Anupam251272's picture
Create app.py
d3a91f4 verified
import PyPDF2
import torch
from transformers import pipeline
import gradio as gr
import logging
from typing import List
import time
import requests
from bs4 import BeautifulSoup
import io
import tempfile
import os
from tqdm import tqdm
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class ContentQuestionGenerator:
def __init__(self):
self.device = "cuda" if torch.cuda.is_available() else "cpu"
logger.info(f"Using device: {self.device}")
self.summarizer = pipeline(
"summarization",
model="facebook/bart-large-cnn",
device=0 if self.device == "cuda" else -1
)
self.question_generator = pipeline(
"text2text-generation",
model="lmqg/t5-base-squad-qg",
device=0 if self.device == "cuda" else -1
)
def process_large_pdf(self, file_obj, chunk_size=50) -> str:
"""Process large PDF files in chunks."""
try:
# Create a temporary file to store the PDF
with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as temp_file:
if isinstance(file_obj, bytes):
temp_file.write(file_obj)
else:
temp_file.write(file_obj.read())
temp_file_path = temp_file.name
# Open the PDF with PyPDF2
with open(temp_file_path, 'rb') as file:
pdf_reader = PyPDF2.PdfReader(file)
total_pages = len(pdf_reader.pages)
logger.info(f"Processing PDF with {total_pages} pages")
all_text = []
# Process pages in chunks
for i in range(0, total_pages, chunk_size):
chunk_text = ""
end_page = min(i + chunk_size, total_pages)
logger.info(f"Processing pages {i+1} to {end_page}")
for page_num in range(i, end_page):
try:
page = pdf_reader.pages[page_num]
chunk_text += page.extract_text() + "\n"
except Exception as e:
logger.warning(f"Error extracting text from page {page_num + 1}: {str(e)}")
continue
if chunk_text.strip():
all_text.append(chunk_text)
# Free up memory
del chunk_text
# Clean up temporary file
os.unlink(temp_file_path)
return "\n".join(all_text)
except Exception as e:
logger.error(f"Error processing large PDF: {str(e)}")
if 'temp_file_path' in locals():
try:
os.unlink(temp_file_path)
except:
pass
raise
def extract_text_from_url(self, url: str) -> str:
"""Extract text content from a webpage."""
try:
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
}
response = requests.get(url, headers=headers, timeout=30)
response.raise_for_status()
soup = BeautifulSoup(response.text, 'html.parser')
# Remove unwanted elements
for element in soup(['script', 'style', 'nav', 'header', 'footer', 'aside']):
element.decompose()
# Handle Wikipedia specifically
if 'wikipedia.org' in url:
main_content = soup.find('div', {'id': 'mw-content-text'})
text = ' '.join([p.get_text() for p in (main_content or soup).find_all('p')])
else:
text = ' '.join([p.get_text() for p in soup.find_all('p')])
text = ' '.join(text.split())
if not text.strip():
raise ValueError("No text content could be extracted from the URL")
return text.strip()
except Exception as e:
logger.error(f"Error extracting text from URL: {str(e)}")
raise ValueError(f"Could not extract text from URL: {str(e)}")
def chunk_text(self, text: str, max_chunk_size: int = 1024) -> List[str]:
"""Split text into chunks for processing."""
chunks = []
current_chunk = []
current_size = 0
for sentence in text.split('.'):
sentence = sentence.strip() + '.'
if current_size + len(sentence) + 1 <= max_chunk_size:
current_chunk.append(sentence)
current_size += len(sentence) + 1
else:
if current_chunk:
chunks.append(' '.join(current_chunk))
current_chunk = [sentence]
current_size = len(sentence) + 1
if current_chunk:
chunks.append(' '.join(current_chunk))
return chunks
def summarize_text(self, text: str) -> str:
"""Summarize text with memory-efficient chunking."""
