Skip to content

The agent engineering platform.

[VectorStore] Improvement: Improve chroma vector store (#28524)

ZhangShenaod26555c68277
diff --git a/libs/partners/chroma/langchain_chroma/vectorstores.py b/libs/partners/chroma/langchain_chroma/vectorstores.py
index 9967ac6d29..afb9191c60 100644
--- a/libs/partners/chroma/langchain_chroma/vectorstores.py
+++ b/libs/partners/chroma/langchain_chroma/vectorstores.py
@@ -758,7 +758,7 @@ class Chroma(VectorStore):
The most similar documents will have the lowest relevance score. Default
relevance score function is euclidean distance. Distance metric must be
- provided in `collection_metadata` during initizalition of Chroma object.
+ provided in `collection_metadata` during initialization of Chroma object.
Example: collection_metadata={"hnsw:space": "cosine"}. Available distance
metrics are: 'cosine', 'l2' and 'ip'.
diff --git a/libs/partners/chroma/tests/unit_tests/test_vectorstores.py b/libs/partners/chroma/tests/unit_tests/test_vectorstores.py
index 84d8637879..66ac3b0622 100644
--- a/libs/partners/chroma/tests/unit_tests/test_vectorstores.py
+++ b/libs/partners/chroma/tests/unit_tests/test_vectorstores.py
@@ -13,3 +13,18 @@ def test_initialization() -> None:
texts=texts,
embedding=FakeEmbeddings(size=10),
)
+
+
+def test_similarity_search() -> None:
+ """Test similarity search by Chroma."""
+ texts = ["foo", "bar", "baz"]
+ metadatas = [{"page": str(i)} for i in range(len(texts))]
+ docsearch = Chroma.from_texts(
+ collection_name="test_collection",
+ texts=texts,
+ embedding=FakeEmbeddings(size=10),
+ metadatas=metadatas,
+ )
+ output = docsearch.similarity_search("foo", k=1)
+ docsearch.delete_collection()
+ assert len(output) == 1