Crumbgrabber avatar

pdf-processing-pro

Production-ready PDF processing with forms, tables, OCR, validation, and batch operations. Use when

作者 Crumbgrabber|オープンソース

PDF Processing Pro

Production-ready PDF processing guidance with comprehensive error handling and support for complex workflows (forms, tables, OCR, batch operations).

Core patterns (tool-agnostic)

  • Text extraction: Use robust libraries (e.g., pdfplumber) and validate output per page.
  • Form workflows: Detect fields, validate data against schemas, then fill; revalidate outputs.
  • Table extraction: Combine multiple extractors and normalize columns; handle merged cells explicitly.
  • OCR for scanned PDFs: Preprocess pages (deskew/denoise) before OCR; store page-level confidence and flag low-confidence regions.
  • Batch operations: Process PDFs page-by-page to control memory; log per-file successes/failures with timestamps and error details.

Reliability practices

  • Wrap PDF operations in try/except; include filename, operation, and stack trace in logs.
  • Validate inputs (paths, expected fields, schemas) before processing.
  • Use structured logging with timestamps, operation, result, duration.
  • Never log sensitive PDF contents; sanitize before sending to external services.
  • For large jobs, checkpoint outputs (per-file artifacts) so you can resume after failures.

Troubleshooting tips

  • File not found/invalid: Confirm path/permissions; verify file is not encrypted/corrupted.
  • OCR quality issues: Increase DPI during rasterization, deskew, and denoise before OCR.
  • Table extraction errors: Try alternate parsers/settings; manually define column boundaries when automated detection fails.
  • Memory issues: Stream pages instead of loading whole documents; free resources after each file.
pdf-processing-pro - Claude Code・Cursor 対応の AIエージェント Skill | Agent Skills