🎨 apply linting

Author: ajndkr

demo/README.md

@@ -41,7 +41,6 @@ http://localhost:8000/chat
 
 ![demo](../assets/demo.gif)
 
-
 ### Sample cURL request for retrieval qa with sources
 
 ```bash

demo/app.py

@@ -8,7 +8,10 @@
 from langchain.chat_models import ChatOpenAI
 from pydantic import BaseModel
 
-from fastapi_async_langchain.responses import LLMChainStreamingResponse, RetrievalQAStreamingResponse
+from fastapi_async_langchain.responses import (
+    LLMChainStreamingResponse,
+    RetrievalQAStreamingResponse,
+)
 
 load_dotenv()
 
@@ -46,32 +49,38 @@ async def chat(
         chain, request.query, media_type="text/event-stream"
     )
 
+
 def retrieval_qa_chain():
     from langchain.chains import RetrievalQAWithSourcesChain
     from langchain.chains.qa_with_sources import load_qa_with_sources_chain
     from langchain.chains.qa_with_sources.stuff_prompt import PROMPT as QA_PROMPT
-    from langchain.vectorstores import FAISS
     from langchain.embeddings import OpenAIEmbeddings
+    from langchain.vectorstores import FAISS
 
     callback_manager = AsyncCallbackManager([])
-    vectorstore = FAISS.load_local(index_name="langchain-python", embeddings=OpenAIEmbeddings(), folder_path="demo/")
+    vectorstore = FAISS.load_local(
+        index_name="langchain-python",
+        embeddings=OpenAIEmbeddings(),
+        folder_path="demo/",
+    )
     retriever = vectorstore.as_retriever()
-    streaming_llm = ChatOpenAI(streaming=True, callback_manager=callback_manager, verbose=True, temperature=0)
-    doc_chain = load_qa_with_sources_chain(llm=streaming_llm,
-                                           chain_type="stuff",
-                                           prompt=QA_PROMPT)
-    return RetrievalQAWithSourcesChain(combine_documents_chain=doc_chain,
-                                       retriever=retriever,
-                                       callback_manager=callback_manager,
-                                       return_source_documents=True,
-                                       verbose=True)
+    streaming_llm = ChatOpenAI(
+        streaming=True, callback_manager=callback_manager, verbose=True, temperature=0
+    )
+    doc_chain = load_qa_with_sources_chain(
+        llm=streaming_llm, chain_type="stuff", prompt=QA_PROMPT
+    )
+    return RetrievalQAWithSourcesChain(
+        combine_documents_chain=doc_chain,
+        retriever=retriever,
+        callback_manager=callback_manager,
+        return_source_documents=True,
+        verbose=True,
+    )
+
 
 @app.post("/retrieval-qa-with-sources")
-async def retrieval_qa_with_sources(
-    request: Request
-) -> RetrievalQAStreamingResponse:
+async def retrieval_qa_with_sources(request: Request) -> RetrievalQAStreamingResponse:
     return RetrievalQAStreamingResponse(
-        chain=retrieval_qa_chain(),
-        inputs=request.query,
-        media_type="text/event-stream"
+        chain=retrieval_qa_chain(), inputs=request.query, media_type="text/event-stream"
     )

fastapi_async_langchain/callbacks/retrieval_qa.py

@@ -1,14 +1,16 @@
 from typing import Any, Dict
+
 from .base import AsyncFastApiStreamingCallback
 
+
 class RetrievalQAFastApiStreamingCallback(AsyncFastApiStreamingCallback):
     """Async Callback handler for FastAPI StreamingResponse to RetrievalQA."""
 
     async def on_chain_end(self, outputs: Dict[str, Any], **kwargs: Any) -> None:
         """Run when chain ends running."""
-        if not outputs['source_documents'] is None:
+        if outputs["source_documents"] is not None:
             await self.send("\n\nSOURCE DOCUMENTS: \n")
-            if not outputs['source_documents'] is None:
-                for doc in outputs['source_documents']:
+            if outputs["source_documents"] is not None:
+                for doc in outputs["source_documents"]:
                     await self.send(f"page content: {doc.page_content} \n")
                     await self.send(f"source: {doc.metadata['source']} \n\n")