@lyra_wang/chrome-frequent-pages 0.4.3 → 0.5.2

package/README.md

@@ -144,3 +144,46 @@ python3 main.py --bookmarks-from-table ~/Desktop/常用网页书签.html
 ```
 
 读取飞书表每个唯一「域名」，生成 `https://<域名>/` 书签 HTML（Chrome 书签管理器导入）。
+
+---
+
+## 开机触发（每周一次）+ 投递到收件箱
+
+数据在本机 Chrome 里，所以执行必须在 Mac 上。定时不用时钟，用**开机作为触发点**：
+
+- `run-if-due.sh`：launchd 的 `RunAtLoad` 调用；读状态文件判断本周是否已跑，
+  没跑则执行 `chrome-frequent-pages --write`，跑完写状态（失败不写，下次开机重试）。
+- `inbox-deliver-mac.py`：把结果写进飞书「任务收件箱」，DSH 收件箱面板会显示为「待读」。
+- `table_snapshot.py`：执行前后各取一次「常用网页」表快照，diff 出本周**新增了哪些页面**，
+  摘要里带上「标题（域名）→ 书签「分类」」；未分类的页面导出书签时归入「其他」。
+
+### 安装
+
+```bash
+npm i -g @lyra_wang/chrome-frequent-pages@0.5.2   # 或 npm i -g file:$PWD
+bash "$(dirname "$(readlink -f "$(command -v chrome-frequent-pages)")")/../install.sh"
+```
+
+`install.sh` 会把 `main.py` / `run-if-due.sh` / `inbox-deliver-mac.py` / `table_snapshot.py`
+复制到 `~/.local/share/chrome-frequent-pages/`，并注册 `RunAtLoad` 的 LaunchAgent（无定时时刻）。
+
+### 必需的 macOS 权限（否则读不到 Chrome 历史）
+
+「完全磁盘访问权限」里需要加入三项（后台进程按二进制逐个鉴权）：
+
+| 加入项 | 作用 |
+|---|---|
+| `/bin/bash` | 启动脚本 |
+| `/usr/local/bin/node` | npm CLI 需要 |
+| `/usr/bin/python3` | 实际读 History 的进程 |
+
+### 行为
+
+| 场景 | 结果 |
+|---|---|
+| 本周第一次开机/登录 | 执行一次 |
+| 同周再次开机 | 状态文件命中，跳过 |
+| 执行失败 | 不写状态，下次开机重试；失败结果也会写收件箱 |
+
+状态文件：`~/.config/chrome-frequent-pages/last-run.json`（`{"week":"2026-43","at":"..."}`）
+日志：`~/.config/chrome-frequent-pages/run-if-due.log`

package/config.example.json

@@ -9,8 +9,8 @@
     "localhost",
     "127.0.0.1"
   ],
-  "base_token": "ZNjDb9TrpalLuCsw0macCeVdnld",
-  "table_id": "tblDejemWsgqKSA2",
+  "base_token": "<feishu-base-token>",
+  "table_id": "<table-id>",
   "identity": "user",
   "rules": {
     "drop_hosts": [

package/inbox-deliver-mac.py

@@ -0,0 +1,66 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""把一条任务结果写进飞书「任务收件箱」（Mac 侧，使用本机 lark-cli 用户身份）。
+
+用法:
+  inbox-deliver-mac.py --task chrome-frequent-pages --result 成功 --summary "本周新增 3 条"
+"""
+import argparse, json, subprocess, sys
+from datetime import datetime
+
+BASE = "NeujbbOfcaPEvMsPIxscaJImn9d"
+INBOX = "tbl9LghPBTjFLH4N"
+TASKS = "tbl4k7ggY1paqoUn"
+LARK = "lark-cli"
+
+def lark(args):
+    cmd = [LARK] + args + ["--as", "user", "--format", "json"]
+    res = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
+    try:
+        payload = json.loads(res.stdout or "{}")
+    except Exception:
+        payload = {}
+    if payload.get("ok") is False:
+        raise SystemExit("lark-cli failed: " + json.dumps(payload.get("error", payload), ensure_ascii=False))
+    if res.returncode != 0:
+        raise SystemExit("lark-cli exited %s: %s" % (res.returncode, (res.stderr or "")[:300]))
+    return payload
+
+def task_record_id(name):
+    payload = lark(["base", "+record-list", "--base-token", BASE, "--table-id", TASKS, "--limit", "200"])
+    meta = payload.get("data") or {}
+    fields = meta.get("fields") or []
