refactor: Phase 1 抽出 shared/skills_data 解耦 daily 与 bot

daily 不再通过 sys.path 导入 bot/skills_service，skills feed 数据层上移到 shared/。

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

bot/env_config.py

@@ -3,11 +3,16 @@
 from __future__ import annotations
 
 import os
+import sys
 from pathlib import Path
 
 from dotenv import load_dotenv
 
 _BOT_DIR = Path(__file__).resolve().parent
+_REPO_ROOT = _BOT_DIR.parent
+if str(_REPO_ROOT) not in sys.path:
+    sys.path.insert(0, str(_REPO_ROOT))
+
 load_dotenv(_BOT_DIR / ".env")
 load_dotenv(_BOT_DIR / ".env.local", override=True)
 

bot/skills_service.py

@@ -1,32 +1,20 @@
-"""skills.sh 数据查询与命令解析。"""
+"""skills.sh 快查命令解析与企微回复格式化。"""
 
 from __future__ import annotations
 
-import json
-import logging
 import re
-import time
+import sys
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Any, Literal
 
-import certifi
+_REPO_ROOT = Path(__file__).resolve().parent.parent
-import httpx
+if str(_REPO_ROOT) not in sys.path:
+    sys.path.insert(0, str(_REPO_ROOT))
 
-logger = logging.getLogger(__name__)
+from shared.skills_data import Board, board_items, format_installs, load_feed, warm_feed_cache
 
-FEED_URLS = [
+__all__ = ["Command", "handle_command", "parse_command", "warm_feed_cache"]
-    # jsDelivr 在国内通常比 raw.githubusercontent.com 更稳定
-    "https://cdn.jsdelivr.net/gh/NeverSight/skills.sh_feed@main/data/feed.json",
-    "https://raw.githubusercontent.com/NeverSight/skills.sh_feed/main/data/feed.json",
-]
-CACHE_TTL_SECONDS = 600
-CACHE_DIR = Path(__file__).resolve().parent / ".cache"
-CACHE_FILE = CACHE_DIR / "feed.json"
-
-_cache: dict[str, Any] = {"data": None, "fetched_at": 0.0}
-
-Board = Literal["trending", "hot", "all"]
 
 
 @dataclass
@@ -37,72 +25,6 @@ class Command:
     query: str = ""
 
 
-def _fetch_json(url: str) -> dict[str, Any]:
-    headers = {
-        "User-Agent": "skills-hot-bot/1.0",
-        "Accept": "application/json",
-    }
-    with httpx.Client(
-        timeout=httpx.Timeout(20.0, connect=10.0),
-        verify=certifi.where(),
-        follow_redirects=True,
-    ) as client:
-        resp = client.get(url, headers=headers)
-        resp.raise_for_status()
-        return resp.json()
-
-
-def _load_disk_cache() -> dict[str, Any] | None:
-    if not CACHE_FILE.exists():
-        return None
-    try:
-        return json.loads(CACHE_FILE.read_text(encoding="utf-8"))
-    except (OSError, json.JSONDecodeError) as exc:
-        logger.warning("读取本地缓存失败: %s", exc)
-        return None
-
-
-def _save_disk_cache(data: dict[str, Any]) -> None:
-    CACHE_DIR.mkdir(parents=True, exist_ok=True)
-    CACHE_FILE.write_text(json.dumps(data, ensure_ascii=False), encoding="utf-8")
-
-
-def load_feed(force: bool = False) -> dict[str, Any]:
-    now = time.time()
-    if not force and _cache["data"] and now - _cache["fetched_at"] < CACHE_TTL_SECONDS:
-        return _cache["data"]
-
-    errors: list[str] = []
-    for url in FEED_URLS:
-        for attempt in range(3):
-            try:
-                data = _fetch_json(url)
-                _cache["data"] = data
-                _cache["fetched_at"] = now
-                _save_disk_cache(data)
-                logger.info("skills 数据已更新: %s", url)
-                return data
-            except Exception as exc:
-                msg = f"{url} (#{attempt + 1}): {exc}"
-                errors.append(msg)
-                logger.debug("拉取失败 %s", msg)
-                time.sleep(0.5 * (attempt + 1))
-
-    stale = _load_disk_cache()
-    if stale:
-        logger.warning("网络不可用，回退到本地缓存")
-        _cache["data"] = stale
-        _cache["fetched_at"] = now
-        return stale
-
-    raise RuntimeError(f"无法获取 skills 数据。最近错误: {errors[-1]}")
-
-
-def warm_feed_cache() -> None:
-    """启动时预加载，避免首条消息才触发网络请求。"""
-    load_feed(force=True)
-
-
 def _normalize_text(text: str) -> str:
     text = re.sub(r"@\S+\s*", "", text)
     return text.strip().lower()
@@ -157,19 +79,6 @@ def parse_command(text: str) -> Command:
     return Command(kind="search", query=raw, limit=5)
 
