change to upload from streamlit

app.py

@@ -201,19 +201,28 @@ def main():
     data = []
     # DB_FAISS_UPLOAD_PATH = "vectorstores/db_faiss"
     st.header("DOCUMENT QUESTION ANSWERING IS2")
-    directory = "data"
-    data_dir = UploadDoc(directory).create_document()
-    data.extend(data_dir)
+    # directory = "data"
+    # data_dir = UploadDoc(directory).create_document()
+    # data.extend(data_dir)
 
-    #create vector from upload 
-    if len(data) > 0 :
-        sp_docs = split_docs(documents = data)
-        st.write(f"This document have {len(sp_docs)} chunks")
+    # #create vector from upload 
+    # if len(data) > 0 :
+    #     sp_docs = split_docs(documents = data)
+    #     st.write(f"This document have {len(sp_docs)} chunks")
+    #     embeddings = load_embeddings()
+    #     with st.spinner('Wait for create vector'):
+    #         db = FAISS.from_documents(sp_docs, embeddings)
+    #         # db.save_local(DB_FAISS_UPLOAD_PATH)
+    #         # st.write(f"Your model is already store in {DB_FAISS_UPLOAD_PATH}")
+    uploaded_file = st.file_uploader('Choose your .pdf file', type="pdf")
+    print(uploaded_file)
+    if uploaded_file is not None:
         embeddings = load_embeddings()
-        with st.spinner('Wait for create vector'):
-            db = FAISS.from_documents(sp_docs, embeddings)
-            # db.save_local(DB_FAISS_UPLOAD_PATH)
-            # st.write(f"Your model is already store in {DB_FAISS_UPLOAD_PATH}")
+        pdf_reader = PdfReader(uploaded_file)
+        text = ""
+        for page in pdf_reader.pages:
+            text += page.extract_text()
+        db = FAISS.from_texts(text, embeddings)
 
     llm = load_llama2_llamaCpp()
     qa_prompt = set_custom_prompt()
@@ -221,52 +230,49 @@ def main():
     #memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
     doc_chain = load_qa_chain(llm, chain_type="stuff", prompt = qa_prompt)
     question_generator = LLMChain(llm=llm, prompt=CONDENSE_QUESTION_PROMPT)
-    qa_chain = ConversationalRetrievalChain(
-        retriever =db.as_retriever(search_type="similarity_score_threshold", search_kwargs={'k':3,  "score_threshold": 0.7}), 
-        question_generator=question_generator,
-        #condense_question_prompt=CONDENSE_QUESTION_PROMPT,
-        combine_docs_chain=doc_chain,
-        return_source_documents=True,
-        memory = memory,
-        #get_chat_history=lambda h :h
-    )
-
-    for message in st.session_state.messages:
-        with st.chat_message(message["role"]):
-            st.markdown(message["content"])
+    if db is not None :
+        qa_chain = ConversationalRetrievalChain(
+            retriever =db.as_retriever(search_type="similarity_score_threshold", search_kwargs={'k':3,  "score_threshold": 0.7}), 
+            question_generator=question_generator,
+            #condense_question_prompt=CONDENSE_QUESTION_PROMPT,
+            combine_docs_chain=doc_chain,
+            return_source_documents=True,
+            memory = memory,
+            #get_chat_history=lambda h :h
+        )
 
-        # Accept user input
-    if query := st.chat_input("What is up?"):
-        # Display user message in chat message container
-        with st.chat_message("user"):
-            st.markdown(query)
-        # Add user message to chat history
-        st.session_state.messages.append({"role": "user", "content": query})
+        for message in st.session_state.messages:
+            with st.chat_message(message["role"]):
+                st.markdown(message["content"])
 
-        start = time.time()
+            # Accept user input
+        if query := st.chat_input("What is up?"):
+            # Display user message in chat message container
+            with st.chat_message("user"):
+                st.markdown(query)
+            # Add user message to chat history
+            st.session_state.messages.append({"role": "user", "content": query})
 
-        response = qa_chain({'question': query})
+            start = time.time()
 
-        url_list = set([i.metadata['source']  for i in response['source_documents']])
-        #print(f"condensed quesion : {question_generator.run({'chat_history': response['chat_history'], 'question' : query})}")
+            response = qa_chain({'question': query})
 
-        with st.chat_message("assistant"):
-            st.markdown(response['answer'])
-        
-        end = time.time()
-        st.write("Respone time:",int(end-start),"sec")
-                
-        # Add assistant response to chat history
-        st.session_state.messages.append({"role": "assistant", "content": response['answer']})
+            url_list = set([i.metadata['source']  for i in response['source_documents']])
+            #print(f"condensed quesion : {question_generator.run({'chat_history': response['chat_history'], 'question' : query})}")
+            
+            with st.chat_message("assistant"):
+                st.markdown(response['answer'])
 
-        with st.expander("See the related documents"):
-            for count, url in enumerate(url_list):
-                #url_reg = regex_source(url)
-                st.write(str(count+1)+":", url)
+            end = time.time()
+            st.write("Respone time:",int(end-start),"sec")
+                    
+            # Add assistant response to chat history
+            st.session_state.messages.append({"role": "assistant", "content": response['answer']})
 
-    # view_messages = st.expander("View the message contents in session state")
-    # with view_messages:
-    #     view_messages.json(st.session_state.langchain_messages)
+            with st.expander("See the related documents"):
+                for count, url in enumerate(url_list):
+                    #url_reg = regex_source(url)
+                    st.write(str(count+1)+":", url)
 
     clear_button = st.button("Start new convo")
     if clear_button :