Spaces:
Sleeping
Sleeping
File size: 3,689 Bytes
a68f74a |
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 |
import os
import requests
from groq import Groq
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS
from langchain.text_splitter import RecursiveCharacterTextSplitter
from PyPDF2 import PdfReader
import streamlit as st
from tempfile import NamedTemporaryFile
# Initialize Groq client
client = Groq(api_key=os.environ['GROQ_API_KEY'])
# Function to extract text from a PDF
def extract_text_from_pdf(pdf_file_path):
pdf_reader = PdfReader(pdf_file_path)
text = ""
for page in pdf_reader.pages:
page_text = page.extract_text()
if page_text:
text += page_text
return text
# Function to split text into chunks
def chunk_text(text, chunk_size=500, chunk_overlap=50):
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=chunk_size, chunk_overlap=chunk_overlap
)
return text_splitter.split_text(text)
# Function to create embeddings and store them in FAISS
def create_embeddings_and_store(chunks, vector_db=None):
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
if vector_db is None:
vector_db = FAISS.from_texts(chunks, embedding=embeddings)
else:
vector_db.add_texts(chunks)
return vector_db
# Function to query the vector database and interact with Groq
def query_vector_db(query, vector_db):
docs = vector_db.similarity_search(query, k=3)
context = "\n".join([doc.page_content for doc in docs])
chat_completion = client.chat.completions.create(
messages=[
{"role": "system", "content": f"Use the following context:\n{context}"},
{"role": "user", "content": query},
],
model="llama3-8b-8192",
)
return chat_completion.choices[0].message.content
# Function to convert Google Drive view link to downloadable link
def get_direct_download_link(view_url):
if "drive.google.com/file/d/" in view_url:
file_id = view_url.split("/file/d/")[1].split("/")[0]
return f"https://drive.google.com/uc?export=download&id={file_id}"
return None
# Function to download and save a PDF from a URL
def download_pdf_from_url(url):
direct_url = get_direct_download_link(url)
if not direct_url:
return None
response = requests.get(direct_url)
if response.status_code == 200:
temp_file = NamedTemporaryFile(delete=False, suffix=".pdf")
temp_file.write(response.content)
temp_file.close()
return temp_file.name
else:
return None
# Streamlit app
st.title("RAG-Based QA on Google Drive PDFs")
# Only fetch from provided links
doc_links = [
"https://drive.google.com/file/d/1YWX-RYxgtcKO1QETnz1N3rboZUhRZwcH/view?usp=sharing",
"https://drive.google.com/file/d/1JPf0XvDhn8QoDOlZDrxCOpu4WzKFESNz/view?usp=sharing",
]
vector_db = None
# Process Google Drive documents
for idx, link in enumerate(doc_links):
st.write(f"📄 Fetching and processing PDF from Link {idx + 1}...")
pdf_path = download_pdf_from_url(link)
if pdf_path:
text = extract_text_from_pdf(pdf_path)
chunks = chunk_text(text)
vector_db = create_embeddings_and_store(chunks, vector_db=vector_db)
st.success(f"✅ Processed document {idx + 1}")
else:
st.error(f"❌ Failed to download or process PDF from Link {idx + 1}")
# User query input
user_query = st.text_input("🔍 Enter your query:")
if user_query and vector_db:
response = query_vector_db(user_query, vector_db)
st.subheader("💬 Response from LLM:")
st.write(response)
elif user_query:
st.warning("⚠️ No documents processed to query.")
|