Spaces:
Running
Running
File size: 5,044 Bytes
9a05fa5 ba8b895 9a05fa5 393c185 9a05fa5 393c185 9a05fa5 393c185 9a05fa5 ba8b895 9a05fa5 393c185 9a05fa5 5a1dbb0 9a05fa5 5a1dbb0 9a05fa5 5a1dbb0 9a05fa5 5a1dbb0 9a05fa5 |
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 |
import gradio as gr
from langchain_community.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain_chroma import Chroma
from langchain_community.vectorstores import FAISS
from langchain_groq import ChatGroq
from langchain.chains import create_retrieval_chain
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain_core.prompts import ChatPromptTemplate
import os
from dotenv import load_dotenv
from helper import SYSTEM_PROMPT
from langchain_google_genai import GoogleGenerativeAIEmbeddings
# from langchain.embeddings import HuggingFaceEmbeddings # open source free embedding
load_dotenv()
class PDFQAProcessor:
SYSTEM_PROMPT = SYSTEM_PROMPT
llm = ChatGroq(
# model_name="deepseek-r1-distill-llama-70b",
model_name="llama-3.3-70b-versatile",
temperature=0.1,
max_tokens=8000,
api_key = os.getenv('GROQ_API_KEY')
)
# Setup RAG chain
prompt = ChatPromptTemplate.from_messages([
("system", SYSTEM_PROMPT),
("human", "{input}"),
])
question_answer_chain = create_stuff_documents_chain(llm, prompt)
# EMBEDDING_MODEL = "intfloat/e5-large-v2"
# embeddings = HuggingFaceEmbeddings(
# model_name=EMBEDDING_MODEL,
# model_kwargs={'device': 'cpu'},
# encode_kwargs={'normalize_embeddings': True}
# )
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
CHUNK_SIZE = 550
CHUNK_OVERLAP = 80
text_splitter = RecursiveCharacterTextSplitter(chunk_size=CHUNK_SIZE,chunk_overlap = CHUNK_OVERLAP)
# persist_directory="./chroma_db"
def __init__(self):
self.vectorstore = None
self.retriever = None
def process_pdfs(self, pdf_files):
"""Processing PDF files and creating vector store"""
if not pdf_files:
return "Please upload PDF files first!"
try:
# Load and split documents
docs = []
for pdf_file in pdf_files:
loader = PyPDFLoader(pdf_file.name)
docs.extend(loader.load())
splits = self.text_splitter.split_documents(docs)
# # Create vector store
# self.vectorstore = Chroma.from_documents(
# documents=splits,
# embedding=self.embeddings,
# # persist_directory = self.persist_directory
# )
# Replace Chroma with:
self.vectorstore = FAISS.from_documents(
splits,
self.embeddings
)
self.retriever = self.vectorstore.as_retriever(search_kwargs={"k": 18})
return "PDFs processed successfully! Ask your questions now."
except Exception as e:
return f"Error processing PDFs: {str(e)}"
def answer_question(self, question):
"""Handling question answering"""
if not self.retriever:
return "Please process PDFs first!", None
try:
# Initialize LLM
rag_chain = create_retrieval_chain(self.retriever, self.question_answer_chain)
response = rag_chain.invoke({"input": question})
# final_response = response["answer"] + "\n\nSources\n\n"
# for info in response["context"]:
# final_response += info.page_content + "\nSource of Info: " + info.metadata['source'] + "\nAt Page No: " + info.metadata['page_label']+"\n\n"
final_response = response["answer"] + "\n\n### Sources\n\n" # Changed to use markdown formatting
for info in response["context"]:
final_response += (
f"{info.page_content}<br>" # Changed to use markdown bold formatting
f"Source of Info: {info.metadata['source']}<br>"
f"At Page No: {info.metadata['page_label']}<br><br>"
)
return final_response
except Exception as e:
return f"Error answering question: {str(e)}", None
processor = PDFQAProcessor()
with gr.Blocks(title="PDF QA Assistant") as demo:
with gr.Tab("Upload PDFs"):
file_input = gr.Files(label="Upload PDFs", file_types=[".pdf"])
process_btn = gr.Button("Process PDFs")
status_output = gr.Textbox(label="Processing Status")
with gr.Tab("Ask Questions"):
question_input = gr.Textbox(label="Your Question")
# answer_output = gr.Textbox(label="Answer", interactive=False)
answer_output = gr.Markdown(label="Answer")
ask_btn = gr.Button("Ask Question")
process_btn.click(
processor.process_pdfs,
inputs=file_input,
outputs=status_output
)
# QA workflow
ask_btn.click(
processor.answer_question,
inputs=question_input,
outputs=[answer_output]
)
if __name__ == "__main__":
demo.launch() |