IP Official Documents Auto-Archiving

Problem Context

IP teams often receive batches of patent, trademark, and software-copyright official PDFs with inconsistent filenames, scattered directories, duplicate files, and mixed legal statuses. Manually renaming documents, verifying application or registration numbers, determining legal state, and maintaining a ledger is easy to get wrong and hard to audit.

How It Works

The skill processes a directory of PDFs in batch: it recursively scans the --input folder, identifies patent, trademark, and software-copyright documents, and extracts fields such as application number, notice name, and rights holder. It then renames files according to naming rules, organizes them into a structured directory, and generates a summary report; each IP item keeps only the latest status record.

Core steps include:
- PDF parsing: prefers pdfplumber, with PyMuPDF or PyPDF2 as fallbacks
- Legal status inference: applies priority rules from certificates, notices, and registers, such as granted, substantive examination, or rejection
- CSV management: detects source encodings and writes UTF-8-BOM for Excel compatibility
- Idempotent archiving: deduplicates by combined keys such as application number, registration number, or record number

Scope and Notes

It fits teams that already have batches of official PDFs and need consistent directory structures plus latest-status summaries. It requires Python 3.10+ and at least one PDF parsing library; scanned PDFs without a text layer, ambiguous fields, or unusual naming conventions still need manual review.