test file_research

app.py

@@ -0,0 +1,109 @@
+import streamlit as st
+import os
+import time
+from dotenv import load_dotenv
+from getpass import getpass
+from langchain.llms import replicate
+from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
+from langchain.prompts import PromptTemplate
+
+from PyPDF2 import PdfReader
+from streamlit_extras.add_vertical_space import add_vertical_space
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+#from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.vectorstores import faiss
+
+load_dotenv()
+REPLICATE_API_TOKEN = os.environ.get("REPLICATE_API_TOKEN")
+
+with st.sidebar:
+    st.title("File Research using LLM")
+    st.markdown(''' Upload your file and ask questions and do Research''')
+    add_vertical_space(5)
+    pdf=st.file_uploader('Upload your file (PDF)', type='pdf')
+    if pdf is not None:
+        pdf_reader=PdfReader(pdf)
+        text=""
+        for page in pdf_reader.pages:
+            text+=page.extract_text()
+        text_splitter=RecursiveCharacterTextSplitter(
+            chunk_size=1000,
+            chunk_overlap=200,
+            length_function=len
+        )
+        chunks=text_splitter.split_text(text)
+        
+        
+           
+        
+    st.write('Made by ALOK')
+    
+
+def main():
+    st.header('Talk to your file')
+    os.environ["REPLICATE_API_TOKEN"]=REPLICATE_API_TOKEN
+    #embeddings=OpenAIEmbeddings()
+    #vectorstore=faiss.FAISS.from_texts(chunks, embedding=embeddings)
+    
+    # The meta/llama-2-70b-chat model can stream output as it's running.
+
+    
+    if "messages" not in st.session_state:
+        st.session_state.messages = []
+
+    # Display chat messages from history on app rerun
+    for message in st.session_state.messages:
+        with st.chat_message(message["role"]):
+            st.markdown(message["content"])
+
+    # React to user input
+    if prompt := st.chat_input("Type Here"):
+        # Display user message in chat message container
+        st.chat_message("user").markdown(prompt)
+        # Add user message to chat history
+        st.session_state.messages.append({"role": "user", "content": prompt})
+        replite_api='r8_4fktoXrDGkgHY8uw1XlVtQJKQlAILKv0iBmPI'
+        
+        
+        # rep = replicate.Client(api_token=replite_api)
+        # output = replicate.run(
+        # "meta/llama-2-70b-chat:02e509c789964a7ea8736978a43525956ef40397be9033abf9fd2badfe68c9e3",
+        # input={"prompt": prompt}
+        # )
+        
+        model="meta/llama-2-70b-chat:02e509c789964a7ea8736978a43525956ef40397be9033abf9fd2badfe68c9e3"
+        llm=replicate.Replicate(
+            streaming=True,
+            callbacks=[StreamingStdOutCallbackHandler()],
+            model=model,
+            model_kwargs={"temperature": 0.75, "max_length": 500, "top_p": 1},
+            replicate_api_token=REPLICATE_API_TOKEN
+        )
+        prompt = """
+        User: Answer the following yes/no question by reasoning step by step. Please don't provide incomplete answer. Can a dog drive a car?
+        Assistant:
+        """
+        
+        
+        # Display assistant response in chat message container
+        with st.chat_message("assistant"):
+            message_placeholder = st.empty()
+        message_placeholder.markdown(llm(prompt) + "▌")
+            
+        # # The predict method returns an iterator, and you can iterate over that output.
+        # response_till_now=''
+        # for item in output:
+        #     response_till_now+=item
+        #     time.sleep(0.03)
+        #     message_placeholder.markdown(response_till_now + "▌")
+        # message_placeholder.markdown(response_till_now)
+        # response = f"AI: {response_till_now}"
+        
+        
+        # Add assistant response to chat history
+        # st.session_state.messages.append({"role": "assistant", "content": response})
+            # https://replicate.com/meta/llama-2-70b-chat/versions/02e509c789964a7ea8736978a43525956ef40397be9033abf9fd2badfe68c9e3/api#output-schema
+            #print(item, end="")
+    
+if __name__=='__main__':
+    main()

requirements.txt

@@ -0,0 +1,7 @@
+langchain
+PyPDF2
+streamlit
+replicate
+python-dotenv
+faiss-cpu
+streamlit-extras