Rename to hkt.sh

projects/news-bot/README.md

@@ -0,0 +1,146 @@
+# 📰 财经快讯 Telegram 机器人
+
+自动抓取金十、华尔街见闻、36氪、新浪财经快讯，AI 评分过滤，推送到 Telegram。
+
+---
+
+## 功能特性
+
+- 🔴 **重磅即时推送** — 评分 ≥8 立即发送
+- 🟡 **普通汇总推送** — 评分 6-7，每 N 分钟汇总一条
+- 📊 **定时总结** — 08:00 / 11:30 / 20:00 三次 AI 总结
+- 🔑 **关键词过滤** — 自定义关注词，精准推送
+- 📡 **多源管理** — 一键开关各信息源
+- ⚙️ **灵活设置** — 推送阈值、汇总频率均可调节
+- 🔄 **去重** — 标题相似度 >70% 自动合并
+
+---
+
+## 环境变量
+
+| 变量 | 说明 | 必填 |
+|------|------|------|
+| `BOT_TOKEN` | Telegram Bot Token（从 @BotFather 获取） | ✅ |
+| `ADMIN_ID` | 管理员 Telegram 用户 ID | ❌（默认 165067365） |
+| `DATA_DIR` | 数据目录 | ❌（默认 `/opt/news-bot/data`） |
+
+---
+
+## 快速部署
+
+### 方法一：直接运行
+
+```bash
+# 1. 进入项目目录
+cd /path/to/news-bot
+
+# 2. 安装依赖
+pip3 install -r requirements.txt
+
+# 3. 创建数据目录
+mkdir -p /opt/news-bot/data
+
+# 4. 启动
+export BOT_TOKEN="your_bot_token_here"
+export ADMIN_ID="165067365"
+python3 bot.py
+```
+
+### 方法二：systemd 服务（Linux）
+
+```ini
+# /etc/systemd/system/news-bot.service
+[Unit]
+Description=Telegram 财经快讯机器人
+After=network.target
+
+[Service]
+Type=simple
+WorkingDirectory=/opt/news-bot
+ExecStart=/usr/bin/python3 /opt/news-bot/bot.py
+Environment=BOT_TOKEN=your_bot_token_here
+Environment=ADMIN_ID=165067365
+Environment=DATA_DIR=/opt/news-bot/data
+Restart=always
+RestartSec=10
+StandardOutput=journal
+StandardError=journal
+
+[Install]
+WantedBy=multi-user.target
+```
+
+```bash
+sudo systemctl daemon-reload
+sudo systemctl enable news-bot
+sudo systemctl start news-bot
+sudo systemctl status news-bot
+```
+
+### 方法三：部署到服务器（38.76.204.161）
+
+```bash
+# 上传文件
+scp -r news-bot/ root@38.76.204.161:/opt/news-bot/
+
+# SSH 登录
+ssh root@38.76.204.161
+
+# 安装依赖
+pip3 install -r /opt/news-bot/requirements.txt
+
+# 创建数据目录
+mkdir -p /opt/news-bot/data
+
+# 配置 systemd（见上方），然后启动
+systemctl start news-bot
+```
+
+---
+
+## 文件结构
+
+```
+news-bot/
+├── bot.py          # 主入口：Telegram bot + 调度器
+├── sources.py      # 信息源抓取（金十/华尔街/36氪/新浪）
+├── scorer.py       # 评分引擎 + 去重
+├── storage.py      # JSON 数据持久化
+├── summarizer.py   # 新闻总结生成
+├── requirements.txt
+├── README.md
+└── data/           # 运行时数据（settings.json / news.json / pushed.json）
+```
+
+---
+
+## 命令菜单
+
+| 命令 | 功能 |
+|------|------|
+| `/start` | 欢迎页 + 主菜单按钮 |
+| `/news` | 最新快讯（分页浏览） |
+| `/summary` | 手动触发总结 |
+| `/sources` | 订阅源开关管理 |
+| `/keywords` | 关键词添加/删除 |
+| `/settings` | 推送阈值/频率设置 |
+
+---
+
+## 评分规则
+
+| 分数 | 含义 | 处理 |
+|------|------|------|
+| ≥ 8 | 重磅/突发 | 🔥 即时推送 |
+| 6–7 | 值得关注 | 🟡 汇总推送 |
+| < 6 | 普通资讯 | ❌ 不推送 |
+
+评分依据：关键词权重 + 来源重要性标记 + 规则引擎（无需外部 AI API）。
+
+---
+
+## 数据文件说明
+
+- `data/settings.json` — 用户配置（关键词、源开关、阈值）
+- `data/news.json` — 新闻历史（最近 2000 条）
+- `data/pushed.json` — 已推送 ID 集合（防重复，最近 5000 条）

projects/news-bot/bot.py

@@ -0,0 +1,776 @@
+"""
+Telegram 财经快讯机器人 - 主入口
+功能：自动抓取财经/科技快讯，AI 评分过滤，推送到 Telegram
+"""
+import os
+import time
+import logging
+import asyncio
+from datetime import datetime
+
+from telegram import Update, InlineKeyboardButton, InlineKeyboardMarkup
+from telegram.ext import (
+    Application, CommandHandler, CallbackQueryHandler,
+    ContextTypes, MessageHandler, filters,
+)
+from telegram.constants import ParseMode
