Enrich Feishu daily digest with auto-refresh status.

Include Gitea web-daily/weekly/patch conclusions and Cloudflare Pages deploys; ship the refresh workflows so those signals exist.

Co-authored-by: Cursor <cursoragent@cursor.com>

notify_site_traffic.py

@@ -1,18 +1,20 @@
-"""Daily site traffic digest -> Feishu webhook.
+"""Daily site digest -> Feishu webhook.
 
-Pulls Cloudflare edge analytics for dota2.refining.dev, estimates human
-homepage visits (known browser UA on `/`, plus RUM beacons), and posts a
-Feishu custom-bot card.
+1) Cloudflare edge analytics for dota2.refining.dev (human homepage estimate)
+2) Auto-refresh status: Gitea Actions (web-daily / web-weekly / web-patch)
+   + Cloudflare Pages production deploys that day
 
 Credentials (env, prefer keyzoo inject names):
   CLOUDFLARE_EMAIL / KEYZOO_ASSET_META_USERNAME
   CLOUDFLARE_API_KEY / KEYZOO_ASSET_SECRET_GLOBAL_API_KEY
   FEISHU_WEBHOOK_URL / KEYZOO_ASSET_SECRET_FEISHU_WEBHOOK_URL
+  GITEA_TOKEN / KEYZOO_ASSET_SECRET_PERSONAL_ACCESS_TOKEN_GITEA_1  (optional;
+    without it the refresh section shows \"未配置 token\")
 
 Usage:
-  python notify_site_traffic.py              # yesterday (UTC date)
+  python notify_site_traffic.py              # yesterday CST
   python notify_site_traffic.py --day 2026-07-27
-  python notify_site_traffic.py --dry-run    # print card, do not POST
+  python notify_site_traffic.py --dry-run
 """
 
 from __future__ import annotations
@@ -20,16 +22,30 @@ from __future__ import annotations
 import argparse
 import json
 import os
+import re
 import sys
 import urllib.error
 import urllib.request
 from dataclasses import dataclass, field
 from datetime import date, datetime, timedelta, timezone
+from pathlib import Path
 from typing import Any
 
 HOST = "dota2.refining.dev"
 ZONE_NAME = "refining.dev"
 SITE_URL = "https://dota2.refining.dev"
+PAGES_PROJECT = "climperor-relations"
+GITEA_URL_DEFAULT = "https://gitea.refining.dev"
+GITEA_OWNER = "refining"
+GITEA_REPO = "climperor"
+CST = timezone(timedelta(hours=8))
+
+# (workflow file basename, short label)
+REFRESH_WORKFLOWS: tuple[tuple[str, str], ...] = (
+    ("web-daily.yml", "每日数据"),
+    ("web-weekly.yml", "每周 Meta"),
+    ("web-patch.yml", "版本检测"),
+)
 
 HUMAN_BROWSERS = frozenset(
     {
@@ -81,17 +97,64 @@ def feishu_url() -> str:
     return url
 
 
-def http_json(url: str, *, method: str = "GET", body: dict | None = None, headers: dict | None = None) -> dict:
+def gitea_token() -> str | None:
+    return _env(
+        "GITEA_TOKEN",
+        "KEYZOO_ASSET_SECRET_PERSONAL_ACCESS_TOKEN_GITEA_1",
+        "KEYZOO_ASSET_SECRET_PERSONAL_ACCESS_TOKEN__GITEA_1",
+        "KEYZOO_ASSET_TOKEN",
+    )
+
+
+def gitea_base() -> str:
+    return (
+        _env("GITEA_URL", "KEYZOO_ASSET_META_URL") or GITEA_URL_DEFAULT
+    ).rstrip("/")
+
+
+def http_json(
+    url: str,
+    *,
+    method: str = "GET",
+    body: dict | None = None,
+    headers: dict | None = None,
+) -> dict | list:
     data = None if body is None else json.dumps(body).encode()
     req = urllib.request.Request(url, data=data, method=method, headers=headers or {})
     try:
         with urllib.request.urlopen(req, timeout=60) as r:
-            return json.loads(r.read().decode())
+            raw = r.read().decode()
+            if not raw:
+                return {}
+            return json.loads(raw)
     except urllib.error.HTTPError as e:
         raw = e.read().decode(errors="replace")
         raise SystemExit(f"HTTP {e.code} {url}: {raw[:800]}") from e
 