+    rows = meta.get("data") or []
+    ids = meta.get("record_id_list") or []
+    if "任务名称" not in fields:
+        return ""
+    idx = fields.index("任务名称")
+    for rid, row in zip(ids, rows):
+        if str(row[idx]).strip() == name:
+            return str(rid)
+    return ""
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument("--task", required=True)
+    ap.add_argument("--result", required=True, choices=["成功", "失败", "超时"])
+    ap.add_argument("--summary", required=True)
+    args = ap.parse_args()
+    summary = args.summary.strip()
+    if len(summary) > 300:
+        summary = summary[:299] + "…"
+    record = {
+        "结果摘要": summary,
+        "结果": [args.result],
+        "状态": ["未读"],
+        "执行任务": int(datetime.now().timestamp() * 1000),
+    }
+    rid = task_record_id(args.task)
+    if rid:
+        record["关联任务"] = [rid]
+    lark(["base", "+record-batch-create", "--base-token", BASE, "--table-id", INBOX,
+          "--json", json.dumps({"create_records": [record]}, ensure_ascii=False)])
+    print(json.dumps({"ok": True, "task": args.task, "linked": bool(rid)}, ensure_ascii=False))
+
+if __name__ == "__main__":
+    main()

package/install.sh

@@ -3,13 +3,15 @@ set -eu
 HERE="$(cd "$(dirname "$0")" && pwd)"
 BIN_DIR="$HOME/.local/bin"
 DATA_DIR="$HOME/.local/share/chrome-frequent-pages"
-mkdir -p "$BIN_DIR" "$DATA_DIR" "$HOME/.config/chrome-frequent-pages"
-# launchd 读不了 ~/Documents（TCC），所以把脚本复制到不受保护的目录再跑
-cp "$HERE/main.py" "$DATA_DIR/main.py"
-chmod +x "$DATA_DIR/main.py"
+CONF_DIR="$HOME/.config/chrome-frequent-pages"
+mkdir -p "$BIN_DIR" "$DATA_DIR" "$CONF_DIR"
+
+# launchd 读不了 ~/Documents（TCC），脚本要放到不受保护的位置
+cp "$HERE/main.py" "$HERE/run-if-due.sh" "$HERE/inbox-deliver-mac.py" "$HERE/table_snapshot.py" "$DATA_DIR/"
+chmod +x "$DATA_DIR/main.py" "$DATA_DIR/run-if-due.sh" "$DATA_DIR/inbox-deliver-mac.py" "$DATA_DIR/table_snapshot.py"
 ln -sf "$DATA_DIR/main.py" "$BIN_DIR/chrome-frequent-pages"
-chmod +x "$HERE/main.py"
-[ -f "$HOME/.config/chrome-frequent-pages/config.json" ] || cp "$HERE/config.example.json" "$HOME/.config/chrome-frequent-pages/config.json"
+[ -f "$CONF_DIR/config.json" ] || cp "$HERE/config.example.json" "$CONF_DIR/config.json"
+
 PLIST="$HOME/Library/LaunchAgents/com.lyrawang.chrome-frequent-pages.plist"
 cat > "$PLIST" <<EOF
 <?xml version="1.0" encoding="UTF-8"?>
@@ -17,14 +19,21 @@ cat > "$PLIST" <<EOF
 <plist version="1.0"><dict>
   <key>Label</key><string>com.lyrawang.chrome-frequent-pages</string>
   <key>ProgramArguments</key><array>
-    <string>/usr/bin/python3</string><string>$DATA_DIR/main.py</string><string>--write</string>
+    <string>/bin/bash</string><string>$DATA_DIR/run-if-due.sh</string>
   </array>
-  <key>StartCalendarInterval</key><dict><key>Weekday</key><integer>1</integer><key>Hour</key><integer>9</integer><key>Minute</key><integer>30</integer></dict>
-  <key>StandardOutPath</key><string>$HOME/.config/chrome-frequent-pages/run.log</string>