+
+import sources
+import scorer
+import storage
+import summarizer
+
+# 日志配置
+logging.basicConfig(
+    format="%(asctime)s [%(name)s] %(levelname)s: %(message)s",
+    level=logging.INFO,
+)
+logger = logging.getLogger(__name__)
+
+# 配置
+BOT_TOKEN = os.environ.get("BOT_TOKEN", "")
+ADMIN_ID = int(os.environ.get("ADMIN_ID", "165067365"))
+
+# 分页状态缓存
+_page_cache = {}
+
+
+# ========== 工具函数 ==========
+
+def _check_admin(user_id: int) -> bool:
+    """检查是否为管理员"""
+    return user_id == ADMIN_ID
+
+
+def _format_news_item(item: dict) -> str:
+    """格式化单条新闻（支持 HTML 链接）"""
+    score = item.get("score", 0)
+    if score >= 9:
+        emoji = "🔥"
+    elif score >= 8:
+        emoji = "⚡"
+    elif score >= 7:
+        emoji = "📌"
+    elif score >= 6:
+        emoji = "🔹"
+    else:
+        emoji = "•"
+    src = item.get("source_name", "")
+    ts = item.get("time_str", "")
+    title = item.get("title", "")
+    url = item.get("url", "")
+    # 转义 HTML 特殊字符
+    import html as html_mod
+    title_safe = html_mod.escape(title)
+    if url:
+        title_display = f'<a href="{url}">{title_safe}</a>'
+    else:
+        title_display = title_safe
+    return f"{emoji} {title_display}\n   📡 {src}  🕐 {ts}  评分:{score}"
+
+
+# ========== /start 命令 ==========
+
+async def cmd_start(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """欢迎消息 + 主菜单"""
+    if not _check_admin(update.effective_user.id):
+        await update.message.reply_text("⛔ 仅限管理员使用")
+        return
+
+    keyboard = [
+        [InlineKeyboardButton("📰 最新快讯", callback_data="news_0"),
+         InlineKeyboardButton("📊 新闻总结", callback_data="summary_menu")],
+        [InlineKeyboardButton("📡 订阅源管理", callback_data="sources_menu"),
+         InlineKeyboardButton("🔑 关键词管理", callback_data="keywords_menu")],
+        [InlineKeyboardButton("⚙️ 设置", callback_data="settings_menu")],
+    ]
+    text = (
+        "👋 <b>欢迎使用财经快讯机器人</b>\n\n"
+        "🤖 自动抓取金十、华尔街见闻、36氪、新浪财经快讯\n"
+        "📊 AI 评分过滤，只推送相关内容\n"
+        "⏰ 定时总结，不错过重要新闻\n\n"
+        "请选择功能："
+    )
+    await update.message.reply_text(
+        text, parse_mode=ParseMode.HTML,
+        reply_markup=InlineKeyboardMarkup(keyboard),
+    )
+
+
+# ========== /news 命令 ==========
+
+async def cmd_news(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """查看最新快讯"""
+    if not _check_admin(update.effective_user.id):
+        return
+    await _show_news_page(update.message, 0)
+
+
+async def _show_news_page(message_or_query, page: int):
+    """显示新闻分页"""
+    settings = storage.get_settings()
+    news_list = storage.get_news(limit=100)
+
+    # 评分
+    scorer.score_and_filter(news_list, settings.get("keywords", []))
+    # 只显示评分 >= 4 的
+    news_list = [n for n in news_list if n.get("score", 0) >= 4]
+    news_list.sort(key=lambda x: x.get("timestamp", 0), reverse=True)
+
+    page_size = 10
+    total_pages = max(1, (len(news_list) + page_size - 1) // page_size)
+    page = max(0, min(page, total_pages - 1))
+    start = page * page_size
+    page_items = news_list[start:start + page_size]
+
+    if not page_items:
+        text = "📭 暂无快讯，等待抓取中..."
