Add GitHub Action for non-blocking translation validation on PRs

.github/scripts/validate_translations.py

@@ -0,0 +1,329 @@
+#!/usr/bin/env python3
+"""
+Translation validation script for Open Food Facts mobile app.
+
+This script validates translation files (.arb) for common issues:
+1. app_name should always be "Open Food Facts" (not translated)
+2. Brand terms (Nutri-Score, NOVA, Open Food Facts, etc.) should be preserved
+3. URLs should be valid and not broken
+4. All ARB files should have required keys
+
+The script outputs validation warnings in a format suitable for GitHub PR comments.
+"""
+
+import json
+import os
+import re
+import sys
+from pathlib import Path
+from typing import Dict, List, Set, Tuple
+
+
+# Brand terms that should be preserved in translations
+BRAND_TERMS = {
+    'Nutri-Score': [
+        'nutriscore_generic',
+        'nutriscore_a',
+        'nutriscore_b',
+        'nutriscore_c',
+        'nutriscore_d',
+        'nutriscore_e',
+        'nutriscore_new_formula',
+        'nutriscore_new_formula_title',
+        'nutriscore_unknown',
+        'nutriscore_unknown_new_formula',
+        'nutriscore_not_applicable',
+        'nutriscore_not_applicable_new_formula',
+        'new_product_title_nutriscore',
+        'product_improvement_add_category',
+        'product_improvement_add_nutrition_facts',
+        'product_improvement_add_nutrition_facts_and_category',
+    ],
+    'NOVA': [
+        'nova_group_1',
+        'nova_group_2',
+        'nova_group_3',
+        'nova_group_4',
+        'nova_group_unknown',
+        'nova_group_generic_new',
+        'new_product_title_nova',
+    ],
+    'Open Food Facts': [
+        'help_improve_country',
+        'contribute_join_skill_pool',
+        'new_product_title_pictures_details',
+        'preferences_page_contribute_project_subtitle',
+    ],
+    'Open Products Facts': [
+        'search_product_filter_visibility_subtitle',
+    ],
+    'Open Beauty Facts': [
+        'search_product_filter_visibility_subtitle',
+    ],
+    'Open Pet Food Facts': [
+        'search_product_filter_visibility_subtitle',
+    ],
+}
+
+# URL keys to validate
+URL_KEYS = [
+    'donate_url',
+    'tiktok_link',
+    'instagram_link',
+    'twitter_link',
+    'mastodon_link',
+    'bsky_link',
+]
+
+
+def load_arb_file(file_path: Path) -> Dict:
+    """Load and parse an ARB file."""
+    try:
+        with open(file_path, 'r', encoding='utf-8') as f:
+            return json.load(f)
+    except Exception as e:
+        print(f"Error loading {file_path}: {e}", file=sys.stderr)
+        return {}
+
+
+def get_arb_files(l10n_dir: Path) -> List[Path]:
+    """Get all ARB files from the l10n directory."""
+    return [f for f in l10n_dir.glob('*.arb') if not f.name.endswith('_template.arb')]
+
+
+def validate_app_name(arb_files: List[Path]) -> List[str]:
+    """Validate that app_name is always 'Open Food Facts'."""
+    issues = []
+
+    for arb_file in arb_files:
+        translations = load_arb_file(arb_file)
+        if 'app_name' in translations:
+            app_name = translations['app_name']
+            if app_name != 'Open Food Facts':
+                issues.append(f"  - `{arb_file.name}`: \"{app_name}\"")
+
+    return issues
+
+
+def validate_brand_terms(arb_files: List[Path], en_translations: Dict) -> Dict[str, List[str]]:
+    """Validate that brand terms are preserved in translations."""
