Track successful sync runs

sync.py

@@ -10,9 +10,11 @@ import subprocess
 import sys
 import tempfile
 import zipfile
+from datetime import datetime, timezone
 from io import BytesIO
 from pathlib import Path
 from urllib.parse import urlencode
+from zoneinfo import ZoneInfo
 import xml.etree.ElementTree as ET
 
 import requests
@@ -21,6 +23,7 @@ import requests
 ROOT = Path(__file__).resolve().parent
 WORKBOOK_ID = os.getenv("WORKBOOK_ID", "1DPbdVsX3SG_kgDXs2j34mDAgplRSfcAWKKG9yqalFmw")
 PUBLISH_PATH = Path("data/merged.html")
+SYNC_LOG_PATH = Path("data/sync-log.json")
 GVIZ_PREFIX = "google.visualization.Query.setResponse("
 
 
@@ -171,18 +174,28 @@ def run_git(args: list[str], env: dict[str, str]) -> subprocess.CompletedProcess
     )
 
 
-def publish(content: bytes) -> None:
-    target = ROOT / PUBLISH_PATH
-    if target.exists() and target.read_bytes() == content:
-        print("data/merged.html is unchanged")
+def publish(content: bytes, sync_log: dict[str, object]) -> None:
+    log_content = (json.dumps(sync_log, ensure_ascii=False, indent=2) + "\n").encode("utf-8")
+    targets = {
+        PUBLISH_PATH: content,
+        SYNC_LOG_PATH: log_content,
+    }
+    changed = False
+    for relative_path, target_content in targets.items():
+        target = ROOT / relative_path
+        if not target.exists() or target.read_bytes() != target_content:
+            changed = True
+        target.parent.mkdir(parents=True, exist_ok=True)
+        target.write_bytes(target_content)
+    if not changed:
+        print("data/merged.html and data/sync-log.json are unchanged")
         return
-    target.parent.mkdir(parents=True, exist_ok=True)
-    target.write_bytes(content)
 
     env = os.environ.copy()
     env["GIT_TERMINAL_PROMPT"] = "0"
     askpass_path: str | None = None
     credential = env.get("GIT_CREDENTIAL_SECRET", "")
+    paths = [str(PUBLISH_PATH), str(SYNC_LOG_PATH)]
     if credential:
         with tempfile.NamedTemporaryFile("w", prefix="ttshop-askpass-", delete=False) as file:
             file.write(
@@ -197,22 +210,23 @@ def publish(content: bytes) -> None:
         env["GIT_ASKPASS"] = askpass_path
 
     try:
-        run_git(["add", "--", str(PUBLISH_PATH)], env)
+        run_git(["add", "--", *paths], env)
         staged = subprocess.run(
-            ["git", "diff", "--cached", "--quiet", "--", str(PUBLISH_PATH)],
+            ["git", "diff", "--cached", "--quiet", "--", *paths],
             cwd=ROOT, env=env, check=False,
         )
         if staged.returncode not in (0, 1):
             raise RuntimeError("could not inspect staged output")
         if staged.returncode == 0:
-            print("data/merged.html is unchanged")
+            print("data/merged.html and data/sync-log.json are unchanged")
             return
         run_git([
             "-c", "user.name=TTSHOP Sync", "-c", "user.email=ttshop-sync@localhost",
-            "commit", "--no-gpg-sign", "-m", "Update merged Google Sheet",
+            "commit", "--only", "--no-gpg-sign", "-m", "Update merged Google Sheet and sync log",
+            "--", *paths,
         ], env)
         run_git(["push", "origin", "HEAD:main"], env)
-        print("published data/merged.html")
+        print("published data/merged.html and data/sync-log.json")
     finally:
         if askpass_path:
             Path(askpass_path).unlink(missing_ok=True)
@@ -222,13 +236,23 @@ def sync() -> None:
     export_url = f"https://docs.google.com/spreadsheets/d/{WORKBOOK_ID}/export?format=xlsx"
     response = requests.get(export_url, timeout=30)
     response.raise_for_status()
+    worksheet_names = sheet_names(response.content)
     sheets = []
-    for name in sheet_names(response.content):
+    for name in worksheet_names:
         headers, rows = normalize_sheet(name, read_gviz(name))
         if headers:
             sheets.append((name, headers, rows))
     headers, rows = merge_sheets(sheets)
-    publish(render_html(headers, rows))
+    content = render_html(headers, rows)
+    success_at = datetime.now(timezone.utc)
+    sync_log = {
+        "last_success_at": success_at.isoformat().replace("+00:00", "Z"),
+        "last_success_at_shanghai": success_at.astimezone(ZoneInfo("Asia/Shanghai")).isoformat(),
+        "worksheet_count": len(worksheet_names),
+        "record_count": len(rows),
+        "artifact_bytes": len(content),
+    }
+    publish(content, sync_log)
 
 
 def notify_failure(error: Exception) -> None: