fixed similarity_search_with_score to really use a score

updated unit test with a test for score threshold
Updated demo notebook

libs/langchain/langchain/vectorstores/vectara.py

@@ -245,6 +245,7 @@ class Vectara(VectorStore):
         k: int = 5,
         lambda_val: float = 0.025,
         filter: Optional[str] = None,
+        score_threshold: Optional[float] = None,
         n_sentence_context: int = 2,
         **kwargs: Any,
     ) -> List[Tuple[Document, float]]:
@@ -258,6 +259,8 @@ class Vectara(VectorStore):
                 filter can be "doc.rating > 3.0 and part.lang = 'deu'"} see
                 https://docs.vectara.com/docs/search-apis/sql/filter-overview
                 for more details.
+            score_threshold: minimal score thresold for the result.
+                If defined, results with score less than this value will be filtered out.
             n_sentence_context: number of sentences before/after the matching segment
                 to add, defaults to 2
 
@@ -305,7 +308,10 @@ class Vectara(VectorStore):
 
         result = response.json()
 
-        responses = result["responseSet"][0]["response"]
+        if score_threshold:
+            responses = [r for r in result["responseSet"][0]["response"] if r["score"] > score_threshold]
+        else:
+            responses = result["responseSet"][0]["response"]
         documents = result["responseSet"][0]["document"]
 
         metadatas = []
@@ -316,7 +322,7 @@ class Vectara(VectorStore):
             md.update(doc_md)
             metadatas.append(md)
 
-        docs = [
+        docs_with_score = [
             (
                 Document(
                     page_content=x["text"],
@@ -327,7 +333,7 @@ class Vectara(VectorStore):
             for x, md in zip(responses, metadatas)
         ]
 
-        return docs
+        return docs_with_score
 
     def similarity_search(
         self,
@@ -358,6 +364,7 @@ class Vectara(VectorStore):
             k=k,
             lambda_val=lambda_val,
             filter=filter,
+            score_threshold=None,
             n_sentence_context=n_sentence_context,
             **kwargs,
         )

libs/langchain/tests/integration_tests/vectorstores/test_vectara.py

@@ -68,6 +68,19 @@ def test_vectara_add_documents() -> None:
     assert output2[0].page_content == "retrieval augmented generation"
     assert output2[0].metadata["abbr"] == "rag"
 
+    # test without filter but with similarity score
+    # this is similar to the first test, but given the score threshold
+    # we only get one result
+    output3 = docsearch.similarity_search_with_score(
+        "large language model",
+        k=2,
+        score_threshold=0.1,
+        n_sentence_context=0,
+    )
+    assert len(output3) == 1
+    assert output3[0][0].page_content == "large language model"
+    assert output3[0][0].metadata["abbr"] == "llm"
+
     for doc_id in doc_ids:
         docsearch._delete_doc(doc_id)