 
+def http_json_soft(
+    url: str,
+    *,
+    method: str = "GET",
+    body: dict | None = None,
+    headers: dict | None = None,
+) -> tuple[int, Any]:
+    data = None if body is None else json.dumps(body).encode()
+    req = urllib.request.Request(url, data=data, method=method, headers=headers or {})
+    try:
+        with urllib.request.urlopen(req, timeout=60) as r:
+            raw = r.read().decode()
+            if not raw:
+                return r.status, {}
+            return r.status, json.loads(raw)
+    except urllib.error.HTTPError as e:
+        raw = e.read().decode(errors="replace")
+        try:
+            return e.code, json.loads(raw)
+        except json.JSONDecodeError:
+            return e.code, raw
+
+
 def gql(query: str, variables: dict) -> dict:
     data = http_json(
         "https://api.cloudflare.com/client/v4/graphql",
@@ -99,11 +162,24 @@ def gql(query: str, variables: dict) -> dict:
         body={"query": query, "variables": variables},
         headers=cf_headers(),
     )
-    if data.get("errors"):
-        raise SystemExit("GraphQL errors: " + json.dumps(data["errors"], ensure_ascii=False)[:1500])
+    if isinstance(data, dict) and data.get("errors"):
+        raise SystemExit(
+            "GraphQL errors: " + json.dumps(data["errors"], ensure_ascii=False)[:1500]
+        )
+    assert isinstance(data, dict)
     return data
 
 
+def read_site_version() -> str:
+    path = Path(__file__).resolve().parent / "web" / "relations" / "config.js"
+    try:
+        text = path.read_text(encoding="utf-8")
+    except OSError:
+        return "?"
+    m = re.search(r'SITE_VERSION\s*=\s*"([^"]+)"', text)
+    return m.group(1) if m else "?"
+
+
 @dataclass
 class DayStats:
     day: date
@@ -115,15 +191,57 @@ class DayStats:
     rum_human: int = 0
     rum_other: int = 0
     bytes: int = 0
-    top_countries: list[tuple[str, int, int]] = field(default_factory=list)  # name, req, vis
+    top_countries: list[tuple[str, int, int]] = field(default_factory=list)
     status: list[tuple[Any, int]] = field(default_factory=list)
 
 
+@dataclass
+class WorkflowDayStatus:
+    file: str
+    label: str
+    runs: list[dict] = field(default_factory=list)  # status/conclusion/started_at
+    note: str = ""  # e.g. missing token / none
+
+    @property
+    def ok_count(self) -> int:
+        return sum(1 for r in self.runs if r.get("conclusion") == "success")
+
+    @property
+    def fail_count(self) -> int:
+        return sum(
+            1
+            for r in self.runs
+            if r.get("conclusion") not in (None, "success")
+            and r.get("status") == "completed"
+        )
+
+    @property
+    def running_count(self) -> int:
+        return sum(1 for r in self.runs if r.get("status") in ("queued", "in_progress", "waiting"))
+
+
+@dataclass
+class PagesDeploy:
+    created_on: datetime
+    status: str
+    commit: str
+    url: str
+
+
+@dataclass
+class DigestExtras:
+    site_version: str
+    workflows: list[WorkflowDayStatus] = field(default_factory=list)
+    deploys: list[PagesDeploy] = field(default_factory=list)
+    gitea_note: str = ""
+
+
 def zone_id() -> str:
     d = http_json(
         f"https://api.cloudflare.com/client/v4/zones?name={ZONE_NAME}",
         headers=cf_headers(),
     )
+    assert isinstance(d, dict)
     rows = d.get("result") or []
     if not rows:
         raise SystemExit(f"zone not found: {ZONE_NAME}")
@@ -230,6 +348,142 @@ def fetch_day(zid: str, day: date) -> DayStats:
     return st
 
