Update app.py

app.py

@@ -45,24 +45,33 @@ qa = RetrievalQA.from_chain_type(llm=Cohere(model='command'), chain_type="stuff"
     retriever=docsearch.as_retriever(search_kwargs={'k':1}),return_source_documents=True)
 
 def predict(message, history):
-    history_langchain_format = []
-    for human, ai in history:
-        history_langchain_format.append(HumanMessage(content=human))
-        history_langchain_format.append(AIMessage(content=ai))
-    history_langchain_format.append(HumanMessage(content=message))
-    gpt_response = qa({'query':history_langchain_format})
-    return gpt_response['result']
+    '''experimenation with memory and conversation retrieval chain has resulted in less
+    performance, usefulness, and more halucination. Hence, this chat bot provides one 
+    shot answers with zero memory. You can use the code in github notebooks to do this
+    experimentation. github.com/mehrdad-es/Amazon-But-Better'''
+    # history_langchain_format = []
+    # for human, ai in history:
+    #     history_langchain_format.append(HumanMessage(content=human))
+    #     history_langchain_format.append(AIMessage(content=ai))
+    # history_langchain_format.append(HumanMessage(content=message))
+    # gpt_response = qa({'query':history_langchain_format})
+    # return gpt_response['result']
+    # gpt_response = qa({'query':''.join(history)+f'.\n given the previous conversation respond using the following prompt:{message}'})
+    # # print(gpt_response)
+    # history.append((f'HumanMessage:{message}',f'AIMessage: {gpt_response},'))
+    # # history=history_langchain_format
+    # return gpt_response['result']
     
-    # message=message+'? just the book title-Author'
-    # result = qa({"query": message})
-    # # r1=docsearch.similarity_search_with_score(query=q,k=3)
-    # # print([(item[-2].metadata,item[-1]) for item in r1],\
-    # #       '\n\n',result['result'],f'|| {result["source_documents"][0].metadata}','\n*****\n')
-    # if result['result'] not in ["I don't know","I don't know."]:
-    #     return result['result']+f'\n---\nAmazon Kindle ebook description is:\n {result["source_documents"][0].page_content}'+\
-    #     f'\nfrom this file: {result["source_documents"][0].metadata}'
-    # else:
-    #     return result['result']
+    message=message+'? just the book title-Author'
+    result = qa({"query": message})
+    # r1=docsearch.similarity_search_with_score(query=q,k=3)
+    # print([(item[-2].metadata,item[-1]) for item in r1],\
+    #       '\n\n',result['result'],f'|| {result["source_documents"][0].metadata}','\n*****\n')
+    if result['result'] not in ["I don't know","I don't know."]:
+        return result['result']+f'\n---\nAmazon Kindle ebook description is:\n {result["source_documents"][0].page_content}'+\
+        f'\nfrom this file: {result["source_documents"][0].metadata}'
+    else:
+        return result['result']
 
 gr.ChatInterface(predict,
     chatbot=gr.Chatbot(height='auto'),