 
-def _format_installs(n: int | float) -> str:
-    if n >= 1_000_000:
-        return f"{n / 1_000_000:.1f}M"
-    if n >= 1_000:
-        return f"{n / 1_000:.1f}K"
-    return str(int(n))
-
-
-def _board_items(feed: dict[str, Any], board: Board) -> list[dict[str, Any]]:
-    key = {"trending": "topTrending", "hot": "topHot", "all": "topAllTime"}[board]
-    return feed.get(key, [])
-
-
 def _board_title(board: Board) -> str:
     return {
         "trending": "Trending（近期增长）",
@@ -180,7 +89,7 @@ def _board_title(board: Board) -> str:
 
 def format_list(board: Board, limit: int) -> str:
     feed = load_feed()
-    items = _board_items(feed, board)[:limit]
+    items = board_items(feed, board)[:limit]
     updated = feed.get("updatedAt", "未知")[:10]
 
     lines = [
@@ -192,7 +101,7 @@ def format_list(board: Board, limit: int) -> str:
     for i, item in enumerate(items, 1):
         title = item.get("title", "?")
         source = item.get("source", "?")
-        installs = _format_installs(item.get("installs", 0))
+        installs = format_installs(item.get("installs", 0))
         desc = item.get("description", "")
         if len(desc) > 80:
             desc = desc[:77] + "..."
@@ -241,7 +150,7 @@ def format_search(query: str, limit: int) -> str:
     for i, item in enumerate(matches, 1):
         title = item.get("title", "?")
         source = item.get("source", "?")
-        installs = _format_installs(item.get("installs", 0))
+        installs = format_installs(item.get("installs", 0))
         link = item.get("link", "")
         lines.append(f"{i}. **{title}** · {installs} · `{source}`")
         if link:
@@ -270,7 +179,7 @@ def format_detail(name: str) -> str:
         [
             f"**{best.get('title', '?')}**",
             f"`{best.get('source', '?')}`",
-            f"安装量：**{_format_installs(best.get('installs', 0))}**",
+            f"安装量：**{format_installs(best.get('installs', 0))}**",
             "",
             desc,
             "",

daily/config.py

@@ -3,13 +3,11 @@
 from __future__ import annotations
 
 import os
-import sys
 from pathlib import Path
 
 from dotenv import load_dotenv
 
 ROOT = Path(__file__).resolve().parent.parent
-BOT_DIR = ROOT / "bot"
 OUTPUT_DIR = ROOT / "output"
 LOG_DIR = ROOT / "logs"
 CACHE_DIR = ROOT / ".cache"
@@ -49,12 +47,6 @@ load_dotenv(ROOT / ".env")
 load_dotenv(ROOT / ".env.local", override=True)
 
 
-def ensure_bot_on_path() -> None:
-    bot = str(BOT_DIR)
-    if bot not in sys.path:
-        sys.path.insert(0, bot)
-
-
 def _clean_env_value(raw: str | None) -> str | None:
     if raw is None:
         return None

daily/generate.py

@@ -22,7 +22,6 @@ from daily.config import (
     LOG_DIR,
     OUTPUT_DIR,
     SNAPSHOT_FILE,
-    ensure_bot_on_path,
     env,
     env_int,
     full_desc_limit,
@@ -58,9 +57,7 @@ from daily.report_data import (
 )
 from daily.skills_board import load_boards
 from daily.skills_group import group_skills_by_source
-
+from shared.skills_data import format_installs, load_feed
-ensure_bot_on_path()
-from skills_service import _format_installs, load_feed  # noqa: E402
 
