OCR/openspec/changes/archive/2025-12-11-use-cellboxes-for-table-rendering/tasks.md

1. Core Algorithm Implementation

1.1 Grid Inference Module

• 1.1.1 Create CellBoxGridInferrer class in pdf_table_renderer.py
• 1.1.2 Implement cluster_values() for Y/X coordinate clustering
• 1.1.3 Implement infer_grid_from_cellboxes() main method
• 1.1.4 Add row_heights and col_widths calculation

1.2 Content Mapping

• 1.2.1 Implement extract_cell_contents() from HTML
• 1.2.2 Implement map_content_to_grid() for row-by-row assignment
• 1.2.3 Handle content count mismatch (more/fewer cells)

2. PDF Generator Integration

2.1 New Rendering Path

• 2.1.1 Add render_from_cellboxes_grid() method to TableRenderer
• 2.1.2 Integrate into draw_table_region() with cellboxes-first check
• 2.1.3 Maintain fallback to existing HTML-based rendering

2.2 Cell Rendering

• 2.2.1 Draw cell borders using cell_boxes coordinates
• 2.2.2 Render text content with proper alignment and padding
• 2.2.3 Handle multi-line text within cells

3. Configuration

3.1 Settings

• 3.1.1 Add table_rendering_prefer_cellboxes: bool = True
• 3.1.2 Add table_cellboxes_row_threshold: float = 15.0
• 3.1.3 Add table_cellboxes_col_threshold: float = 15.0

4. Testing

4.1 Unit Tests

• 4.1.1 Test grid inference with various cell_box configurations
• 4.1.2 Test content mapping edge cases
• 4.1.3 Test coordinate clustering accuracy

4.2 Integration Tests

• 4.2.1 Test with scan.pdf Table 7 (the problematic case)
• 4.2.2 Verify no regression on existing table tests
• 4.2.3 Visual comparison of output PDFs

5. Documentation

• 5.1 Update inline code comments
• 5.2 Update spec with new table rendering requirement