Understand

Understand is a Google Chrome extension that converts the images inside PDFs to text. This is incredibly useful for searching through scanned books.

Chrome extension for PDF optical character recognition

TODO

Features

• Stitch all images into 1 to reduce # of API requests
• PDF download progress
• Re-enable pages > 1
• Fix "All images processed indicator"
• Order "recognize-ed" text properly (vertically)
• Improve highlight positioning
• Understand local files (file://....) - ask user to select file using <input type="file">, alternative: native app (these 2 are necessary if they don't have "allow file URLs" checked for the extension)
• Offer to load normally (especially if there was an error)
• Loading progress bars (instead of just spinner)
• Port to other browsers

Developer Tooling

• SCSS pre-processing
• Use a UI library (React?)
• TypeScript on the server
• Unit tests
• Integration tests
• Continuous integration
• Docker

Other

• Code cleanup (split app.ts into smaller files)

Design Decisions

See this document for a list of critical design decisions.

Contributing

See this document for contribution guidelines, development environment setup, and build processes.