-  <key>StandardErrorPath</key><string>$HOME/.config/chrome-frequent-pages/run.log</string>
+  <key>RunAtLoad</key><true/>
+  <key>StandardOutPath</key><string>$CONF_DIR/launchd.out.log</string>
+  <key>StandardErrorPath</key><string>$CONF_DIR/launchd.err.log</string>
+  <key>ProcessType</key><string>Background</string>
 </dict></plist>
 EOF
 launchctl bootout "gui/$(id -u)/com.lyrawang.chrome-frequent-pages" 2>/dev/null || true
 launchctl bootstrap "gui/$(id -u)" "$PLIST" 2>/dev/null || launchctl load "$PLIST" 2>/dev/null || true
 echo "✓ 已安装：~/.local/bin/chrome-frequent-pages"
-echo "✓ 已注册每周一 09:30 定时任务（launchd）"
+echo "✓ 已注册开机触发（launchd RunAtLoad，每周首次开机跑一次）"
+cat <<'NOTE'
+
+⚠️ 首次安装需在「系统设置 → 隐私与安全性 → 完全磁盘访问权限」中加入：
+   /bin/bash、/usr/local/bin/node、/usr/bin/python3
+   否则读不到 Chrome History（会写一条「失败」摘要到收件箱）。
+NOTE

package/package.json

@@ -1,17 +1,40 @@
 {
   "name": "@lyra_wang/chrome-frequent-pages",
-  "version": "0.4.3",
+  "version": "0.5.2",
   "description": "Chrome 高频网页 → 飞书「常用网页」表 + 按分类导出 Chrome 书签（根域名 / 10 天 / >5 次）",
   "license": "MIT",
   "type": "module",
-  "keywords": ["chrome", "history", "bookmarks", "feishu", "lark", "cli"],
+  "keywords": [
+    "chrome",
+    "history",
+    "bookmarks",
+    "feishu",
+    "lark",
+    "cli"
+  ],
   "repository": {
     "type": "git",
     "url": "git+https://github.com/LyraWang6688/ai-capability-store.git",
     "directory": "external-capabilities/cli/chrome-frequent-pages"
   },
-  "bin": { "chrome-frequent-pages": "bin/cli.js" },
-  "files": ["main.py", "bin/", "config.example.json", "README.md", "LICENSE", "install.sh"],
-  "publishConfig": { "access": "public" },
-  "engines": { "node": ">=20" }
+  "bin": {
+    "chrome-frequent-pages": "bin/cli.js"
+  },
+  "files": [
+    "main.py",
+    "run-if-due.sh",
+    "inbox-deliver-mac.py",
+    "table_snapshot.py",
+    "bin/",
+    "config.example.json",
+    "README.md",
+    "LICENSE",
+    "install.sh"
+  ],
+  "publishConfig": {
+    "access": "public"
+  },
+  "engines": {
+    "node": ">=20"
+  }
 }

package/run-if-due.sh

@@ -0,0 +1,58 @@
+#!/bin/bash
+# 开机触发：每周只跑一次（ISO 周去重）。由 launchd 的 RunAtLoad 调用。
+set -u
+STATE_DIR="$HOME/.config/chrome-frequent-pages"
+STATE="$STATE_DIR/last-run.json"
+LOG="$STATE_DIR/run-if-due.log"
+HERE="$(cd "$(dirname "$0")" && pwd)"
+export PATH="$HOME/.local/bin:/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
+WEEK="$(date +%G-%V)"
+mkdir -p "$STATE_DIR"
+log() { echo "$(date '+%F %T') $*" >> "$LOG"; }
+
+LAST=""
+if [ -f "$STATE" ]; then
+  LAST="$(python3 -c 'import json,sys;print(json.load(open(sys.argv[1])).get("week",""))' "$STATE" 2>/dev/null || true)"
+fi
+if [ "$LAST" = "$WEEK" ]; then
+  log "skip: week $WEEK already ran"
+  exit 0
+fi
+
+SNAP_BEFORE="$(python3 "$HERE/table_snapshot.py" 2>/dev/null || echo '{}')"
+OUT="$(chrome-frequent-pages --write 2>&1)"
+CODE=$?
