Support similarity search by vector (in FAISS) (#961)

Alternate implementation to PR #960 Again - only FAISS is implemented.
If accepted can add this to other vectorstores or leave as
NotImplemented? Suggestions welcome...

docs/modules/utils/combine_docs_examples/vectorstores.ipynb

@@ -297,6 +297,26 @@
     "docs_and_scores[0]"
    ]
   },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "id": "d5170563",
+   "metadata": {},
+   "source": [
+    "It is also possible to do a search for documents similar to a given embedding vector using `similarity_search_by_vector` which accepts an embedding vector as a parameter instead of a string."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7675b0aa",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "embedding_vector = embeddings.embed_query(query)\n",
+    "docs_and_scores = docsearch.similarity_search_by_vector(embedding_vector)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "b386dbb8",

langchain/vectorstores/base.py

@@ -31,6 +31,20 @@ def similarity_search(
     ) -> List[Document]:
         """Return docs most similar to query."""
 
+    def similarity_search_by_vector(
+        self, embedding: List[float], k: int = 4, **kwargs: Any
+    ) -> List[Document]:
+        """Return docs most similar to embedding vector.
+
+        Args:
+            embedding: Embedding to look up documents similar to.
+            k: Number of Documents to return. Defaults to 4.
+
+        Returns:
+            List of Documents most similar to the query vector.
+        """
+        raise NotImplementedError
+
     def max_marginal_relevance_search(
         self, query: str, k: int = 4, fetch_k: int = 20
     ) -> List[Document]:
@@ -49,6 +63,24 @@ def max_marginal_relevance_search(
         """
         raise NotImplementedError
 
+    def max_marginal_relevance_search_by_vector(
+        self, embedding: List[float], k: int = 4, fetch_k: int = 20
+    ) -> List[Document]:
+        """Return docs selected using the maximal marginal relevance.
+
+        Maximal marginal relevance optimizes for similarity to query AND diversity
+        among selected documents.
+
+        Args:
+            embedding: Embedding to look up documents similar to.
+            k: Number of Documents to return. Defaults to 4.
+            fetch_k: Number of Documents to fetch to pass to MMR algorithm.
+
+        Returns:
+            List of Documents selected by maximal marginal relevance.
+        """
+        raise NotImplementedError
+
     @classmethod
     def from_documents(
         cls,

langchain/vectorstores/faiss.py

@@ -92,8 +92,8 @@ def add_texts(
         self.index_to_docstore_id.update(index_to_id)
         return [_id for _, _id, _ in full_info]
 
-    def similarity_search_with_score(
-        self, query: str, k: int = 4
+    def similarity_search_with_score_by_vector(
+        self, embedding: List[float], k: int = 4
     ) -> List[Tuple[Document, float]]:
         """Return docs most similar to query.
 
@@ -104,7 +104,6 @@ def similarity_search_with_score(
         Returns:
             List of Documents most similar to the query and score for each
         """
-        embedding = self.embedding_function(query)
         scores, indices = self.index.search(np.array([embedding], dtype=np.float32), k)
         docs = []
         for j, i in enumerate(indices[0]):
@@ -118,6 +117,37 @@ def similarity_search_with_score(
             docs.append((doc, scores[0][j]))
         return docs
 
+    def similarity_search_with_score(
+        self, query: str, k: int = 4
+    ) -> List[Tuple[Document, float]]:
+        """Return docs most similar to query.
+
+        Args:
+            query: Text to look up documents similar to.
+            k: Number of Documents to return. Defaults to 4.
+
+        Returns:
+            List of Documents most similar to the query and score for each
+        """
+        embedding = self.embedding_function(query)
+        docs = self.similarity_search_with_score_by_vector(embedding, k)
+        return docs
+
+    def similarity_search_by_vector(
+        self, embedding: List[float], k: int = 4, **kwargs: Any
+    ) -> List[Document]:
+        """Return docs most similar to embedding vector.
+
+        Args:
+            embedding: Embedding to look up documents similar to.
+            k: Number of Documents to return. Defaults to 4.
+
+        Returns:
+            List of Documents most similar to the embedding.
+        """
+        docs_and_scores = self.similarity_search_with_score_by_vector(embedding, k)
+        return [doc for doc, _ in docs_and_scores]
+
     def similarity_search(
         self, query: str, k: int = 4, **kwargs: Any
     ) -> List[Document]:
@@ -133,27 +163,28 @@ def similarity_search(
         docs_and_scores = self.similarity_search_with_score(query, k)
         return [doc for doc, _ in docs_and_scores]
 
