File size: 4,060 Bytes
f8817e5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os 
import base64
from io import BytesIO
from PIL import Image
import streamlit as st
from app_config import SYSTEM_PROMPT,MODEL,MAX_TOKENS,TRANSFORMER_MODEL
from langchain.memory import ConversationSummaryBufferMemory
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_groq import ChatGroq
from streamlit_pdf_viewer import pdf_viewer
from pydantic import BaseModel
from langchain.chains import LLMChain
from langchain.prompts import ChatPromptTemplate
from langchain_community.vectorstores import FAISS
from sentence_transformers import SentenceTransformer 
from typing import Any

st.title("Hitachi Support Bot")

class Element(BaseModel):
    type: str
    text: Any

# llm = ChatGoogleGenerativeAI(
#     model=MODEL,
#     max_tokens=MAX_TOKENS
# )
llm = ChatGroq(model=MODEL,api_key='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')


prompt = ChatPromptTemplate.from_template(SYSTEM_PROMPT)
qa_chain = LLMChain(llm=llm,prompt=prompt)
embeddings = HuggingFaceEmbeddings(model_name=TRANSFORMER_MODEL)
db = FAISS.load_local("faiss_index",embeddings,allow_dangerous_deserialization=True)

st.markdown(
    """

<style>

    .st-emotion-cache-janbn0 {

        flex-direction: row-reverse;

        text-align: right;

    }

</style>

""",
    unsafe_allow_html=True,
)

def response_generator(question):
    relevant_docs = db.similarity_search_with_relevance_scores(question,k=5)
    context = ""
    relevant_images = []
    for d,score in relevant_docs:
        if score > 0:
            if d.metadata['type'] == 'text':
                context += str(d.metadata['original_content'])
            elif d.metadata['type'] == 'table':
                context += str(d.metadata['original_content'])
            elif d.metadata['type'] == 'image':
                context += d.page_content
                relevant_images.append(d.metadata['original_content'])
    result =  qa_chain.run({'context':context,"question":question})
    return result,relevant_images

with st.sidebar:
    st.header("Hitachi Support Bot")
    button = st.toggle("View Doc file.")

if button:
    pdf_viewer("GPT OUTPUT.pdf")
else:
    if "messages" not in st.session_state:
        st.session_state.messages=[{"role": "system", "content": SYSTEM_PROMPT}]
        
    if "llm" not in st.session_state:
        st.session_state.llm = llm
    if "rag_memory" not in st.session_state:
        st.session_state.rag_memory = ConversationSummaryBufferMemory(llm=st.session_state.llm, max_token_limit= 5000)

    container =  st.container(height=700)
    for message in st.session_state.messages:
        if message["role"] != "system":
            if message["role"] == "user":
                with container.chat_message(message["role"]):
                    st.write(message["content"])
            if message["role"] == "assistant":
                with container.chat_message(message["role"]):
                    st.write(message["content"])
                    for i in range(len(message["images"])):
                        st.image(Image.open(BytesIO(base64.b64decode(message["images"][i].encode('utf-8')))))

    if prompt := st.chat_input("Enter your query here... "):
        with container.chat_message("user"):
            st.write(prompt)
        st.session_state.messages.append({"role":"user" , "content":prompt})
        with container.chat_message("assistant"):  
            response,images = response_generator(prompt)
            st.write(response)
            for i in range(len(images)):
                st.markdown("""---""")
                st.image(Image.open(BytesIO(base64.b64decode(images[i].encode('utf-8')))))
            st.markdown("""---""")

                
        st.session_state.rag_memory.save_context({'input': prompt}, {'output': response})
        st.session_state.messages.append({"role":"assistant" , "content":response,'images':images})