Add multilingual audit CI pipeline + extract mandelblog_content_guard

2026-03-29 20:49:42 +02:00
parent 643aca26d0
commit e3bafd3a73
104 changed files with 3372 additions and 6 deletions
--- a/scripts/multilingual_audit_ci.py
+++ b/scripts/multilingual_audit_ci.py
@@ -0,0 +1,99 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import argparse
+import json
+from pathlib import Path
+
+
+def load_json(path: Path) -> dict:
+    return json.loads(path.read_text())
+
+
+def locale_rows(payload: dict) -> list[tuple[str, dict]]:
+    summary = payload.get("summary", {})
+    return [(locale, data) for locale, data in summary.items() if locale != "snippets"]
+
+
+def print_error(payload: dict) -> int:
+    error = payload.get("error")
+    if error:
+        print(f"AUDIT ERROR: {error}")
+        return 2
+    return 0
+
+
+def print_summary(payload: dict) -> tuple[int, int]:
+    total_block = 0
+    total_warn = 0
+    for locale, data in locale_rows(payload):
+        sev = data.get("by_severity", {})
+        block = int(sev.get("block", 0) or 0)
+        warn = int(sev.get("warn", 0) or 0)
+        log = int(sev.get("log", 0) or 0)
+        total_block += block
+        total_warn += warn
+        print(
+            f"LOCALE {locale}: issues_found={data.get('issues_found', 0)} "
+            f"issues_remaining={data.get('remaining_issues', 0)} "
+            f"block={block} warn={warn} log={log}"
+        )
+    return total_block, total_warn
+
+
+def print_regressions(current: dict, previous: dict) -> None:
+    prev_summary = {locale: data for locale, data in locale_rows(previous)}
+    regressions = []
+    for locale, data in locale_rows(current):
+        prev = prev_summary.get(locale, {})
+        cur_remaining = int(data.get("remaining_issues", 0) or 0)
+        prev_remaining = int(prev.get("remaining_issues", 0) or 0)
+        cur_sev = data.get("by_severity", {})
+        prev_sev = prev.get("by_severity", {})
+        delta = {
+            "remaining": cur_remaining - prev_remaining,
+            "block": int(cur_sev.get("block", 0) or 0) - int(prev_sev.get("block", 0) or 0),
+            "warn": int(cur_sev.get("warn", 0) or 0) - int(prev_sev.get("warn", 0) or 0),
+            "log": int(cur_sev.get("log", 0) or 0) - int(prev_sev.get("log", 0) or 0),
+        }
+        if any(value > 0 for value in delta.values()):
+            regressions.append((locale, delta))
+    if regressions:
+        print("REGRESSIONS:")
+        for locale, delta in regressions:
+            print(
+                f"- {locale}: remaining={delta['remaining']:+d} block={delta['block']:+d} "
+                f"warn={delta['warn']:+d} log={delta['log']:+d}"
+            )
+    else:
+        print("REGRESSIONS: none")
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--json", required=True, help="Current multilingual audit JSON file")
+    parser.add_argument("--previous-json", help="Optional previous audit JSON file for regression comparison")
+    args = parser.parse_args()
+
+    current = load_json(Path(args.json))
+    error_status = print_error(current)
+    if error_status:
+        return error_status
+    total_block, total_warn = print_summary(current)
+
+    if args.previous_json:
+        prev_path = Path(args.previous_json)
+        if prev_path.exists():
+            print_regressions(current, load_json(prev_path))
+        else:
+            print("REGRESSIONS: previous artifact not found")
+
+    if total_block > 0:
+        return 2
+    if total_warn > 0:
+        return 1
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())
--- a/scripts/run_remote_multilingual_audit.sh
+++ b/scripts/run_remote_multilingual_audit.sh
@@ -0,0 +1,72 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+: "${STAGING_AUDIT_HOST:?STAGING_AUDIT_HOST is required}"
+: "${STAGING_AUDIT_PROJECT_DIR:?STAGING_AUDIT_PROJECT_DIR is required}"
+: "${STAGING_AUDIT_MANAGE:?STAGING_AUDIT_MANAGE is required}"
+
+mkdir -p artifacts
+SSH_OPTS=${SSH_OPTS:-"-o StrictHostKeyChecking=accept-new"}
+if [[ -n "${STAGING_SSH_KEYFILE:-}" ]]; then
+  SSH_OPTS="$SSH_OPTS -i ${STAGING_SSH_KEYFILE}"
+fi
+AUDIT_TIMEOUT_SECONDS=${AUDIT_TIMEOUT_SECONDS:-300}
+OUT_FILE="artifacts/multilingual-audit.json"
+TMP_FILE="${OUT_FILE}.tmp"
+
+write_failure_json() {
+  python3 - <<PY > "$OUT_FILE"
+import json
+print(json.dumps({
+  "run_id": None,
+  "total_urls_checked": 0,
+  "issues_found": 0,
+  "summary": {},
+  "issues": {},
+  "error": ${1@Q}
+}, indent=2))
+PY
+}
+
+REMOTE_CMD="cd '${STAGING_AUDIT_PROJECT_DIR}' && '${STAGING_AUDIT_MANAGE}' audit_locales --format=json"
+set +e
+SSH_OPTS="$SSH_OPTS" STAGING_AUDIT_HOST="$STAGING_AUDIT_HOST" REMOTE_CMD="$REMOTE_CMD" AUDIT_TIMEOUT_SECONDS="$AUDIT_TIMEOUT_SECONDS" python3 - <<'PY' > "$TMP_FILE"
+import os
+import shlex
+import subprocess
+import sys
+
+ssh_opts = shlex.split(os.environ["SSH_OPTS"])
+cmd = ["ssh", *ssh_opts, os.environ["STAGING_AUDIT_HOST"], os.environ["REMOTE_CMD"]]
+try:
+    proc = subprocess.run(
+        cmd,
+        check=True,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.PIPE,
+        text=True,
+        timeout=int(os.environ["AUDIT_TIMEOUT_SECONDS"]),
+    )
+    sys.stdout.write(proc.stdout)
+    sys.stderr.write(proc.stderr)
+except subprocess.TimeoutExpired as exc:
+    sys.stderr.write(exc.stderr or "")
+    raise SystemExit(124)
+except subprocess.CalledProcessError as exc:
+    sys.stdout.write(exc.stdout or "")
+    sys.stderr.write(exc.stderr or "")
+    raise SystemExit(exc.returncode)
+PY
+rc=$?
+set -e
+if [[ $rc -eq 0 ]]; then
+  mv "$TMP_FILE" "$OUT_FILE"
+  exit 0
+fi
+rm -f "$TMP_FILE"
+if [[ $rc -eq 124 ]]; then
+  write_failure_json "Remote multilingual audit timed out after ${AUDIT_TIMEOUT_SECONDS}s"
+else
+  write_failure_json "Remote multilingual audit failed with exit status ${rc}"
+fi
+exit $rc