix chroma linting; rm numpy (#1063)
-
-Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
+- Refactor: fix chroma linting; do not use numpy (#1063)
 - Filters in chroma integration (#1072)
 
 ### 🧪 Testing
 
 - Do not retry tests in `hatch run test` command (#954)
 
-### ⚙️ Miscellaneous Tasks
+### 🧹 Chores
 
 - Chroma - ruff update, don't ruff tests (#983)
 - Update ruff linting scripts and settings (#1105)
 
+### 🌀 Miscellaneous
+
+- Chore: ChromaDocumentStore lint fix (#1065)
+
 ## [integrations/chroma-v0.21.1] - 2024-07-17
 
 ### 🐛 Bug Fixes
 
 - `ChromaDocumentStore` - discard `meta` items when the type of their value is not supported in Chroma (#907)
 
+
 ## [integrations/chroma-v0.21.0] - 2024-07-16
 
 ### 🚀 Features
 
 - Add metadata parameter to ChromaDocumentStore. (#906)
 
+
 ## [integrations/chroma-v0.20.1] - 2024-07-15
 
 ### 🚀 Features
@@ -64,15 +91,32 @@ Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
 - Allow search in ChromaDocumentStore without metadata (#863)
 - `Chroma` - Fallback to default filter policy when deserializing retrievers without the init parameter (#897)
 
-### ⚙️ Miscellaneous Tasks
+### ⚙️ CI
 
 - Retry tests to reduce flakyness (#836)
+
+### 🧹 Chores
+
 - Update ruff invocation to include check parameter (#853)
 
+### 🌀 Miscellaneous
+
+- Ci: install `pytest-rerunfailures` where needed; add retry config to `test-cov` script (#845)
+- Chore: Minor retriever pydoc fix (#884)
+
 ## [integrations/chroma-v0.18.0] - 2024-05-31
 
+### 🌀 Miscellaneous
+
+- Chore: pin chromadb>=0.5.0 (#777)
+
 ## [integrations/chroma-v0.17.0] - 2024-05-10
 
+### 🌀 Miscellaneous
+
+- Chore: change the pydoc renderer class (#718)
+- Implement filters for chromaQueryTextRetriever via existing haystack filters logic (#705)
+
 ## [integrations/chroma-v0.16.0] - 2024-05-02
 
 ### 📚 Documentation
@@ -80,15 +124,27 @@ Co-authored-by: Silvano Cerza <3314350+silvanocerza@users.noreply.github.com>
 - Small consistency improvements (#536)
 - Disable-class-def (#556)
 
+### 🌀 Miscellaneous
+
+- Make tests show coverage (#566)
+- Refactor tests (#574)
+- Remove references to Python 3.7 (#601)
+- Make Document Stores initially skip `SparseEmbedding` (#606)
+- Pin databind-core (#619)
+- Chore: add license classifiers (#680)
+- Feature/bump chromadb dep to 0.5.0 (#700)
+
 ## [integrations/chroma-v0.15.0] - 2024-03-01
 
+### 🌀 Miscellaneous
+
+- Release chroma on python 3.8 (#512)
+
 ## [integrations/chroma-v0.14.0] - 2024-02-29
 
 ### 🐛 Bug Fixes
 
 - Fix order of API docs (#447)
-
-This PR will also push the docs to Readme
 - Serialize the path to the local db (#506)
 
 ### 📚 Documentation
@@ -96,35 +152,59 @@ This PR will also push the docs to Readme
 - Update category slug (#442)
 - Review chroma integration (#501)
 
+### 🌀 Miscellaneous
+
+- Small improvements (#443)
+- Fix: make write_documents compatible with the DocumentStore protocol (#505)
+
 ## [integrations/chroma-v0.13.0] - 2024-02-13
 
+### 🌀 Miscellaneous
+
+- Chroma: rename retriever (#407)
+
 ## [integrations/chroma-v0.12.0] - 2024-02-06
 