 
+def day_window_utc(day: date) -> tuple[datetime, datetime]:
+    """Inclusive start / exclusive end of CST calendar day, as aware UTC datetimes."""
+    start_cst = datetime(day.year, day.month, day.day, tzinfo=CST)
+    end_cst = start_cst + timedelta(days=1)
+    return start_cst.astimezone(timezone.utc), end_cst.astimezone(timezone.utc)
+
+
+def parse_gitea_time(s: str | None) -> datetime | None:
+    if not s:
+        return None
+    try:
+        # Gitea may return +08:00 or Z
+        if s.endswith("Z"):
+            return datetime.fromisoformat(s.replace("Z", "+00:00"))
+        return datetime.fromisoformat(s)
+    except ValueError:
+        return None
+
+
+def fetch_workflow_statuses(day: date) -> tuple[list[WorkflowDayStatus], str]:
+    tok = gitea_token()
+    if not tok:
+        return (
+            [
+                WorkflowDayStatus(file=f, label=lab, note="未配置 GITEA_TOKEN")
+                for f, lab in REFRESH_WORKFLOWS
+            ],
+            "未配置 GITEA_TOKEN，跳过 Actions 查询",
+        )
+
+    headers = {
+        "Authorization": f"token {tok}",
+        "Accept": "application/json",
+        "Content-Type": "application/json",
+    }
+    start_utc, end_utc = day_window_utc(day)
+    code, payload = http_json_soft(
+        f"{gitea_base()}/api/v1/repos/{GITEA_OWNER}/{GITEA_REPO}/actions/runs?limit=50",
+        headers=headers,
+    )
+    if code != 200:
+        return (
+            [
+                WorkflowDayStatus(file=f, label=lab, note=f"API {code}")
+                for f, lab in REFRESH_WORKFLOWS
+            ],
+            f"Gitea runs API HTTP {code}",
+        )
+
+    if isinstance(payload, dict):
+        items = payload.get("workflow_runs") or payload.get("runs") or []
+    elif isinstance(payload, list):
+        items = payload
+    else:
+        items = []
+
+    by_file: dict[str, list[dict]] = {f: [] for f, _ in REFRESH_WORKFLOWS}
+    for x in items:
+        path = str(x.get("path") or "")
+        # path like "web-daily.yml@refs/heads/main"
+        base = path.split("@", 1)[0]
+        if base not in by_file:
+            continue
+        ts = parse_gitea_time(x.get("started_at") or x.get("completed_at"))
+        if ts is None:
+            continue
+        ts_utc = ts.astimezone(timezone.utc)
+        if not (start_utc <= ts_utc < end_utc):
+            continue
+        by_file[base].append(
+            {
+                "id": x.get("id"),
+                "status": x.get("status"),
+                "conclusion": x.get("conclusion"),
+                "event": x.get("event"),
+                "started_at": x.get("started_at"),
+                "completed_at": x.get("completed_at"),
+                "html_url": x.get("html_url"),
+            }
+        )
+
+    out: list[WorkflowDayStatus] = []
+    for f, lab in REFRESH_WORKFLOWS:
+        runs = by_file.get(f) or []
+        note = "" if runs else "当日无运行"
+        out.append(WorkflowDayStatus(file=f, label=lab, runs=runs, note=note))
+    return out, ""
+
+
+def cf_account_id() -> str:
+    d = http_json(
+        "https://api.cloudflare.com/client/v4/accounts",
+        headers=cf_headers(),
+    )
+    assert isinstance(d, dict)
+    rows = d.get("result") or []
+    if not rows:
+        raise SystemExit("no Cloudflare accounts")
+    return rows[0]["id"]
+
+
+def fetch_pages_deploys(day: date) -> list[PagesDeploy]:
+    aid = cf_account_id()
+    d = http_json(
+        f"https://api.cloudflare.com/client/v4/accounts/{aid}/pages/projects/"
+        f"{PAGES_PROJECT}/deployments",
+        headers=cf_headers(),
+    )
+    assert isinstance(d, dict)
+    start_utc, end_utc = day_window_utc(day)
+    out: list[PagesDeploy] = []
+    for x in d.get("result") or []:
+        created = parse_gitea_time(x.get("created_on"))
+        if created is None:
+            continue
+        created_utc = created.astimezone(timezone.utc)
+        if not (start_utc <= created_utc < end_utc):
+            continue
+        if (x.get("environment") or "") != "production":
+            continue
+        stage = x.get("latest_stage") or {}
+        trig = x.get("deployment_trigger") or {}
+        meta = trig.get("metadata") or {}
+        commit = str(meta.get("commit_hash") or "")[:10]
+        out.append(
+            PagesDeploy(
+                created_on=created.astimezone(CST),
+                status=str(stage.get("status") or "?"),
+                commit=commit,
+                url=str(x.get("url") or ""),
+            )
+        )
+    out.sort(key=lambda p: p.created_on, reverse=True)
+    return out
+
+
 def fmt_bytes(n: int) -> str:
     x = float(n)
     for unit in ("B", "KB", "MB", "GB"):
@@ -239,39 +493,110 @@ def fmt_bytes(n: int) -> str:
     return f"{x:.1f} TB"
 
