Skip to main content

Best practices

Start lossless, measure first​

  • Run filerepack doctor after install so missing binaries are obvious.
  • Use --dryrun (and --stats) on a sample before rewriting a tree.
  • Keep JPEG/PNG/PDF lossless unless you explicitly want --lossy or a quality flag.
  • Do not pass --allow-grow unless you are debugging a packer.

Nested files and Office​

  • Leave --deep on so images and XML inside .docx / .xlsx / ZIP are packed.
  • Install Info-ZIP zip for OOXML so Word/Excel stay happy.
  • Use --no-images when you only want XML/JSON minify and archive rewrite.

Bulk directories​

  • Prefer --jobs auto on multi-core machines.
  • Filter with --include-ext / --exclude-ext and --min-size so you do not touch tiny files.
  • Default skipped directories already cover .git, virtualenvs, and node_modules. Add more with --exclude-dir.
  • bulk skips top-level .zip by default (--no-skip-zip to include them). Nested zips inside other archives are still walked.
  • Pass --progress (and filerepack[progress]) for long jobs.

PDFs and scans​

  • Lossless: pip install 'filerepack[pdf]' then filerepack repack file.pdf.
  • Scanned pages: --lossy (Ghostscript /ebook). Use --pdf-profile prepress for print-quality Ghostscript.
  • --jpeg-quality also re-encodes images inside PDFs.

Data and media extras​

  • Parquet: filerepack[parquet] or [data]; --ultra is zstd level 22.
  • Cover art: filerepack[media].
  • Fonts: filerepack[fonts].

Configuration​

Put tool paths in ~/.config/filerepack/config.toml or environment variables (FILEREPACK_7ZZ, …). See External tools.