From 77fc2f76442b922764436cc267220a5df8b7a022 Mon Sep 17 00:00:00 2001
From: Sian Cao <yinshuiboy@gmail.com>
Date: Thu, 19 Oct 2023 14:51:28 +0800
Subject: [PATCH] fix: impl missing embeddings method (#10823)

FAISS does not implement embeddings method and use embed_query to
embedding texts which is wrong for some embedding models.

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
---
 .../langchain/langchain/vectorstores/faiss.py | 48 ++++++++++++++-----
 1 file changed, 37 insertions(+), 11 deletions(-)

diff --git a/libs/langchain/langchain/vectorstores/faiss.py b/libs/langchain/langchain/vectorstores/faiss.py
index 2fc3bb68eb..0b867d2709 100644
--- a/libs/langchain/langchain/vectorstores/faiss.py
+++ b/libs/langchain/langchain/vectorstores/faiss.py
@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import logging
 import operator
 import os
 import pickle
@@ -15,6 +16,7 @@ from typing import (
     Optional,
     Sized,
     Tuple,
+    Union,
 )
 
 import numpy as np
@@ -26,6 +28,8 @@ from langchain.schema.embeddings import Embeddings
 from langchain.schema.vectorstore import VectorStore
 from langchain.vectorstores.utils import DistanceStrategy, maximal_marginal_relevance
 
+logger = logging.getLogger(__name__)
+
 
 def dependable_faiss_import(no_avx2: Optional[bool] = None) -> Any:
     """
@@ -82,7 +86,7 @@ class FAISS(VectorStore):
 
     def __init__(
         self,
-        embedding_function: Callable,
+        embedding_function: Union[Callable, Embeddings],
         index: Any,
         docstore: Docstore,
         index_to_docstore_id: Dict[int, str],
@@ -91,6 +95,11 @@ class FAISS(VectorStore):
         distance_strategy: DistanceStrategy = DistanceStrategy.EUCLIDEAN_DISTANCE,
     ):
         """Initialize with necessary components."""
+        if not isinstance(embedding_function, Embeddings):
+            logger.warning(
+                "`embedding_function` is expected to be an Embeddings object, support "
+                "for passing in a function will soon be removed."
+            )
         self.embedding_function = embedding_function
         self.index = index
         self.docstore = docstore
@@ -108,6 +117,26 @@ class FAISS(VectorStore):
                 )
             )
 
+    @property
+    def embeddings(self) -> Optional[Embeddings]:
+        return (
+            self.embedding_function
+            if isinstance(self.embedding_function, Embeddings)
+            else None
+        )
+
+    def _embed_documents(self, texts: List[str]) -> List[List[float]]:
+        if isinstance(self.embedding_function, Embeddings):
+            return self.embedding_function.embed_documents(texts)
+        else:
+            return [self.embedding_function(text) for text in texts]
+
+    def _embed_query(self, text: str) -> List[float]:
+        if isinstance(self.embedding_function, Embeddings):
+            return self.embedding_function.embed_query(text)
+        else:
+            return self.embedding_function(text)
+
     def __add(
         self,
         texts: Iterable[str],
@@ -163,7 +192,8 @@ class FAISS(VectorStore):
         Returns:
             List of ids from adding the texts into the vectorstore.
         """
-        embeddings = [self.embedding_function(text) for text in texts]
+        texts = list(texts)
+        embeddings = self._embed_documents(texts)
         return self.__add(texts, embeddings, metadatas=metadatas, ids=ids)
 
     def add_embeddings(
@@ -272,7 +302,7 @@ class FAISS(VectorStore):
             List of documents most similar to the query text with
             L2 distance in float. Lower score represents more similarity.
         """
-        embedding = self.embedding_function(query)
+        embedding = self._embed_query(query)
         docs = self.similarity_search_with_score_by_vector(
             embedding,
             k,
@@ -465,7 +495,7 @@ class FAISS(VectorStore):
         Returns:
             List of Documents selected by maximal marginal relevance.
         """
-        embedding = self.embedding_function(query)
+        embedding = self._embed_query(query)
         docs = self.max_marginal_relevance_search_by_vector(
             embedding,
             k=k,
@@ -561,7 +591,7 @@ class FAISS(VectorStore):
             # Default to L2, currently other metric types not initialized.
             index = faiss.IndexFlatL2(len(embeddings[0]))
         vecstore = cls(
-            embedding.embed_query,
+            embedding,
             index,
             InMemoryDocstore(),
             {},
@@ -696,9 +726,7 @@ class FAISS(VectorStore):
         # load docstore and index_to_docstore_id
         with open(path / "{index_name}.pkl".format(index_name=index_name), "rb") as f:
             docstore, index_to_docstore_id = pickle.load(f)
-        return cls(
-            embeddings.embed_query, index, docstore, index_to_docstore_id, **kwargs
-        )
+        return cls(embeddings, index, docstore, index_to_docstore_id, **kwargs)
 
     def serialize_to_bytes(self) -> bytes:
         """Serialize FAISS index, docstore, and index_to_docstore_id to bytes."""
@@ -713,9 +741,7 @@ class FAISS(VectorStore):
     ) -> FAISS:
         """Deserialize FAISS index, docstore, and index_to_docstore_id from bytes."""
         index, docstore, index_to_docstore_id = pickle.loads(serialized)
-        return cls(
-            embeddings.embed_query, index, docstore, index_to_docstore_id, **kwargs
-        )
+        return cls(embeddings, index, docstore, index_to_docstore_id, **kwargs)
 
     def _select_relevance_score_fn(self) -> Callable[[float], float]:
         """