 
-def build_card(st: DayStats) -> dict:
+def _fmt_workflow_line(w: WorkflowDayStatus) -> str:
+    if w.note and not w.runs:
+        return f"- **{w.label}**（`{w.file}`）：{w.note}"
+    parts: list[str] = []
+    if w.ok_count:
+        parts.append(f"成功 {w.ok_count}")
+    if w.fail_count:
+        parts.append(f"失败 {w.fail_count}")
+    if w.running_count:
+        parts.append(f"进行中 {w.running_count}")
+    if not parts:
+        parts.append(w.note or "无结论")
+    # Show latest conclusion time if any
+    latest = ""
+    if w.runs:
+        r0 = w.runs[0]
+        t = r0.get("completed_at") or r0.get("started_at") or ""
+        if t:
+            dt = parse_gitea_time(t)
+            if dt:
+                latest = " · 最近 " + dt.astimezone(CST).strftime("%H:%M")
+    return f"- **{w.label}**：{' / '.join(parts)}{latest}"
+
+
+def build_card(st: DayStats, extras: DigestExtras) -> dict:
     """Feishu interactive card (msg_type=interactive)."""
     countries = "、".join(
         f"{name} {vis}访/{req}次" for name, req, vis in st.top_countries[:5]
     ) or "—"
     status_line = "、".join(f"{code}×{n}" for code, n in st.status[:6]) or "—"
-    now = datetime.now(timezone(timedelta(hours=8))).strftime("%Y-%m-%d %H:%M CST")
+    now = datetime.now(CST).strftime("%Y-%m-%d %H:%M CST")
 
-    # Free plan: no BotScore. Human ≈ known-browser homepage visits.
     estimate = st.human_home_vis
     rum = st.rum_human
 
+    traffic_md = (
+        f"**真实访问（估）**：约 **{estimate}** 次首页打开\n"
+        f"（已知浏览器 UA 打开 `/`；SPA Hash 路由不计路径；不含爬虫）\n\n"
+        f"**RUM 页面浏览**：{rum}"
+        + (f" + {st.rum_other} 其他" if st.rum_other else "")
+        + "\n"
+        f"**边缘总量**：{st.all_req} 请求 / {st.all_vis} visits / {fmt_bytes(st.bytes)}\n"
+        f"**非人类首页命中**：{st.other_home_vis} visits"
+    )
+
+    wf_lines = [_fmt_workflow_line(w) for w in extras.workflows]
+    if extras.gitea_note:
+        wf_lines.append(f"- _{extras.gitea_note}_")
+
+    ok_deploys = [p for p in extras.deploys if p.status == "success"]
+    fail_deploys = [p for p in extras.deploys if p.status != "success"]
+    if extras.deploys:
+        deploy_md = (
+            f"**Pages 生产部署**：成功 **{len(ok_deploys)}**"
+            + (f" / 异常 {len(fail_deploys)}" if fail_deploys else "")
+            + "\n"
+        )
+        for p in extras.deploys[:5]:
+            mark = "[ok]" if p.status == "success" else "[!]"
+            deploy_md += (
+                f"- {mark} {p.created_on.strftime('%H:%M')} "
+                f"`{p.commit or '?'}` {p.status}\n"
+            )
+    else:
+        deploy_md = "**Pages 生产部署**：当日无新部署（数据未变则 refresh 会跳过 deploy）\n"
+
+    refresh_md = (
+        f"**站点版本**：`v{extras.site_version}`\n\n"
+        f"**自动刷新（Gitea Actions）**\n"
+        + "\n".join(wf_lines)
+        + "\n\n"
+        + deploy_md.strip()
+    )
+
+    any_refresh_ok = any(w.ok_count > 0 for w in extras.workflows) or bool(ok_deploys)
+    any_refresh_fail = any(w.fail_count > 0 for w in extras.workflows) or bool(
+        fail_deploys
+    )
+
+    if any_refresh_fail:
+        header_color = "red"
+    elif estimate > 0 or any_refresh_ok:
+        header_color = "blue"
+    else:
+        header_color = "grey"
+
     elements: list[dict] = [
         {
             "tag": "div",
-            "text": {
-                "tag": "lark_md",
-                "content": (
-                    f"**真实访问（估）**：约 **{estimate}** 次首页打开\n"
-                    f"（已知浏览器 UA 打开 `/`；不含 Unknown/Headless/爬虫）\n\n"
-                    f"**RUM 页面浏览**：{rum}（人类浏览器执行 JS）"
-                    + (f" + {st.rum_other} 其他" if st.rum_other else "")
-                    + "\n"
-                    f"**边缘总量**：{st.all_req} 请求 / {st.all_vis} visits / {fmt_bytes(st.bytes)}\n"
-                    f"**非人类首页命中**：{st.other_home_vis} visits"
-                ),
-            },
+            "text": {"tag": "lark_md", "content": "**访问**\n" + traffic_md},
+        },
+        {
+            "tag": "div",
+            "text": {"tag": "lark_md", "content": "**自动更新**\n" + refresh_md},
         },
         {
             "tag": "div",
             "text": {
                 "tag": "lark_md",
-                "content": f"**首页地区 Top**：{countries}\n**状态码**：{status_line}",
+                "content": (
+                    f"**首页地区 Top**：{countries}\n"
+                    f"**状态码**：{status_line}\n"
+                    f"_模块：英雄 / 排行 / 主播 / 走势 / 物品 / 版本 / 机制_"
+                ),
             },
         },
         {
@@ -280,7 +605,8 @@ def build_card(st: DayStats) -> dict:
                 {
                     "tag": "plain_text",
                     "content": (
-                        f"{HOST} · Free 档无 BotScore，按 UA+路径估算 · {now}"
+                        f"{HOST} · Free 档无 BotScore · "
+                        f"refresh=web-daily/weekly/patch · {now}"
                     ),
                 }
             ],
@@ -293,7 +619,13 @@ def build_card(st: DayStats) -> dict:
                     "text": {"tag": "plain_text", "content": "打开站点"},
                     "type": "primary",
                     "url": SITE_URL,
-                }
+                },
+                {
+                    "tag": "button",
+                    "text": {"tag": "plain_text", "content": "Gitea Actions"},
+                    "type": "default",
+                    "url": f"{gitea_base()}/{GITEA_OWNER}/{GITEA_REPO}/actions",
+                },
             ],
         },
     ]
@@ -304,9 +636,9 @@ def build_card(st: DayStats) -> dict:
             "header": {
                 "title": {
                     "tag": "plain_text",
-                    "content": f"上分帝 Web 日活 · {st.day.isoformat()}",
+                    "content": f"上分帝 Web 日报 · {st.day.isoformat()}",
                 },
-                "template": "blue" if estimate > 0 else "grey",
+                "template": header_color,
             },
             "elements": elements,
         },
@@ -315,19 +647,19 @@ def build_card(st: DayStats) -> dict:
 
