glyph vacuum

glyph-vacuum

Measure every image your forms stream before you write a glyph rule. It walks the .dfm / .fmx files under one or more root folders, decodes each Picture.Data / Glyph.Data blob (wrapper class, format, width, height, bits per pixel, palette) and pairs it with its count property (NumGlyphs and kin). Use it when you convert a component whose pictures or glyph strips must survive the conversion and you need to know what is out there first.

Running it from the CLI

drag-lint glyph-vacuum --root DIR [--root DIR ...] --output DIR [--append] [--db PATH ...]

--root is a folder to scan and may be repeated. --output is the folder that receives the result files; --append adds this run to a measurement already there instead of replacing it. --db may be repeated.

What you get

In the output folder: instances.tsv (one row per streamed graphic), classes.tsv (one row per wrapper class), skipped.tsv (blobs it could not decode, with the reason) and an images\ folder with the decoded pictures.

Reaching it in the IDE

No IDE surface -- this is a CLI-only feature. The glyph expressions of a rule book (G[I/N]) are authored in the ConvRulesEditor; see ConvRulesEditor.

What it needs

The folders to scan. An index is optional (--db).

Where it fits

Run glyph-vacuum first, read classes.tsv to see which wrapper classes and sizes occur, then write the glyph rule in your book and check it with convert-validate.