-    def max_marginal_relevance_search(
-        self, query: str, k: int = 4, fetch_k: int = 20
+    def max_marginal_relevance_search_by_vector(
+        self, embedding: List[float], k: int = 4, fetch_k: int = 20
     ) -> List[Document]:
         """Return docs selected using the maximal marginal relevance.
 
         Maximal marginal relevance optimizes for similarity to query AND diversity
         among selected documents.
 
         Args:
-            query: Text to look up documents similar to.
+            embedding: Embedding to look up documents similar to.
             k: Number of Documents to return. Defaults to 4.
             fetch_k: Number of Documents to fetch to pass to MMR algorithm.
 
         Returns:
             List of Documents selected by maximal marginal relevance.
         """
-        embedding = self.embedding_function(query)
         _, indices = self.index.search(np.array([embedding], dtype=np.float32), fetch_k)
         # -1 happens when not enough docs are returned.
         embeddings = [self.index.reconstruct(int(i)) for i in indices[0] if i != -1]
-        mmr_selected = maximal_marginal_relevance(embedding, embeddings, k=k)
+        mmr_selected = maximal_marginal_relevance(
+            np.array([embedding], dtype=np.float32), embeddings, k=k
+        )
         selected_indices = [indices[0][i] for i in mmr_selected]
         docs = []
         for i in selected_indices:
@@ -164,6 +195,26 @@ def max_marginal_relevance_search(
             docs.append(doc)
         return docs
 
+    def max_marginal_relevance_search(
+        self, query: str, k: int = 4, fetch_k: int = 20
+    ) -> List[Document]:
+        """Return docs selected using the maximal marginal relevance.
+
+        Maximal marginal relevance optimizes for similarity to query AND diversity
+        among selected documents.
+
+        Args:
+            query: Text to look up documents similar to.
+            k: Number of Documents to return. Defaults to 4.
+            fetch_k: Number of Documents to fetch to pass to MMR algorithm.
+
+        Returns:
+            List of Documents selected by maximal marginal relevance.
+        """
+        embedding = self.embedding_function(query)
+        docs = self.max_marginal_relevance_search_by_vector(embedding, k, fetch_k)
+        return docs
+
     @classmethod
     def from_texts(
         cls,

tests/integration_tests/vectorstores/test_faiss.py

@@ -27,6 +27,24 @@ def test_faiss() -> None:
     assert output == [Document(page_content="foo")]
 
 
+def test_faiss_vector_sim() -> None:
+    """Test vector similarity."""
+    texts = ["foo", "bar", "baz"]
+    docsearch = FAISS.from_texts(texts, FakeEmbeddings())
+    index_to_id = docsearch.index_to_docstore_id
+    expected_docstore = InMemoryDocstore(
+        {
+            index_to_id[0]: Document(page_content="foo"),
+            index_to_id[1]: Document(page_content="bar"),
+            index_to_id[2]: Document(page_content="baz"),
+        }
+    )
+    assert docsearch.docstore.__dict__ == expected_docstore.__dict__
+    query_vec = FakeEmbeddings().embed_query(text="foo")
+    output = docsearch.similarity_search_by_vector(query_vec, k=1)
+    assert output == [Document(page_content="foo")]
+
+
 def test_faiss_with_metadatas() -> None:
     """Test end to end construction and search."""
     texts = ["foo", "bar", "baz"]