 THEME_RULES: list[tuple[str, str, list[str]]] = [
     ("🎬", "AI 多媒体 / 视频", ["runcomfy", "remotion", "video", "seedance", "inpaint", "lipsync"]),
@@ -232,7 +229,7 @@ def _prepare_skill_item(item: dict[str, Any], prev_ids: set[str], rank: int) ->
         badge = "🆕"
     elif rank == 1:
         badge = "👑"
-    installs_fmt = item.get("installs_fmt") or _format_installs(item.get("installs", 0))
+    installs_fmt = item.get("installs_fmt") or format_installs(item.get("installs", 0))
     title = item.get("source", "?") if item.get("cluster") else item.get("title", "?")
     desc = item.get("wecom_desc") or item.get("description") or item.get("cluster_titles") or ""
     limit = wecom_skill_desc_limit()
@@ -303,7 +300,7 @@ def _build_highlights(
             )
     if trending:
         t0 = trending[0]
-        points.append(f"📈 Skills 榜首 **{t0.get('title')}**（{_format_installs(t0.get('installs', 0))}）")
+        points.append(f"📈 Skills 榜首 **{t0.get('title')}**（{format_installs(t0.get('installs', 0))}）")
     if github_trending:
         g0 = github_trending[0]
         stars = g0.get("stars_today_fmt", "")
@@ -315,7 +312,7 @@ def _build_highlights(
         points.append(f"🌱 新兴 [{e0['repo']}]({e0['url']})（⭐ {e0.get('total_stars_fmt', '?')}）")
     elif hot:
         h0 = hot[0]
-        points.append(f"🔥 Skills Hot 榜首 **{h0.get('title')}**（1H {_format_installs(h0.get('installs', 0))}）")
+        points.append(f"🔥 Skills Hot 榜首 **{h0.get('title')}**（1H {format_installs(h0.get('installs', 0))}）")
     while len(points) < 3 and len(trending) > len(points):
         item = trending[len(points)]
         points.append(f"✨ **{item.get('title')}** · `{item.get('source')}`")
@@ -397,7 +394,7 @@ def _format_skill_section(items: list[dict[str, Any]], *, hot: bool = False) ->
     for i, item in enumerate(items, 1):
         skill_id = item.get("id") or f"{item.get('source', '?')}/{item.get('title', '?')}"
         link = item.get("link", "")
-        installs = _format_installs(item.get("installs", 0))
+        installs = format_installs(item.get("installs", 0))
         meta = f"1H {installs}" if hot else f"总安装 {installs}"
         if link:
             lines.append(f"{i}. **[{skill_id}]({link})** · {meta}")

daily/llm_client.py

@@ -74,11 +74,14 @@ def _cursor_chat(system: str, user: str) -> str:
     api_key = (env("CURSOR_API_KEY") or "").strip()
     if not api_key:
         return ""
+    import sys
+
+    from daily.config import ROOT
     from cursor_sdk import Agent, AgentOptions, CursorAgentError, LocalAgentOptions
 
-    from daily.config import ensure_bot_on_path
+    _bot = str(ROOT / "bot")
-
+    if _bot not in sys.path:
-    ensure_bot_on_path()
+        sys.path.insert(0, _bot)
     try:
         from bridge_manager import warm_cursor_bridge
     except ImportError:

shared/__init__.py

@@ -0,0 +1,10 @@
+"""Repo-root shared modules (skills feed data, etc.)."""
+
+from shared.skills_data import (
+    Board,
+    format_installs,
+    load_feed,
+    warm_feed_cache,
+)
+
+__all__ = ["Board", "format_installs", "load_feed", "warm_feed_cache"]

