How happyshadowcoder/image-to-table works
TableLens is a Flask-based table-extraction application. A user works in the browser with an image, chooses local or cloud vision, and obtains structured table data for Excel or CSV export. The README also describes nested-header handling, one-shot extraction, and model settings. The main Python source was not excerpted, so internal runtime wiring is represented only where the README documents the workflow; implementation details remain uncertain.
Main entry points: app.py Image extraction · index.html Extraction interface · dashboard.html Model settings · docs.html Help pages
Components 16
Web interface
- Flask application
app.py - Extraction interface
templates/index.html - Model settings
templates/dashboard.html - Help pages
templates/docs.html
Table processing
- Image extraction
app.py - One-shot mode
app.py - Table structuring
app.py
Vision providers
- Local Ollama
app.py - Cloud APIs
app.py
Export
- Excel export
app.py - CSV export
app.py
Other components
- User
- Browser
- Ollama service
- OpenAI API
- Gemini API
Connections 15
- User → Browser: uses
README.md - Browser → Flask application: opens
README.md - Flask application → Extraction interface: presents
- Flask application → Model settings: presents
- Flask application → Help pages: presents
- Browser → Image extraction: submits image
README.md - Image extraction → Local Ollama: uses local model
README.md - Image extraction → Cloud APIs: uses cloud model
README.md - Local Ollama → Ollama service: calls
README.md - Cloud APIs → OpenAI API: calls
README.md - Cloud APIs → Gemini API: calls
README.md - Image extraction → One-shot mode: supports
README.md - Image extraction → Table structuring: structures tables
README.md - Table structuring → Excel export: exports
README.md - Table structuring → CSV export: exports
README.md