+    all_issues = {}
+
+    for brand_term, keys_to_check in BRAND_TERMS.items():
+        issues = []
+
+        for arb_file in arb_files:
+            if arb_file.name == 'app_en.arb':
+                continue
+
+            translations = load_arb_file(arb_file)
+
+            for key in keys_to_check:
+                if key not in translations:
+                    continue
+
+                en_value = en_translations.get(key)
+                if not en_value or not isinstance(en_value, str):
+                    continue
+
+                if brand_term not in en_value or en_value.startswith('{'):
+                    continue
+
+                translated_value = translations[key]
+                if not isinstance(translated_value, str) or translated_value.startswith('{'):
+                    continue
+
+                if brand_term not in translated_value:
+                    issues.append(f"  - `{arb_file.name}`: `{key}`")
+                    issues.append(f"    - English: \"{en_value}\"")
+                    issues.append(f"    - Translation: \"{translated_value}\"")
+
+        if issues:
+            all_issues[brand_term] = issues
+
+    return all_issues
+
+
+def validate_urls(arb_files: List[Path]) -> Tuple[List[str], List[str]]:
+    """Validate URLs in translation files."""
+    url_pattern = re.compile(r'^https?://[^\s<>"{}|\\^`\[\]]+$', re.IGNORECASE)
+    url_in_text_pattern = re.compile(r'https?://[^\s<>"{}|\\^`\[\]]+', re.IGNORECASE)
+    domain_pattern = re.compile(r'^https?://[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}')
+
+    url_key_issues = []
+    url_text_issues = []
+
+    for arb_file in arb_files:
+        translations = load_arb_file(arb_file)
+
+        # Check URL keys
+        for url_key in URL_KEYS:
+            if url_key in translations:
+                value = translations[url_key]
+                if value and not url_pattern.match(value):
+                    url_key_issues.append(f"  - `{arb_file.name}`: `{url_key}` = \"{value}\"")
+
+        # Check URLs in text
+        for key, value in translations.items():
+            if key.startswith('@') or not isinstance(value, str):
+                continue
+
+            matches = url_in_text_pattern.findall(value)
+            for url in matches:
+                clean_url = re.sub(r'[.,;!?]+$', '', url)
+                if not domain_pattern.match(clean_url):
+                    url_text_issues.append(f"  - `{arb_file.name}`: `{key}`")
+                    url_text_issues.append(f"    - URL: \"{clean_url}\"")
+
+    return url_key_issues, url_text_issues
+
+
+def validate_required_keys(arb_files: List[Path]) -> List[str]:
+    """Validate that all ARB files have required keys."""
+    missing = []
+
+    for arb_file in arb_files:
+        translations = load_arb_file(arb_file)
+        if 'app_name' not in translations:
+            missing.append(f"  - `{arb_file.name}`")
+
+    return missing
+
+
+def format_validation_comment(
+    app_name_issues: List[str],
+    brand_term_issues: Dict[str, List[str]],
+    url_key_issues: List[str],
+    url_text_issues: List[str],
+    missing_keys: List[str]
+) -> str:
+    """Format validation results as a GitHub PR comment."""
+    sections = []
+
+    # Header
+    sections.append("## 🔍 Translation Validation Report")
+    sections.append("")
+    sections.append("This is an automated check of translation quality. Issues found are informational and don't block the PR.")
+    sections.append("")
+
+    has_issues = False
+
+    # App name issues
+    if app_name_issues:
+        has_issues = True
+        sections.append("### ⚠️ App Name Should Not Be Translated")
+        sections.append("")
+        sections.append("The `app_name` key should always be `\"Open Food Facts\"` (not translated).")
