WIP: Backend PDF Parsing

[harshkhandeparkar]

Fixes #146

Description

WIP: Created a (testing sandbox) binary that takes a PDF, renders its pages to images, parses each one using tesseract, and then detects QPs in them.

Planned features

• Detect multiple QPs in one PDF
• Remove OCR on the frontend, which takes too long, especially on mobile phones
  • The upload endpoint should create DB entries and dump all files in a directory
  • The PDF parser looks for changes in the uploaded papers directory and parses them one by one
  • If duplicate papers are found, add them unapproved and show on the Admin dashboard
  • If the detected paper is not duplicate, directly approve it (unless important details are missing, in which case also add to the admin dashboard)
• FUTURE: Report paper feature for incorrect detections

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
iqps	Ready	Preview, Comment	Apr 19, 2026 5:02am