+    else:
+        lines = [f"📰 <b>最新快讯</b> (第{page+1}/{total_pages}页)\n"]
+        for item in page_items:
+            lines.append(_format_news_item(item))
+        text = "\n".join(lines)
+
+    # 分页按钮
+    buttons = []
+    if page > 0:
+        buttons.append(InlineKeyboardButton("⬅️ 上一页", callback_data=f"news_{page-1}"))
+    buttons.append(InlineKeyboardButton("🔄 刷新", callback_data=f"news_{page}"))
+    if page < total_pages - 1:
+        buttons.append(InlineKeyboardButton("➡️ 下一页", callback_data=f"news_{page+1}"))
+    keyboard = [buttons, [InlineKeyboardButton("🏠 主菜单", callback_data="main_menu")]]
+
+    if hasattr(message_or_query, "edit_message_text"):
+        try:
+            await message_or_query.edit_message_text(
+                text, parse_mode=ParseMode.HTML,
+                reply_markup=InlineKeyboardMarkup(keyboard),
+            )
+        except Exception:
+            await message_or_query.edit_message_text(
+                text, reply_markup=InlineKeyboardMarkup(keyboard),
+            )
+    else:
+        await message_or_query.reply_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+
+# ========== /summary 命令 ==========
+
+async def cmd_summary(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """手动触发新闻总结"""
+    if not _check_admin(update.effective_user.id):
+        return
+    await _show_summary_menu(update.message)
+
+
+async def _show_summary_menu(message_or_query):
+    """显示总结时间范围选择"""
+    keyboard = [
+        [InlineKeyboardButton("⏱ 最近1小时", callback_data="summary_最近1小时")],
+        [InlineKeyboardButton("🌅 上午", callback_data="summary_上午")],
+        [InlineKeyboardButton("📅 全天", callback_data="summary_全天")],
+        [InlineKeyboardButton("🏠 主菜单", callback_data="main_menu")],
+    ]
+    text = "📊 <b>新闻总结</b>\n\n请选择时间范围："
+    if hasattr(message_or_query, "edit_message_text"):
+        await message_or_query.edit_message_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+    else:
+        await message_or_query.reply_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+
+# ========== /sources 命令 ==========
+
+async def cmd_sources(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """查看/管理订阅源"""
+    if not _check_admin(update.effective_user.id):
+        return
+    await _show_sources_menu(update.message)
+
+
+async def _show_sources_menu(message_or_query):
+    """显示订阅源管理菜单"""
+    settings = storage.get_settings()
+    srcs = settings.get("sources", {})
+    source_info = {
+        "jin10":      "金十数据",
+        "wallstreet": "华尔街见闻",
+        "kr36":       "36氪",
+        "sina":       "新浪财经",
+    }
+    keyboard = []
+    for key, name in source_info.items():
+        enabled = srcs.get(key, True)
+        status = "✅" if enabled else "❌"
+        keyboard.append([InlineKeyboardButton(
+            f"{status} {name}", callback_data=f"toggle_src_{key}"
+        )])
+    keyboard.append([InlineKeyboardButton("🏠 主菜单", callback_data="main_menu")])
+    text = "📡 <b>订阅源管理</b>\n\n点击切换开/关："
+    if hasattr(message_or_query, "edit_message_text"):
+        await message_or_query.edit_message_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+    else:
+        await message_or_query.reply_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+
+# ========== /keywords 命令 ==========
+
+async def cmd_keywords(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """管理关注关键词"""
+    if not _check_admin(update.effective_user.id):
+        return
+    await _show_keywords_menu(update.message)
+
+
+async def _show_keywords_menu(message_or_query):
+    """显示关键词管理菜单"""
+    settings = storage.get_settings()
+    kw_list = settings.get("keywords", [])
+    # 每行显示3个关键词删除按钮
+    keyboard = []
+    row = []
+    for kw in kw_list[:30]:
+        row.append(InlineKeyboardButton(f"❌ {kw}", callback_data=f"del_kw_{kw}"))
+        if len(row) == 3:
+            keyboard.append(row)
+            row = []
+    if row:
+        keyboard.append(row)
+    keyboard.append([InlineKeyboardButton("➕ 添加关键词", callback_data="add_kw_prompt")])
+    keyboard.append([InlineKeyboardButton("🏠 主菜单", callback_data="main_menu")])
+    text = f"🔑 <b>关键词管理</b>\n\n当前 {len(kw_list)} 个关键词\n（点击关键词可删除）"
+    if hasattr(message_or_query, "edit_message_text"):
+        await message_or_query.edit_message_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+    else:
+        await message_or_query.reply_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+
+# ========== /settings 命令 ==========
+
+async def cmd_settings(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """设置菜单"""
+    if not _check_admin(update.effective_user.id):
+        return
+    await _show_settings_menu(update.message)
+
+
+async def _show_settings_menu(message_or_query):
