Shields.io Badge JSON Endpoints

giellalt/giella-core

Shields.io Badge JSON Endpoints

This document describes the JSON files used as endpoints for shields.io badges displayed in language repository GitHub pages.

File Location

Badge JSON files are generated in the docs/badgedata/ directory of each language repository during the build process. These files are served from the gh-pages branch.

File Naming Convention

Files follow the pattern: prefix-datatype.json.

Prefixes

• fst: Lexicon and FST-related data
• pkg: Package-level data, not specific to the FST
• speller: Spell checker related data
• gramcheck: Grammar checker related data

Current Badge Files

FST-related

• fst-lemmacount.json: Number of lemmas in the lexicon

Package-related

• pkg-maturity.json: Maturity level of the language resources
• pkg-version.json: Package version number (from AC_INIT in configure.ac)

Speller-related

• speller-suggestions.json: Spell checker suggestion quality metrics
  • First position accuracy (%)
  • Top 5 accuracy (%)
  • Test set size
• speller-version.json: Speller version number (from SPELLERVERSION in configure.ac)

Hyphenation-related

• hyphenator-version.json: Hyphenator version number (from HYPHVERSION in configure.ac)

Grammar checker-related

• gramcheck-version.json: Grammar checker version number (from GRAMCHECKVERSION in configure.ac)
• gramcheck-rules.json: Number of error-detection rules in tools/grammarcheckers/grammarchecker.cg3

TTS text processing-related

• tts-textproc-version.json: TTS text processing version number (from TTSTEXTPROCVERSION in configure.ac)

Shields.io Endpoint JSON Schema

These JSON files follow the shields.io endpoint JSON schema:

{
  "schemaVersion": 1,
  "label": "Badge Label",
  "message": "Badge Message",
  "color": "badge-color"
}

Color Conventions

Colors typically follow this pattern based on quality/maturity:

• black: Experimental/very poor
• red: Poor quality
• yellow: Medium quality (or pre-release version 0.x.y)
• green: Good quality (or stable version 1.x.y+)
• grey: N/A or no data

Build System

These files are automatically generated during the build process via rules in giella-core/am-shared/docs-dir-include.am using the FORCE target to ensure they are always regenerated.

Generation Scripts

Badge JSON files are generated by shell scripts in giella-core/scripts/:

• make-lemmacount.json.sh: Generates FST lemma count badge
• make-maturity.json.sh: Generates maturity level badge
• make-spellerbadge-json.sh: Generates speller suggestion quality badge
• make-version-json.sh: Generates version badges (FST, speller, hyphenator, gramcheck, TTS textproc, etc.)
  • Usage: make-version-json.sh INPUTDIR VARNAME [--label LABEL]
  • VARNAME can be: FST, SPELLER, HYPH, GRAMCHECK, TTSTEXTPROC, etc.
• make-gramcheck-rules-json.sh: Generates grammar checker error-detection rule count badge
  • Counts rules via count-gramcheck-rules.sh, which looks for rules of the form RULETYPE:rule-name (&errortag) ... in tools/grammarcheckers/grammarchecker.cg3

Using in README

To display a badge in a repository README, use shields.io endpoint syntax:

![Badge Label](https://img.shields.io/endpoint?url=https://giellalt.github.io/lang-xxx/badgedata/prefix-datatype.json)

Replace lang-xxx with the appropriate language code.

Last updated: September 25, 2026 Edit on GitHub