GSoC 2026 Proposal: Bulk OCR Improvements
Contact Information
Name: Kaung Sithu Tun
Phabricator Username: Kaung_Sithu_Tun
Gerrit Username: KaungSithuTun
Location/Timezone: Bangkok/Asia
Typical Working Hours: 11:00 - 23:00
Project Overview
The Bulk OCR feature in Wikisource is essential for digitizing large volumes of text. My goal is to improve the user interface and reliability of this process, ensuring that authorized users can efficiently process multiple pages while receiving clear, actionable feedback on the status of each page.
Goals
- Implement a robust UI for tracking bulk OCR progress.
- Improve error handling so users can clearly see which pages failed and why (fixing issues like T411157).
- Refine the permission logic to ensure only authorized user groups can trigger bulk actions.
- Optimize the frontend performance when handling large lists of pages.
Technical Skills & Experience
- Frontend Development: Experienced in JavaScript, HTML, and CSS.
- Modern Frameworks: Strong background in Next.js and TypeScript.
- State Management: Experience managing complex UI states and API integrations (Supabase).
- Open Source: Familiar with Git and collaborative development workflows.
Proposed Timeline (350 Hour Project)
Phase 1: Community Bonding (May 4 - May 24)
- Set up the local MediaWiki development environment.
- Engage with mentors (@theprotonade and @SGill) to refine the technical design.
- Explore the existing Wikisource extension codebase and the Wikimedia OCR API.
Phase 2: Implementation - First Half (May 25 - July 6)
- Week 1-2: Refactor the current Bulk OCR trigger logic to handle large page arrays efficiently.
- Week 3-4: Develop the new Progress UI (ProgressBar/Status Indicators).
- Week 5-6: Implement the "Error Summary" view to replace the current buggy notification system.
Phase 3: Implementation - Second Half (July 14 - August 17)
- Week 7-8: Integrate permission-based UI elements (hiding/showing features for authorized users).
- Week 9-10: Testing and Bug Fixing. Ensure compatibility across different Wikisource skins.
- Week 11-12: Documentation and Final Polish. Prepare the final code for production deployment.
Why Me?
I am a dedicated frontend developer with a focus on building functional, user-centric tools. I have experience building educational platforms and managing complex data flows. Contributing to Wikimedia allows me to apply my skills to a global project that provides free knowledge to everyone.