+        sections.append("")
+        sections.append(f"Found {len(app_name_issues)} file(s) with translated app_name:")
+        sections.append("")
+        sections.extend(app_name_issues[:10])  # Limit to 10 examples
+        if len(app_name_issues) > 10:
+            sections.append(f"  - ... and {len(app_name_issues) - 10} more")
+        sections.append("")
+
+    # Brand term issues
+    for brand_term, issues in brand_term_issues.items():
+        if issues:
+            has_issues = True
+            num_issues = len(issues) // 3  # Each issue has 3 lines
+            sections.append(f"### ⚠️ Brand Term \"{brand_term}\" Should Be Preserved")
+            sections.append("")
+            sections.append(f"Found {num_issues} translation(s) where \"{brand_term}\" was not preserved:")
+            sections.append("")
+            sections.extend(issues[:15])  # Limit to 5 examples (3 lines each)
+            if len(issues) > 15:
+                sections.append(f"  - ... and {num_issues - 5} more")
+            sections.append("")
+
+    # URL issues
+    if url_key_issues:
+        has_issues = True
+        sections.append("### ⚠️ Invalid URLs in URL Fields")
+        sections.append("")
+        sections.append(f"Found {len(url_key_issues)} invalid URL(s):")
+        sections.append("")
+        sections.extend(url_key_issues[:10])
+        if len(url_key_issues) > 10:
+            sections.append(f"  - ... and {len(url_key_issues) - 10} more")
+        sections.append("")
+
+    if url_text_issues:
+        has_issues = True
+        num_issues = len(url_text_issues) // 2
+        sections.append("### ⚠️ Invalid URLs in Text Strings")
+        sections.append("")
+        sections.append(f"Found {num_issues} issue(s):")
+        sections.append("")
+        sections.extend(url_text_issues[:10])
+        if len(url_text_issues) > 10:
+            sections.append(f"  - ... and {num_issues - 5} more")
+        sections.append("")
+
+    # Missing keys
+    if missing_keys:
+        has_issues = True
+        sections.append("### ⚠️ Missing Required Keys")
+        sections.append("")
+        sections.append(f"Found {len(missing_keys)} file(s) without `app_name` key:")
+        sections.append("")
+        sections.extend(missing_keys[:10])
+        if len(missing_keys) > 10:
+            sections.append(f"  - ... and {len(missing_keys) - 10} more")
+        sections.append("")
+
+    # Footer
+    if has_issues:
+        sections.append("---")
+        sections.append("")
+        sections.append("**Note:** These are informational warnings to help improve translation quality. ")
+        sections.append("Some variations may be acceptable based on language requirements. ")
+        sections.append("Please review and decide if any action is needed.")
+    else:
+        sections.append("### ✅ No Issues Found")
+        sections.append("")
+        sections.append("All translations look good!")
+
+    return "\n".join(sections)
+
+
+def main():
+    """Main function to run translation validation."""
+    # Find the l10n directory
+    repo_root = Path(__file__).parent.parent.parent
+    l10n_dir = repo_root / 'packages' / 'smooth_app' / 'lib' / 'l10n'
+
+    if not l10n_dir.exists():
+        print(f"Error: l10n directory not found at {l10n_dir}", file=sys.stderr)
+        sys.exit(1)
+
+    # Get all ARB files
+    arb_files = get_arb_files(l10n_dir)
+
+    # Load English translations for reference
+    en_file = l10n_dir / 'app_en.arb'
+    en_translations = load_arb_file(en_file)
+
+    # Run validations
+    app_name_issues = validate_app_name(arb_files)
+    brand_term_issues = validate_brand_terms(arb_files, en_translations)
+    url_key_issues, url_text_issues = validate_urls(arb_files)
+    missing_keys = validate_required_keys(arb_files)
+
+    # Format and output comment
+    comment = format_validation_comment(
+        app_name_issues,
+        brand_term_issues,
+        url_key_issues,
+        url_text_issues,
+        missing_keys
+    )
+
+    print(comment)
+
+    # Exit with 0 (success) - we don't want to block PRs
+    sys.exit(0)
+
+
+if __name__ == '__main__':
+    main()

.github/workflows/translation-validation.yml

@@ -0,0 +1,46 @@
+name: Translation Validation
+
+permissions:
+  contents: read
+  pull-requests: write
+
+on:
+  pull_request:
+    paths:
+      - 'packages/smooth_app/lib/l10n/app_*.arb'
+    types: [opened, synchronize, reopened]
+
+jobs:
+  validate-translations:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v5
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.x'
+
+      - name: Run translation validation
+        id: validation
+        run: |
+          python .github/scripts/validate_translations.py > validation_report.md
+          echo "REPORT_FILE=validation_report.md" >> $GITHUB_OUTPUT
+
+      - name: Find existing comment
+        uses: peter-evans/find-comment@v3
+        id: find-comment
+        with:
+          issue-number: ${{ github.event.pull_request.number }}
+          comment-author: 'github-actions[bot]'
+          body-includes: '🔍 Translation Validation Report'
+
+      - name: Create or update comment
+        uses: peter-evans/create-or-update-comment@v4
+        with:
+          comment-id: ${{ steps.find-comment.outputs.comment-id }}
+          issue-number: ${{ github.event.pull_request.number }}
+          body-path: validation_report.md
+          edit-mode: replace