Files
grm/scripts/ci/check_translations.py
T

199 lines
7.0 KiB
Python

#!/usr/bin/env python3
"""Check translation files for gaps, dead keys, and missing languages.
Validates two separate translation sets:
1. GRM tool translations: ``src/gitea_runner_manager/translations.json``
— keys used by ``src/gitea_runner_manager/*.py``
2. CI/dev tool translations: ``scripts/translations.json``
— keys used by ``scripts/**/*.py``
Checks performed (all fail with exit code 1 on error):
- **Missing keys**: a ``_()`` call in code has no entry in the corresponding
translations file.
- **Dead keys**: a key in a translations file is not used in any code.
- **Missing languages**: a key exists but is missing one of the 5 supported
languages (en, bg, de, ru, zh). Reported as a warning, not an error.
Usage::
python3 scripts/ci/check_translations.py
python3 scripts/ci/check_translations.py --strict # warnings are errors
"""
from __future__ import annotations
import ast
import json
import sys
from dataclasses import dataclass, field
from pathlib import Path
import click
REPO_ROOT = Path(__file__).resolve().parent.parent.parent
SUPPORTED_LANGS = ("en", "bg", "de", "ru", "zh")
GRM_SRC_DIR = REPO_ROOT / "src" / "gitea_runner_manager"
GRM_TRANS_FILE = GRM_SRC_DIR / "translations.json"
CI_SRC_DIR = REPO_ROOT / "scripts"
CI_TRANS_FILE = CI_SRC_DIR / "translations.json"
# Functions that wrap _() and receive a translation key as first arg.
# Their string-literal arguments should be treated as translation keys.
_I18N_WRAPPERS = {"_handle_errors"}
# Known dynamic keys used via _(variable) that can't be detected by AST.
# These are status strings set as variable values and passed to _().
DYNAMIC_KEYS = {"completed", "pending", "in_progress", "failed", "active", "inactive", "unknown"}
@dataclass
class TranslationCheckResult:
"""Result of a translation check for one translation set."""
name: str
src_dir: Path
trans_file: Path
used_keys: set[str] = field(default_factory=set)
defined_keys: set[str] = field(default_factory=set)
missing_keys: set[str] = field(default_factory=set)
dead_keys: set[str] = field(default_factory=set)
missing_langs: dict[str, list[str]] = field(default_factory=dict)
errors: list[str] = field(default_factory=list)
warnings: list[str] = field(default_factory=list)
def extract_keys(filepath: Path) -> set[str]:
"""Extract translation keys from a Python file using AST.
Detects:
- ``_("key")`` calls with string-literal first argument
- ``_handle_errors("key")`` and other wrapper calls (see ``_I18N_WRAPPERS``)
"""
try:
tree = ast.parse(filepath.read_text(encoding="utf-8"), filename=str(filepath))
except SyntaxError:
return set()
keys: set[str] = set()
for node in ast.walk(tree):
if isinstance(node, ast.Call):
func = node.func
if (
isinstance(func, ast.Name)
and func.id in ("_", *_I18N_WRAPPERS)
and node.args
and isinstance(node.args[0], ast.Constant)
and isinstance(node.args[0].value, str)
):
keys.add(node.args[0].value)
return keys
def collect_keys(src_dir: Path) -> set[str]:
"""Collect all translation keys from .py files in a directory tree.
Also includes known dynamic keys (see ``DYNAMIC_KEYS``) that are used
via ``_(variable)`` and can't be detected by AST scanning.
"""
keys: set[str] = set()
for pyfile in src_dir.rglob("*.py"):
if pyfile.name == "i18n.py":
continue
keys |= extract_keys(pyfile)
# Add dynamic keys for the GRM source directory
if src_dir == GRM_SRC_DIR:
keys |= DYNAMIC_KEYS
return keys
def check_translation_set(name: str, src_dir: Path, trans_file: Path) -> TranslationCheckResult:
"""Check one translation set for gaps and dead keys."""
result = TranslationCheckResult(name=name, src_dir=src_dir, trans_file=trans_file)
# Collect used keys from source code
result.used_keys = collect_keys(src_dir)
# Load defined keys from translations file
if not trans_file.exists():
result.errors.append(f"Translations file not found: {trans_file}")
return result
translations = json.loads(trans_file.read_text(encoding="utf-8"))
result.defined_keys = set(translations.keys())
# Check for missing keys (used in code but not in translations)
result.missing_keys = result.used_keys - result.defined_keys
for key in sorted(result.missing_keys):
result.errors.append(f"Missing key in {name}: {key!r}")
# Check for dead keys (in translations but not used in code)
result.dead_keys = result.defined_keys - result.used_keys
for key in sorted(result.dead_keys):
result.warnings.append(f"Dead key in {name}: {key!r}")
# Check for missing languages
for key, langs in translations.items():
missing = [lang for lang in SUPPORTED_LANGS if lang not in langs]
if missing:
result.missing_langs[key] = missing
result.warnings.append(f"Missing languages {missing} for key {key!r} in {name}")
return result
def print_result(result: TranslationCheckResult) -> None:
"""Print check results in a human-readable format."""
click.echo(f"\n=== {result.name} ===")
click.echo(f" Source dir: {result.src_dir}")
click.echo(f" Translations: {result.trans_file}")
click.echo(f" Used keys: {len(result.used_keys)}")
click.echo(f" Defined keys: {len(result.defined_keys)}")
click.echo(f" Missing keys: {len(result.missing_keys)}")
click.echo(f" Dead keys: {len(result.dead_keys)}")
click.echo(f" Missing langs: {len(result.missing_langs)} keys")
for err in result.errors:
click.echo(f" ERROR: {err}", err=True)
for warn in result.warnings:
click.echo(f" WARN: {warn}", err=True)
if not result.errors and not result.warnings:
click.echo(" All good!")
@click.command()
@click.option("--strict", is_flag=True, default=False, help="Treat warnings as errors.")
def main(strict: bool) -> None:
"""Check translation files for gaps, dead keys, and missing languages."""
results = [
check_translation_set("GRM tool", GRM_SRC_DIR, GRM_TRANS_FILE),
check_translation_set("CI/dev tools", CI_SRC_DIR, CI_TRANS_FILE),
]
has_errors = False
has_warnings = False
for result in results:
print_result(result)
if result.errors:
has_errors = True
if result.warnings:
has_warnings = True
click.echo()
if has_errors:
click.echo("FAIL: Translation check found errors.", err=True)
sys.exit(1)
if strict and has_warnings:
click.echo("FAIL: Translation check found warnings (--strict mode).", err=True)
sys.exit(1)
if has_warnings:
click.echo("PASS with warnings: Translation check passed (warnings present).")
else:
click.echo("PASS: All translations are complete and up to date.")
if __name__ == "__main__": # pragma: no cover
main() # pragma: no cover