 def post_feishu(payload: dict) -> dict:
     url = feishu_url()
-    return http_json(
+    resp = http_json(
         url,
         method="POST",
         body=payload,
         headers={"Content-Type": "application/json"},
     )
+    assert isinstance(resp, dict)
+    return resp
 
 
 def parse_day(s: str | None) -> date:
     if not s:
-        # Yesterday in China time (site audience).
-        cst = timezone(timedelta(hours=8))
-        return (datetime.now(cst) - timedelta(days=1)).date()
+        return (datetime.now(CST) - timedelta(days=1)).date()
     return date.fromisoformat(s)
 
 
@@ -344,11 +676,23 @@ def main(argv: list[str] | None = None) -> int:
     day = parse_day(args.day)
     zid = zone_id()
     st = fetch_day(zid, day)
-    payload = build_card(st)
+    workflows, gitea_note = fetch_workflow_statuses(day)
+    deploys = fetch_pages_deploys(day)
+    extras = DigestExtras(
+        site_version=read_site_version(),
+        workflows=workflows,
+        deploys=deploys,
+        gitea_note=gitea_note,
+    )
+    payload = build_card(st, extras)
 
     print(
         f"{HOST} {day}: human_home_vis={st.human_home_vis} "
-        f"rum_human={st.rum_human} all_req={st.all_req}",
+        f"rum_human={st.rum_human} all_req={st.all_req} "
+        f"deploys={len(deploys)} "
+        f"wf_ok={sum(w.ok_count for w in workflows)} "
+        f"wf_fail={sum(w.fail_count for w in workflows)} "
+        f"v={extras.site_version}",
         flush=True,
     )
 
@@ -357,10 +701,13 @@ def main(argv: list[str] | None = None) -> int:
         return 0
 
     resp = post_feishu(payload)
-    # Feishu returns {code:0, msg:success} on OK
     code = resp.get("code", resp.get("StatusCode"))
     if code not in (0, None) and code != 200:
-        print("Feishu error:", json.dumps(resp, ensure_ascii=False)[:800], file=sys.stderr)
+        print(
+            "Feishu error:",
+            json.dumps(resp, ensure_ascii=False)[:800],
+            file=sys.stderr,
+        )
         return 1
     print("feishu ok:", json.dumps(resp, ensure_ascii=False)[:200])
     return 0