feat: add pdf-to-markdown tool

[fdematos]

Summary

• add pdf-to-markdown tool implementation and core converter
• register new tool in plugin exports and tool map
• add pdf-to-markdown tests and PDF parsing dependency

[opencode-agent]

Great! The tests are passing. Now let me analyze the code quality and potential issues:

Code Review Analysis

Overall Assessment

The PDF-to-Markdown tool implementation is well-structured and functional. The code follows good TypeScript practices, has comprehensive tests, and integrates properly with the existing plugin architecture.

Strengths

1. Good Type Safety: Proper TypeScript interfaces and type guards are used throughout
2. Comprehensive Testing: Tests cover core functionality including headings, paragraphs, and lists
3. Error Handling: Proper error handling for file reading and PDF parsing
4. Modular Design: Core conversion logic is separated from the tool definition
5. Performance Considerations: Efficient text grouping and line merging algorithms

Issues and Improvements

1. Missing Type Definitions for PDF.js

The TypeScript compilation fails due to missing bun-types. This needs to be addressed in the tsconfig.json.

2. PDF.js Configuration Issues

The tests show warnings about missing standardFontDataUrl parameter. This should be configured for better PDF parsing reliability.

3. Edge Case Handling

• Empty PDFs: The code doesn't explicitly handle empty PDFs with no text content
• Very Large PDFs: No protection against extremely large PDFs that could cause memory issues
• Corrupted PDFs: More robust error handling for corrupted PDF files

4. Code Quality Improvements

src/tools/pdf-to-markdown-core.ts:216-231
The page processing loop could benefit from better error handling around individual page parsing failures.

src/tools/pdf-to-markdown-core.ts:206-207
The getDocument call could use better error handling for invalid PDF data.

5. Performance Considerations

The current implementation loads the entire PDF into memory. For very large PDFs, this could be problematic. Consider:

• Streaming processing for large documents
• Memory limits and chunking

6. Documentation

The tool description could be more specific about:

• What types of PDFs work best (text-based vs scanned)
• Limitations (complex layouts, tables, images)
• Performance characteristics

7. Testing Improvements

Additional test cases would be beneficial:

• Multi-page documents
• Documents with tables
• Documents with mixed content types
• Edge cases like very small/large font sizes
• Non-English text handling

Specific Code Issues

1. src/tools/pdf-to-markdown-core.ts:206

   const loadingTask = getDocument({ data } as DocumentInitParameters)

   The type assertion could be more specific about what properties are expected.

2. src/tools/pdf-to-markdown-core.ts:218-220
   The getTextContent call uses a type assertion that could be more precise.

3. Error Messages: Some error messages could be more user-friendly and actionable.

Recommendations

1. Fix TypeScript Configuration: Remove or properly configure bun-types in tsconfig.json
2. Add PDF.js Configuration: Set up proper standardFontDataUrl for better PDF parsing
3. Enhance Error Handling: Add more specific error handling for various PDF parsing scenarios
4. Add Memory Limits: Consider adding configuration options for maximum PDF size
5. Expand Test Coverage: Add tests for edge cases and complex document structures
6. Improve Documentation: Add more detailed documentation about capabilities and limitations

The implementation is solid overall and the tests pass, but these improvements would make it more robust and production-ready.

opencode session  |  github run

[opencode-agent]

Great! The tests pass and TypeScript compilation succeeds. Now let me analyze the code quality and potential issues:

Code Review Analysis

Overall Assessment

The PDF-to-Markdown tool implementation is well-structured and functional. The code follows good TypeScript practices, has comprehensive tests, and integrates properly with the existing plugin architecture.

Strengths

1. Good Type Safety: Proper TypeScript interfaces and type guards are used throughout
2. Comprehensive Testing: Tests cover core functionality including headings, paragraphs, and lists
3. Error Handling: Proper error handling for file reading and PDF parsing
4. Modular Design: Core conversion logic is separated from the tool definition
5. Performance Considerations: Efficient text grouping and line merging algorithms

Issues and Improvements

1. PDF.js Configuration Issues

The tests show warnings about missing standardFontDataUrl parameter. This should be configured for better PDF parsing reliability.

2. Edge Case Handling

• Empty PDFs: The code doesn't explicitly handle empty PDFs with no text content
• Very Large PDFs: No protection against extremely large PDFs that could cause memory issues
• Corrupted PDFs: More robust error handling for corrupted PDF files

3. Code Quality Improvements

src/tools/pdf-to-markdown-core.ts:216-231
The page processing loop could benefit from better error handling around individual page parsing failures.

src/tools/pdf-to-markdown-core.ts:206-207
The getDocument call could use better error handling for invalid PDF data.

4. Performance Considerations

The current implementation loads the entire PDF into memory. For very large PDFs, this could be problematic. Consider:

• Streaming processing for large documents
• Memory limits and chunking

5. Documentation

The tool description could be more specific about:

• What types of PDFs work best (text-based vs scanned)
• Limitations (complex layouts, tables, images)
• Performance characteristics

6. Testing Improvements

Additional test cases would be beneficial:

• Multi-page documents
• Documents with tables
• Documents with mixed content types
• Edge cases like very small/large font sizes
• Non-English text handling

Specific Code Issues

1. src/tools/pdf-to-markdown-core.ts:206

   const loadingTask = getDocument({ data } as DocumentInitParameters)

   The type assertion could be more specific about what properties are expected.

2. src/tools/pdf-to-markdown-core.ts:218-220
   The getTextContent call uses a type assertion that could be more precise.

3. Error Messages: Some error messages could be more user-friendly and actionable.

Recommendations

1. Add PDF.js Configuration: Set up proper standardFontDataUrl for better PDF parsing
2. Enhance Error Handling: Add more specific error handling for various PDF parsing scenarios
3. Add Memory Limits: Consider adding configuration options for maximum PDF size
4. Expand Test Coverage: Add tests for edge cases and complex document structures
5. Improve Documentation: Add more detailed documentation about capabilities and limitations

The implementation is solid overall and the tests pass, but these improvements would make it more robust and production-ready.

opencode session  |  github run