chunks = self.chunk_text(text)
summaries = []
for chunk in tqdm(chunks, desc="Summarizing text"):
if len(chunk.strip()) > 50:
try:
summary = self.summarizer(chunk,
max_length=150,
min_length=40,
do_sample=False)[0]['summary_text']
summaries.append(summary)
except Exception as e:
logger.warning(f"Error summarizing chunk: {str(e)}")
continue
# Free up memory
torch.cuda.empty_cache() if torch.cuda.is_available() else None
return " ".join(summaries)
def generate_questions(self, text: str, num_questions: int = 20) -> List[str]:
"""Generate diverse questions with memory management."""
try:
all_questions = set() # Use set to ensure uniqueness
sentences = text.split('.')
for sentence in tqdm(sentences, desc="Generating questions"):
if len(all_questions) >= num_questions * 2:
break
if len(sentence.strip()) > 30:
try:
generated = self.question_generator(
sentence.strip(),
max_length=64,
num_return_sequences=2,
do_sample=True,
temperature=0.8
)
for gen in generated:
question = gen['generated_text'].strip()
if question.endswith('?') and len(question.split()) > 3:
all_questions.add(question)
# Free up memory
torch.cuda.empty_cache() if torch.cuda.is_available() else None
except Exception as e:
logger.warning(f"Error generating question: {str(e)}")
continue
# Convert to list and randomize
questions_list = list(all_questions)
import random
random.shuffle(questions_list)
return questions_list[:num_questions]
except Exception as e:
logger.error(f"Error generating questions: {str(e)}")
raise
def process_input(self, input_data) -> str:
"""Process either PDF file or URL with progress tracking."""
try:
start_time = time.time()
# Extract text based on input type
if isinstance(input_data, str) and (input_data.startswith('http://') or input_data.startswith('https://')):
logger.info("Processing URL content...")
text = self.extract_text_from_url(input_data)
else:
logger.info("Processing PDF content...")
text = self.process_large_pdf(input_data)
logger.info(f"Extracted {len(text)} characters of text")
# Process in chunks with memory management
logger.info("Summarizing content...")
summarized_text = self.summarize_text(text)
logger.info(f"Summarized to {len(summarized_text)} characters")
logger.info("Generating questions...")
questions = self.generate_questions(summarized_text)
logger.info(f"Generated {len(questions)} questions")
if not questions:
return "Could not generate any valid questions from the content."
formatted_output = "\n".join(f"{i+1}. {q}" for i, q in enumerate(questions))
processing_time = time.time() - start_time
logger.info(f"Total processing time: {processing_time:.2f} seconds")
return formatted_output
except Exception as e:
error_msg = f"Error processing input: {str(e)}"
logger.error(error_msg)
return f"An error occurred: {error_msg}"
def create_gradio_interface():
"""Create and configure Gradio interface."""
generator = ContentQuestionGenerator()
def process_input(file, url):
if file is None and not url:
return "Please provide either a PDF file or a webpage URL."
if file is not None and url:
return "Please provide either a PDF file or a URL, not both."
try:
if url:
if not (url.startswith('http://') or url.startswith('https://')):
return "Please provide a valid URL starting with http:// or https://"
return generator.process_input(url)
return generator.process_input(file)
except Exception as e:
logger.error("Error processing input:", exc_info=True)
return f"Error processing input: {str(e)}"
interface = gr.Interface(
fn=process_input,
inputs=[
gr.File(
label="Upload PDF Document",
type="binary",
file_types=[".pdf"],
file_count="single"
),
gr.Textbox(
label="Or enter webpage URL",
placeholder="https://example.com/page or https://en.wikipedia.org/wiki/Topic"
)
],
outputs=gr.Textbox(
label="Generated Questions",
lines=20
),
title="Content Question Generator",
description="""
Upload any size PDF document or provide a webpage URL to generate relevant questions.
Features:
- Supports large PDF files (100MB+)
- Works with any webpage URL
- Special handling for Wikipedia pages
- Generates 20 unique random questions
- Shows progress during processing
Note: Large files may take several minutes to process.
""",
allow_flagging="never"
)
return interface
if __name__ == "__main__":
interface = create_gradio_interface()
interface.queue().launch(share=True)