Skip to content

index.md

Latest commit

 

History

History
99 lines (74 loc) · 4.12 KB

index.md

File metadata and controls

99 lines (74 loc) · 4.12 KB

OVOS Localize

GitHub-native translation platform for OpenVoiceOS locale files with contextual cues extracted from skill source code.

Problem

GitLocalize treats OVOS locale files as flat key-value pairs. Translators work blind — they don't know if a file is a Padatious training set (needs 10+ diverse examples) or an Adapt keyword list (needs short keywords). OVOS Localize makes every native file type a first-class citizen with context from skill code.

Architecture

Fully GitHub-native: no server, no database, no Docker.

Component Tech Purpose
Data generation scripts/generate_data.py Clones skills, scans, outputs JSON to data/
Scheduling GitHub Actions cron Daily data refresh + auto-commit
Frontend index.html (static SPA) Tailwind + vanilla JS, served via GitHub Pages
CLI ovos-localize-cli CI pipeline validation

Key Modules

Module Path Description
parsers/ ovos_localize/parsers/ Per-file-type parsers (.intent, .voc, .dialog, .entity, .rx, .value, skill.json, settingsmeta)
analyzers/ ovos_localize/analyzers/ AST analysis of skill Python source → SkillAnalysis, ContextCard
validators/ ovos_localize/validators/ Per-type validation rules (slot preservation, regex compilation, diversity scoring)
sync/ ovos_localize/sync/ Git clone/pull, locale directory scanning
cli/ ovos_localize/cli/ Standalone validation CLI for CI
enums.py ovos_localize/enums.py FileType, IntentSystem enums (pure stdlib)

Data Pipeline

  1. skills.txt lists org/repo entries (one per line)
  2. scripts/generate_data.py clones each repo, scans locale dirs, runs AST analysis + validation
  3. Outputs to data/:
    • repos.json — skill index
    • coverage.json — language × skill coverage matrix
    • validation.json — aggregated validation results
    • skills/{id}.json — per-skill detail with entries, context cards, edit URLs

Static SPA Views

Route View Data Source
#/ Dashboard + heatmap repos.json, coverage.json
#/skill/{id} Skill detail + file list skills/{id}.json
#/skill/{id}/{file}/{lang} Three-panel translation viewer skills/{id}.json

Supported File Types

Extension Type Parser Key Validations
.intent Padatious training IntentParser Min 20 lines, slot preservation, diversity ≥0.25
.voc Adapt keywords VocabParser Non-empty, warn >5 words
.dialog TTS response variants DialogParser Variable preservation, ≥2 variants
.entity Slot examples EntityParser ≥5 examples
.rx Regex extraction RegexParser Compiles, named groups match source
.value Display→system CSV ValueParser Valid CSV, system values preserved
skill.json Metadata SkillJsonParser Valid JSON, required keys
settingsmeta.* Settings UI SettingsMetaParser Structure preserved

Quick Start

# Install (pure Python, zero dependencies)
uv pip install -e ".[dev]"

# Generate data (subset for quick test)
echo "OpenVoiceOS/ovos-skill-hello-world" > skills.txt
python scripts/generate_data.py

# Serve locally
python -m http.server 8000
# Open http://localhost:8000

# Validate a single skill
ovos-localize-cli --repo /path/to/skill --report-format github

See local-development.md for the full local workflow.

CI Integration

Skills can use the reusable workflow from gh-automations:

# .github/workflows/validate-translations.yml
name: Validate Translations
on: [pull_request]
jobs:
  validate:
    uses: OpenVoiceOS/gh-automations/.github/workflows/validate-translations.yml@dev

Cross-references

  • ovos-workshop — resource_files.py for file type resolution, skills/ovos.py for intent registration
  • gh-automations — Reusable CI workflows including validate-translations.yml
  • lang-support-tracker — Sister project using same GitHub-native pattern (cron → JSON → SPA)