shared/skills_data.py

@@ -0,0 +1,109 @@
+"""skills.sh feed.json 拉取、缓存与榜单数据。"""
+
+from __future__ import annotations
+
+import json
+import logging
+import time
+from pathlib import Path
+from typing import Any, Literal
+
+import certifi
+import httpx
+
+logger = logging.getLogger(__name__)
+
+ROOT = Path(__file__).resolve().parent.parent
+FEED_URLS = [
+    "https://cdn.jsdelivr.net/gh/NeverSight/skills.sh_feed@main/data/feed.json",
+    "https://raw.githubusercontent.com/NeverSight/skills.sh_feed/main/data/feed.json",
+]
+CACHE_TTL_SECONDS = 600
+CACHE_DIR = ROOT / ".cache"
+CACHE_FILE = CACHE_DIR / "skills-feed.json"
+USER_AGENT = "Mozilla/5.0 (compatible; daily-robots/1.0; +https://skills.sh)"
+
+_cache: dict[str, Any] = {"data": None, "fetched_at": 0.0}
+
+Board = Literal["trending", "hot", "all"]
+
+BOARD_FEED_KEYS = {
+    "trending": "topTrending",
+    "hot": "topHot",
+    "all": "topAllTime",
+}
+
+
+def _fetch_json(url: str) -> dict[str, Any]:
+    headers = {"User-Agent": USER_AGENT, "Accept": "application/json"}
+    with httpx.Client(
+        timeout=httpx.Timeout(20.0, connect=10.0),
+        verify=certifi.where(),
+        follow_redirects=True,
+    ) as client:
+        resp = client.get(url, headers=headers)
+        resp.raise_for_status()
+        return resp.json()
+
+
+def _load_disk_cache() -> dict[str, Any] | None:
+    if not CACHE_FILE.exists():
+        return None
+    try:
+        return json.loads(CACHE_FILE.read_text(encoding="utf-8"))
+    except (OSError, json.JSONDecodeError) as exc:
+        logger.warning("读取 skills feed 本地缓存失败: %s", exc)
+        return None
+
+
+def _save_disk_cache(data: dict[str, Any]) -> None:
+    CACHE_DIR.mkdir(parents=True, exist_ok=True)
+    CACHE_FILE.write_text(json.dumps(data, ensure_ascii=False), encoding="utf-8")
+
+
+def load_feed(force: bool = False) -> dict[str, Any]:
+    now = time.time()
+    if not force and _cache["data"] and now - _cache["fetched_at"] < CACHE_TTL_SECONDS:
+        return _cache["data"]
+
+    errors: list[str] = []
+    for url in FEED_URLS:
+        for attempt in range(3):
+            try:
+                data = _fetch_json(url)
+                _cache["data"] = data
+                _cache["fetched_at"] = now
+                _save_disk_cache(data)
+                logger.info("skills 数据已更新: %s", url)
+                return data
+            except Exception as exc:
+                msg = f"{url} (#{attempt + 1}): {exc}"
+                errors.append(msg)
+                logger.debug("拉取失败 %s", msg)
+                time.sleep(0.5 * (attempt + 1))
+
+    stale = _load_disk_cache()
+    if stale:
+        logger.warning("网络不可用，回退到本地 skills feed 缓存")
+        _cache["data"] = stale
+        _cache["fetched_at"] = now
+        return stale
+
+    raise RuntimeError(f"无法获取 skills 数据。最近错误: {errors[-1]}")
+
+
+def warm_feed_cache() -> None:
+    """启动时预加载，避免首条消息才触发网络请求。"""
+    load_feed(force=True)
+
+
+def format_installs(n: int | float) -> str:
+    if n >= 1_000_000:
+        return f"{n / 1_000_000:.1f}M"
+    if n >= 1_000:
+        return f"{n / 1_000:.1f}K"
+    return str(int(n))
+
+
+def board_items(feed: dict[str, Any], board: Board) -> list[dict[str, Any]]:
+    return feed.get(BOARD_FEED_KEYS[board], [])