fix(build): consolidate pdf parsing dependencies, remove extraneous html deps

[waleedlatif1]

Summary

consolidate pdf parsing dependencies, remove extraneous html deps, we had multiple pdf parse libraries but only really need one. we also explicitly declared a downstream dependency that we didn't need to declare, which caused conflicts and let to the failed gh action to publish the image

Type of Change

• Bug fix

Testing

Tested manually.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

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

Project	Deployment	Preview	Comments	Updated (UTC)
docs	Ready	Preview	Comment	Sep 1, 2025 5:19pm
sim	Building	Preview	Comment	Sep 1, 2025 5:19pm

[greptile-apps]

Greptile Summary

This PR consolidates PDF parsing dependencies and removes extraneous HTML dependencies to fix build conflicts that were preventing GitHub Actions from successfully publishing Docker images. The changes eliminate redundant PDF parsing libraries and streamline the file parsing architecture.

Key Changes Made:

1. PDF Parser Consolidation: Removed the complex dual PDF parsing approach that used both pdf-lib and a custom RawPdfParser class. The codebase now uses only the pdf-parse library through a simplified PdfParser implementation.

2. Dependency Cleanup: Updated package.json to remove pdf-lib and add @types/pdf-parse for TypeScript support. Also added entities for HTML entity handling and sharp for image processing, while pinning @types/html-to-text to an exact version.

3. Simplified Error Handling: Removed the complex fallback mechanism in the file parsing API route (apps/sim/app/api/files/parse/route.ts) that would attempt multiple PDF parsers. The system now relies on a single parsing approach with cleaner error handling.

4. Code Cleanup: Removed the entire 467-line RawPdfParser class and updated the file parser index to only load the main PdfParser. Also removed an unnecessary comment in the Outlook polling service.

5. Test Updates: Updated test mocks to reflect the removal of the RawPdfParser dependency.

The changes integrate well with the existing file processing system, maintaining the same API surface while reducing complexity. The file parser loader (apps/sim/lib/file-parsers/index.ts) continues to support multiple formats (PDF, CSV, DOCX, TXT, MD, PPTX, HTML) but now uses a simpler, more maintainable architecture for PDF processing.

Confidence score: 3/5

• This PR addresses a real build issue but removes fallback mechanisms that may have provided better error recovery for problematic PDF files
• Score reflects the trade-off between simplified maintenance and potentially reduced robustness in PDF parsing edge cases
• Pay close attention to apps/sim/lib/file-parsers/pdf-parser.ts and apps/sim/app/api/files/parse/route.ts for potential PDF parsing failures with complex documents

_{7 files reviewed, no comments}

_{Edit Code Review Bot Settings | Greptile}