+if [ $CODE -ne 0 ]; then
+  log "run failed (exit $CODE): $(echo "$OUT" | tail -3 | tr '\n' ' ')"
+  python3 "$HERE/inbox-deliver-mac.py" --task chrome-frequent-pages --result 失败 \
+    --summary "chrome-frequent-pages 执行失败：$(echo "$OUT" | tail -1 | cut -c1-70)" >> "$LOG" 2>&1 || true
+  exit 1
+fi
+
+LINE="$(echo "$OUT" | grep -o '飞书写入完成：.*' | tail -1)"
+[ -z "$LINE" ] && LINE="$(echo "$OUT" | tail -1)"
+SNAP_AFTER="$(python3 "$HERE/table_snapshot.py" 2>/dev/null || echo '{}')"
+DETAIL="$(python3 - "$SNAP_BEFORE" "$SNAP_AFTER" <<'PYD'
+import json, sys
+before = json.loads(sys.argv[1] or "{}")
+after = json.loads(sys.argv[2] or "{}")
+new = [(h, v) for h, v in after.items() if h not in before]
+if not new:
+    print("本周无新增页面")
+else:
+    parts = []
+    for host, v in new[:5]:
+        cat = v.get("cat") or "其他"
+        tag = "未分类→其他" if cat == "其他" else cat
+        parts.append(f"{v.get('title') or host}（{host}）→ 书签「{tag}」")
+    more = "" if len(new) <= 5 else f" 等 {len(new)} 条"
+    print("新增：" + "；".join(parts) + more)
+PYD
+)"
+python3 "$HERE/inbox-deliver-mac.py" --task chrome-frequent-pages --result 成功 \
+  --summary "$DETAIL；$LINE" >> "$LOG" 2>&1 || log "inbox deliver failed"
+python3 - "$STATE" "$WEEK" <<'PY'
+import json, sys, datetime
+json.dump({"week": sys.argv[2], "at": datetime.datetime.now().isoformat(timespec="seconds")}, open(sys.argv[1], "w"), ensure_ascii=False)
+PY
+log "done week=$WEEK result=$LINE"

package/table_snapshot.py

@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""导出「常用网页」表的 域名 -> {标题, 分类} 快照，供 run-if-due.sh 做前后 diff。"""
+import json, subprocess, sys
+
+BASE = "ZNjDb9TrpalLuCsw0macCeVdnld"
+TABLE = "tblDejemWsgqKSA2"
+
+def main():
+    out = subprocess.run(
+        ["lark-cli", "base", "+record-list", "--base-token", BASE, "--table-id", TABLE,
+         "--limit", "200", "--as", "user", "--format", "json"],
+        capture_output=True, text=True, timeout=120)
+    try:
+        data = json.loads(out.stdout)
+    except Exception:
+        print("{}")
+        return 1
+    meta = (data.get("data") or {})
+    cols = meta.get("fields") or []
+    rows = meta.get("data") or []
+    if "域名" not in cols:
+        print("{}")
+        return 1
+    idx = {name: i for i, name in enumerate(cols)}
+    snap = {}
+    for row in rows:
+        host = str(row[idx["域名"]] or "").strip()
+        if not host:
+            continue
+        cat = "其他"
+        if "分类" in idx:
+            raw = row[idx["分类"]]
+            if isinstance(raw, list) and raw:
+                cat = str(raw[0])
+            elif isinstance(raw, str) and raw.strip():
+                cat = raw.strip()
+        title = ""
+        if "标题" in idx and row[idx["标题"]]:
+            title = str(row[idx["标题"]]).strip()
+        snap[host] = {"title": title or host, "cat": cat}
+    print(json.dumps(snap, ensure_ascii=False))
+    return 0
+
+if __name__ == "__main__":
+    sys.exit(main())