File size: 6,128 Bytes
b379775 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 |
import streamlit as st
import os
from langchain_groq import ChatGroq
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain_core.prompts import ChatPromptTemplate
from langchain.chains import create_retrieval_chain
from langchain_community.vectorstores import FAISS
from langchain_community.document_loaders import PyPDFLoader
from dotenv import load_dotenv
from PyPDF2 import PdfReader
import time
load_dotenv()
##CSS for the background and sidebar styling
st.markdown(
"""
<style>
.stApp {
background-image: url('https://www.transparenttextures.com/patterns/white-leather.png');
background-size: cover;
}
.sidebar .sidebar-content {
padding: 20px;
background-image: url('https://www.transparenttextures.com/patterns/asfalt-light.png');
background-size: cover;
border-radius: 10px;
box-shadow: 0px 0px 10px rgba(0, 0, 0, 0.1);
}
.sidebar .bottom-button {
position: fixed;
bottom: 20px;
left: 20px;
width: calc(100% - 40px);
}
</style>
""",
unsafe_allow_html=True
)
os.environ["HF_TOKEN"] = os.getenv("HF_TOKEN")
groq_api_key = os.getenv("GROQ_API_KEY")
#documentloader -> text splitter -> embeddings -> vector store -> use retriever chains
embeddings = HuggingFaceEmbeddings(model_name = "all-MiniLM-L6-v2")
llm = ChatGroq(model = "Llama3-8b-8192",api_key = groq_api_key)
prompt_template = ChatPromptTemplate.from_template("""
Answer the following question from the provided context only.
Please provide the most accurate response based on the question
<context>
{context}
</context>
Question : {input}
""")
def get_pdf_text(pdf_docs):
text=""
for pdf in pdf_docs:
pdf_reader= PdfReader(pdf)
for page in pdf_reader.pages:
text+= page.extract_text()
return text
def create_vector_embeddings(pdfText):
if "vectors" not in st.session_state:
st.session_state.docs = get_pdf_text(pdfText)
st.session_state.splitter = RecursiveCharacterTextSplitter(chunk_size=1200,chunk_overlap=400)
st.session_state.final_docs = st.session_state.splitter.split_text(st.session_state.docs)
st.session_state.vectors = FAISS.from_texts(st.session_state.final_docs, embeddings)
if "options" not in st.session_state:
st.session_state.options = ["Select a query"]
if "user_prompt" not in st.session_state:
st.session_state.user_prompt = ""
def autopopulate_promptsbydoctype(uploaded_text):
if uploaded_text and uploaded_text[0].name.endswith("pdf"):
#autopopulate all the questions in pdf
itemsToAppend = ["get all the programme details including rights and tape content etc in pointwise manner, dont miss any info",
"give a structured short summary of the programmes and details",
"give me programme package with programme details listed"]
for itemToAppend in itemsToAppend:
if itemToAppend not in st.session_state.options:
st.session_state.options.append(itemToAppend)
st.title("Basic Document QnA")
with st.sidebar:
st.title("Menu:")
#if "uploaded_text" not in st.session_state:
st.session_state.uploaded_text = st.file_uploader("Upload your Files and Click on the Submit & Process Button", accept_multiple_files=True)
if st.button("Click To Process File"):
with st.spinner("Processing..."):
create_vector_embeddings(st.session_state.uploaded_text)
st.write("Vector Database is ready")
autopopulate_promptsbydoctype(st.session_state.uploaded_text)
# st.markdown('<div class="bottom-button">', unsafe_allow_html=True)
# params = ['docs', 'splitter','final_docs']
# if st.button("Clean Current Document Settings") and st.session_state.keys():
# with st.spinner("Cleaning In Progress...."):
# for param in params:
# if param in st.session_state:
# del st.session_state[param]
# st.session_state['uploaded_text'] = ""
# st.write("Cleanup completed..")
# st.markdown('</div>', unsafe_allow_html=True)
new_option = st.text_input("Or type your query here:")
if new_option and new_option not in st.session_state.options:
st.session_state.options.append(new_option)
st.session_state.user_prompt = new_option
if st.session_state.uploaded_text and "Technical" not in st.session_state.uploaded_text[0].name:
st.session_state.user_prompt= st.selectbox("Enter/Select your query from the document", st.session_state.options,
index=st.session_state.options.index(st.session_state.user_prompt) if st.session_state.user_prompt in st.session_state.options else 0)
if st.session_state.user_prompt and st.session_state.user_prompt != "Select a query":
#st.write(st.session_state.user_prompt)
document_chain = create_stuff_documents_chain(llm=llm, prompt= prompt_template)
retriever = st.session_state.vectors.as_retriever()
retrieval_chain=create_retrieval_chain(retriever,document_chain)
start = time.process_time()
response = retrieval_chain.invoke({"input": st.session_state.user_prompt})
print(f"Response time :{time.process_time()-start}")
st.write(response['answer'])
## With a streamlit expander
with st.expander("Document similarity Search"):
for i,doc in enumerate(response['context']):
st.write(doc.page_content)
st.write('------------------------')
|