How semani01/abstractive-summarization works
This is a Flask web app for abstractive summarization with PEGASUS. Users submit text, PDFs, or URLs through the browser; Flask extracts or accepts text, runs model inference, and returns a summary for display. The graph separates the input paths from shared inference and the web interface. Training and evaluation are described in the README, but the available source evidence does not establish them as runtime subsystems.
Main entry points: app.py Flask routes · index.html Browser UI · scripts.js Request handlers
Components 10
Web interface
- Browser UI
app/templates/index.html - Request handlers
app/static/scripts.js - Flask routes
app/app.py
Input processing
- Text input
app/app.py - PDF extraction
app/app.py - Web extraction
app/app.py
Summarization
- Text summarizer
app/app.py - PEGASUS inference
app/app.py
Other components
- User
- Remote web content
Connections 14
- User → Browser UI: provides input
README.md - Browser UI → Request handlers: triggers
app/static/scripts.js - Request handlers → Flask routes: posts requests
app/static/scripts.js - Flask routes → Browser UI: renders page
app/app.py - Flask routes → Text input: accepts text
app/app.py - Flask routes → PDF extraction: extracts pages
app/app.py - Flask routes → Web extraction: fetches and extracts
app/app.py - Web extraction → Remote web content: requests page
app/app.py - Flask routes → Text summarizer: summarizes text
app/app.py - Text input → Text summarizer: passes text
app/app.py - PDF extraction → Text summarizer: passes text
app/app.py - Web extraction → Text summarizer: passes text
app/app.py - Text summarizer → PEGASUS inference: generates summary
app/app.py - Request handlers → Browser UI: displays summary
app/static/scripts.js