Upload 2 files
Browse files- requirements.txt +15 -0
- test.py +91 -0
requirements.txt
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit
|
2 |
+
google-generativeai
|
3 |
+
python-dotenv
|
4 |
+
langchain
|
5 |
+
PyPDF2
|
6 |
+
chromadb
|
7 |
+
faiss-cpu
|
8 |
+
pdf2image
|
9 |
+
langchain
|
10 |
+
PyPDF2
|
11 |
+
chromadb
|
12 |
+
faiss-cpu
|
13 |
+
langchain_google_genai
|
14 |
+
pdfplumber
|
15 |
+
pickle
|
test.py
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from dotenv import load_dotenv
|
2 |
+
load_dotenv() # Load all env variables
|
3 |
+
|
4 |
+
import streamlit as st
|
5 |
+
from PyPDF2 import PdfReader
|
6 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
7 |
+
import os
|
8 |
+
from langchain_google_genai import GoogleGenerativeAIEmbeddings
|
9 |
+
import google.generativeai as genai
|
10 |
+
from langchain.vectorstores import FAISS
|
11 |
+
from langchain_google_genai import ChatGoogleGenerativeAI
|
12 |
+
from langchain.chains.question_answering import load_qa_chain
|
13 |
+
from langchain.prompts import PromptTemplate
|
14 |
+
|
15 |
+
## Function to load gemini pro model and get responses
|
16 |
+
model = genai.GenerativeModel("gemini-pro")
|
17 |
+
def get_gemini_response(question):
|
18 |
+
response=model.generate_content(question)
|
19 |
+
return response.text
|
20 |
+
|
21 |
+
def get_pdf_text(pdf_docs):
|
22 |
+
text=""
|
23 |
+
for pdf in pdf_docs:
|
24 |
+
pdf_reader= PdfReader(pdf)
|
25 |
+
for page in pdf_reader.pages:
|
26 |
+
text+= page.extract_text()
|
27 |
+
return text
|
28 |
+
|
29 |
+
def get_text_chunks(text):
|
30 |
+
text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
|
31 |
+
chunks = text_splitter.split_text(text)
|
32 |
+
return chunks
|
33 |
+
|
34 |
+
def get_vector_store(text_chunks):
|
35 |
+
embeddings = GoogleGenerativeAIEmbeddings(model = "models/embedding-001")
|
36 |
+
vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
|
37 |
+
vector_store.save_local("faiss_index")
|
38 |
+
|
39 |
+
def get_conversational_chain():
|
40 |
+
prompt_template = """
|
41 |
+
Answer the question as detailed as possible from the provided context, make sure to provide all the details, if the answer is not in
|
42 |
+
provided context just say, "answer is not available in the context", don't provide the wrong answer\n\n
|
43 |
+
Context:\n {context}?\n
|
44 |
+
Question: \n{question}\n
|
45 |
+
|
46 |
+
Answer:
|
47 |
+
"""
|
48 |
+
model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
|
49 |
+
prompt = PromptTemplate(template = prompt_template, input_variables = ["context", "question"])
|
50 |
+
chain = load_qa_chain(model, chain_type="stuff", prompt=prompt)
|
51 |
+
return chain
|
52 |
+
|
53 |
+
def user_input(user_question):
|
54 |
+
embeddings = GoogleGenerativeAIEmbeddings(model = "models/embedding-001")
|
55 |
+
new_db = FAISS.load_local("faiss_index", embeddings,allow_dangerous_deserialization=True)
|
56 |
+
docs = new_db.similarity_search(user_question)
|
57 |
+
chain = get_conversational_chain()
|
58 |
+
response = chain({"input_documents":docs, "question": user_question}, return_only_outputs=True)
|
59 |
+
return response["output_text"]
|
60 |
+
|
61 |
+
def main():
|
62 |
+
st.set_page_config(page_title='Q&A Demo')
|
63 |
+
st.header("Combined Application")
|
64 |
+
|
65 |
+
app_mode = st.sidebar.selectbox("Choose the App Mode", ["Gemini Q&A", "PDF Q&A"])
|
66 |
+
|
67 |
+
if app_mode == "Gemini Q&A":
|
68 |
+
st.subheader("Gemini LLM Application")
|
69 |
+
user_question = st.text_input("Input")
|
70 |
+
if st.button("Ask the question"):
|
71 |
+
response = get_gemini_response(user_question)
|
72 |
+
st.subheader("The Response is :")
|
73 |
+
st.write(response)
|
74 |
+
|
75 |
+
elif app_mode == "PDF Q&A":
|
76 |
+
st.subheader("Chat with PDF using Gemini💁")
|
77 |
+
user_question = st.text_input("Ask a Question from the PDF Files")
|
78 |
+
pdf_docs = st.file_uploader("Upload your PDF Files and Click on the Submit & Process Button", accept_multiple_files=True)
|
79 |
+
if st.button("Submit & Process"):
|
80 |
+
with st.spinner("Processing..."):
|
81 |
+
raw_text = get_pdf_text(pdf_docs)
|
82 |
+
text_chunks = get_text_chunks(raw_text)
|
83 |
+
get_vector_store(text_chunks)
|
84 |
+
st.success("Done")
|
85 |
+
|
86 |
+
if user_question:
|
87 |
+
response = user_input(user_question)
|
88 |
+
st.write("Reply: ", response)
|
89 |
+
|
90 |
+
if __name__ == "__main__":
|
91 |
+
main()
|