File size: 4,366 Bytes
060c9d8
 
 
 
 
 
 
 
 
 
 
 
 
fee98e6
060c9d8
 
 
156c3a1
060c9d8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c729c7f
060c9d8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1ab4f11
 
 
 
060c9d8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
import streamlit as st
from dotenv import load_dotenv
from PyPDF2 import PdfReader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings import HuggingFaceInstructEmbeddings
from langchain.vectorstores import FAISS
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationalRetrievalChain
from htmlTemplates import css, bot_template, user_template
from langchain.llms import HuggingFaceHub
import os
import numpy as np

#EMBEDDINGS_FILE = "embeddings.npy"
INDEX_FILE = "index.faiss"

def save_embeddings_and_index(index):
    #np.save(EMBEDDINGS_FILE, embeddings)
    index.save_local(INDEX_FILE)

def load_embeddings_and_index():
    if os.path.exists(INDEX_FILE):
        embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl")
        index = FAISS.load_local(INDEX_FILE, embeddings)
        return index
    return None


def get_pdf_text(pdf):
    text = ""
    pdf_reader = PdfReader(pdf)
    for page in pdf_reader.pages:
        text += page.extract_text()
    return text


def get_files(text_doc):
    text = ""
    for file in text_doc:
        if file.type == "text/plain":
            # Read the text directly from the file
            text += file.getvalue().decode("utf-8")
        elif file.type == "application/pdf":
            text += get_pdf_text(file)
    return text


def get_text_chunks(text):
    text_splitter = RecursiveCharacterTextSplitter(
        chunk_size=900,
        chunk_overlap=0,
        separators="\n",
        add_start_index = True,
        length_function= len
    )
    chunks = text_splitter.split_text(text)
    return chunks


def get_vectorstore(text_chunks, index):
    if index is None:
        embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl")
        vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
        return vectorstore
    else:
        index.add_texts(texts=text_chunks)
        return index
    


def get_conversation_chain(vectorstore):
    llm = HuggingFaceHub(repo_id="mistralai/Mistral-7B-v0.1", model_kwargs={"temperature":0.2, "max_length":1024})

    memory = ConversationBufferMemory(
        memory_key='chat_history', return_messages=True)
    conversation_chain = ConversationalRetrievalChain.from_llm(
        llm=llm,
        retriever=vectorstore.as_retriever(),
        memory=memory
    )
    return conversation_chain


def handle_userinput(user_question):
    response = st.session_state.conversation({'question': user_question})
    st.session_state.chat_history = response['chat_history']

    for i, message in enumerate(st.session_state.chat_history):
        if i % 2 == 0:
            st.write(user_template.replace(
                "{{MSG}}", message.content), unsafe_allow_html=True)
        else:
            st.write(bot_template.replace(
                "{{MSG}}", message.content), unsafe_allow_html=True)


def main():
    load_dotenv()
    st.set_page_config(page_title="ChatBot")
    st.write(css, unsafe_allow_html=True)

    if "conversation" not in st.session_state:
        index = load_embeddings_and_index()
        if index==None:
            st.session_state.conversation = None
        else:
            st.session_state.conversation = get_conversation_chain(index)
    if "chat_history" not in st.session_state:
        st.session_state.chat_history = None

    st.header("Chat Bot")
    user_question = st.text_input("Ask a question:")
    if user_question:
        handle_userinput(user_question)

    with st.sidebar:
        st.subheader("Your documents")
        pdf_docs = st.file_uploader(
            "Upload your PDFs here and click on 'Process'", accept_multiple_files=True)
        if st.button("Process"):
            with st.spinner("Processing"):
                index = load_embeddings_and_index()
                raw_text = get_files(pdf_docs)
                text_chunks = get_text_chunks(raw_text)
                # Load a new faiss index or append to existing (if it exists)
                index = get_vectorstore(text_chunks, index)
                # save updated faiss index
                save_embeddings_and_index(index)

                # create conversation chain
                st.session_state.conversation = get_conversation_chain(index)


if __name__ == '__main__':
    main()