 ### 🚀 Features
 
 - Generate API docs (#262)
 
+### 🌀 Miscellaneous
+
+- Add typing_extensions pin to Chroma integration (#295)
+- Allows filters and persistent document stores for Chroma (#342)
+
 ## [integrations/chroma-v0.11.0] - 2024-01-18
 
 ### 🐛 Bug Fixes
 
 - Chroma DocumentStore creation for pre-existing collection name (#157)
 
+### 🌀 Miscellaneous
+
+- Mount chroma integration under `haystack_integrations.*` (#193)
+- Remove ChromaSingleQueryRetriever (#240)
+
 ## [integrations/chroma-v0.9.0] - 2023-12-20
 
 ### 🐛 Bug Fixes
 
-- Fix project urls (#96)
+- Fix project URLs (#96)
 
 ### 🚜 Refactor
 
 - Use `hatch_vcs` to manage integrations versioning (#103)
 
+### 🌀 Miscellaneous
+
+- Chore: pin chroma version (#104)
+- Fix: update to the latest Document format (#127)
+
 ## [integrations/chroma-v0.8.1] - 2023-12-05
 
 ### 🐛 Bug Fixes
 
-- Fix import and increase version (#77)
+- Document Stores: fix protocol import (#77)
 
 ## [integrations/chroma-v0.8.0] - 2023-12-04
 
@@ -132,4 +212,11 @@ This PR will also push the docs to Readme
 
 - Fix license headers
 
+### 🌀 Miscellaneous
+
+- Reorganize repository (#62)
+- Update import paths (#64)
+- Patch chroma filters tests (#67)
+- Remove Document Store decorator (#76)
+
 <!-- generated by git-cliff -->

{chroma_haystack-2.0.0 → chroma_haystack-2.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chroma-haystack
-Version: 2.0.0
+Version: 2.0.1
 Project-URL: Documentation, https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/chroma#readme
 Project-URL: Issues, https://github.com/deepset-ai/haystack-core-integrations/issues
 Project-URL: Source, https://github.com/deepset-ai/haystack-core-integrations/tree/main/integrations/chroma

{chroma_haystack-2.0.0 → chroma_haystack-2.0.1}/src/haystack_integrations/document_stores/chroma/document_store.py

@@ -9,6 +9,7 @@ from chromadb.api.types import GetResult, QueryResult
 from haystack import default_from_dict, default_to_dict
 from haystack.dataclasses import Document
 from haystack.document_stores.types import DuplicatePolicy
+from numpy import ndarray
 
 from .filters import _convert_filters
 from .utils import get_embedding_function
@@ -208,18 +209,18 @@ class ChromaDocumentStore:
         self._ensure_initialized()
         assert self._collection is not None
 
+        kwargs: Dict[str, Any] = {"include": ["embeddings", "documents", "metadatas"]}
+
         if filters:
             chroma_filter = _convert_filters(filters)
-            kwargs: Dict[str, Any] = {"where": chroma_filter.where}
+            kwargs["where"] = chroma_filter.where
 
             if chroma_filter.ids:
                 kwargs["ids"] = chroma_filter.ids
             if chroma_filter.where_document:
                 kwargs["where_document"] = chroma_filter.where_document
 
-            result = self._collection.get(**kwargs)
-        else:
-            result = self._collection.get()
+        result = self._collection.get(**kwargs)
 
         return self._get_result_to_documents(result)
 
@@ -249,11 +250,19 @@ class ChromaDocumentStore:
             if doc.content is None:
                 logger.warning(
                     "ChromaDocumentStore cannot store documents with `content=None`. "
-                    "`array`, `dataframe` and `blob` are not supported. "
                     "Document with id %s will be skipped.",
                     doc.id,
                 )
                 continue
+            elif (hasattr(doc, "dataframe") and doc.dataframe is not None) or (
+                hasattr(doc, "blob") and doc.blob is not None
+            ):
+                logger.warning(
+                    "Document with id %s contains `dataframe` or `blob` fields. "
+                    "ChromaDocumentStore cannot store `dataframe` or `blob` fields. "
+                    "These fields will be ignored. ",
+                    doc.id,
+                )
             data = {"ids": [doc.id], "documents": [doc.content]}
 
             if doc.meta:
@@ -416,8 +425,11 @@ class ChromaDocumentStore:
                 document_dict["meta"] = result_metadata[i]
 
             result_embeddings = result.get("embeddings")
-            if result_embeddings:
-                document_dict["embedding"] = list(result_embeddings[i])
+            if result_embeddings is not None:
+                if isinstance(result_embeddings[i], ndarray):
+                    document_dict["embedding"] = result_embeddings[i].tolist()
+                else:
+                    document_dict["embedding"] = result_embeddings[i]
 
             retval.append(Document.from_dict(document_dict))
 

{chroma_haystack-2.0.0 → chroma_haystack-2.0.1}/tests/test_document_store.py

@@ -11,7 +11,7 @@ from unittest import mock
 import numpy as np
 import pytest
 from chromadb.api.types import Documents, EmbeddingFunction, Embeddings
-from haystack import Document
+from haystack.dataclasses import ByteStream, Document
 from haystack.testing.document_store import (
     TEST_EMBEDDING_1,
     TEST_EMBEDDING_2,
@@ -179,7 +179,13 @@ class TestDocumentStore(CountDocumentsTest, DeleteDocumentsTest, FilterDocuments
         document_store.write_documents([doc])
         document_store.delete_documents(["non_existing"])
         filters = {"operator": "==", "field": "id", "value": doc.id}
-        assert document_store.filter_documents(filters=filters) == [doc]
+
+        assert document_store.filter_documents(filters=filters)[0].id == doc.id
+
+    def test_filter_documents_return_embeddings(self, document_store: ChromaDocumentStore):
+        document_store.write_documents([Document(content="test doc", embedding=TEST_EMBEDDING_1)])
+
+        assert document_store.filter_documents()[0].embedding == pytest.approx(TEST_EMBEDDING_1)
 
     def test_search(self):
         document_store = ChromaDocumentStore()
@@ -227,6 +233,20 @@ class TestDocumentStore(CountDocumentsTest, DeleteDocumentsTest, FilterDocuments
         assert written_docs[1].meta == {}
         assert written_docs[2].meta == {"ok": 123}
 
+    def test_documents_with_content_none_are_not_stored(self, document_store: ChromaDocumentStore):
+        document_store.write_documents([Document(content=None)])
+        assert document_store.filter_documents() == []
+
+    def test_dataframe_and_blob_are_not_stored(self, document_store: ChromaDocumentStore):
+        bs = ByteStream(data=b"test")
+        doc_mixed = Document(content="test", blob=bs)
+        doc_mixed.dataframe = "non null content"  # just a placeholder
+
+        document_store.write_documents([doc_mixed])
+
+        expected_doc = Document(id=doc_mixed.id, content="test")
+        assert document_store.filter_documents() == [expected_doc]
+
     @pytest.mark.integration
     def test_to_dict(self, request):
         ds = ChromaDocumentStore(
@@ -406,18 +426,6 @@ class TestDocumentStore(CountDocumentsTest, DeleteDocumentsTest, FilterDocuments
             ],
         )
 
-    @pytest.mark.skip(reason="Filter on dataframe contents is not supported.")
-    def test_comparison_equal_with_dataframe(
-        self, document_store: ChromaDocumentStore, filterable_docs: List[Document]
-    ):
-        pass
-
-    @pytest.mark.skip(reason="Filter on dataframe contents is not supported.")
-    def test_comparison_not_equal_with_dataframe(
-        self, document_store: ChromaDocumentStore, filterable_docs: List[Document]
-    ):
-        pass
-
     @pytest.mark.skip(reason="Chroma does not support comparison with null values")
     def test_comparison_equal_with_none(self, document_store, filterable_docs):
         pass