+    """显示设置菜单"""
+    settings = storage.get_settings()
+    push_on = settings.get("push_enabled", True)
+    min_score = settings.get("min_score", 6)
+    instant_score = settings.get("instant_score", 8)
+    interval = settings.get("batch_interval", 30)
+
+    push_txt = "✅ 推送已开" if push_on else "❌ 推送已关"
+    keyboard = [
+        [InlineKeyboardButton(push_txt, callback_data="toggle_push")],
+        [
+            InlineKeyboardButton(f"最低评分: {min_score} ➖", callback_data="score_min_down"),
+            InlineKeyboardButton(f"最低评分: {min_score} ➕", callback_data="score_min_up"),
+        ],
+        [
+            InlineKeyboardButton(f"即时阈值: {instant_score} ➖", callback_data="score_instant_down"),
+            InlineKeyboardButton(f"即时阈值: {instant_score} ➕", callback_data="score_instant_up"),
+        ],
+        [
+            InlineKeyboardButton(f"汇总间隔: {interval}min ➖", callback_data="interval_down"),
+            InlineKeyboardButton(f"汇总间隔: {interval}min ➕", callback_data="interval_up"),
+        ],
+        [InlineKeyboardButton("🏠 主菜单", callback_data="main_menu")],
+    ]
+    text = (
+        f"⚙️ <b>设置</b>\n\n"
+        f"推送状态: {'开启' if push_on else '关闭'}\n"
+        f"最低推送评分: {min_score}\n"
+        f"即时推送阈值: {instant_score}\n"
+        f"汇总推送间隔: {interval} 分钟"
+    )
+    if hasattr(message_or_query, "edit_message_text"):
+        await message_or_query.edit_message_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+    else:
+        await message_or_query.reply_text(
+            text, parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+
+# ========== CallbackQuery 处理器 ==========
+
+# 等待用户输入关键词的状态
+_waiting_kw_add = set()
+
+
+async def handle_callback(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """统一处理所有按钮回调"""
+    query = update.callback_query
+    await query.answer()
+    data = query.data
+    user_id = query.from_user.id
+
+    if not _check_admin(user_id):
+        return
+
+    # ---------- 主菜单 ----------
+    if data == "main_menu":
+        keyboard = [
+            [InlineKeyboardButton("📰 最新快讯", callback_data="news_0"),
+             InlineKeyboardButton("📊 新闻总结", callback_data="summary_menu")],
+            [InlineKeyboardButton("📡 订阅源管理", callback_data="sources_menu"),
+             InlineKeyboardButton("🔑 关键词管理", callback_data="keywords_menu")],
+            [InlineKeyboardButton("⚙️ 设置", callback_data="settings_menu")],
+        ]
+        await query.edit_message_text(
+            "🏠 <b>主菜单</b>\n\n请选择功能：",
+            parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup(keyboard),
+        )
+
+    # ---------- 新闻分页 ----------
+    elif data.startswith("news_"):
+        page = int(data.split("_", 1)[1])
+        await _show_news_page(query, page)
+
+    # ---------- 总结菜单 ----------
+    elif data == "summary_menu":
+        await _show_summary_menu(query)
+
+    elif data.startswith("summary_"):
+        period = data[len("summary_"):]
+        start_ts, end_ts = summarizer.get_period_range(period)
+        settings = storage.get_settings()
+        news_list = storage.get_news(limit=500, since_ts=start_ts)
+        news_list = [n for n in news_list if n.get("timestamp", 0) <= end_ts]
+        scorer.score_and_filter(news_list, settings.get("keywords", []))
+        news_list = [n for n in news_list if n.get("score", 0) >= settings.get("min_score", 6)]
+        text = summarizer.generate_summary(news_list, period)
+        back_btn = InlineKeyboardMarkup([[
+            InlineKeyboardButton("🔙 返回", callback_data="summary_menu"),
+            InlineKeyboardButton("🏠 主菜单", callback_data="main_menu"),
+        ]])
+        try:
+            await query.edit_message_text(text, parse_mode=ParseMode.HTML, reply_markup=back_btn)
+        except Exception:
+            await query.edit_message_text(text, reply_markup=back_btn)
+
+    # ---------- 订阅源 ----------
+    elif data == "sources_menu":
+        await _show_sources_menu(query)
+
+    elif data.startswith("toggle_src_"):
+        src = data[len("toggle_src_"):]
+        new_state = storage.toggle_source(src)
+        state_txt = "开启" if new_state else "关闭"
+        await query.answer(f"已{state_txt} {sources.SOURCE_NAMES.get(src, src)}", show_alert=False)
+        await _show_sources_menu(query)
+
+    # ---------- 关键词 ----------
+    elif data == "keywords_menu":
+        await _show_keywords_menu(query)
+
+    elif data.startswith("del_kw_"):
+        kw = data[len("del_kw_"):]
+        storage.remove_keyword(kw)
+        await query.answer(f"已删除关键词: {kw}", show_alert=False)
+        await _show_keywords_menu(query)
+
+    elif data == "add_kw_prompt":
+        _waiting_kw_add.add(user_id)
+        await query.edit_message_text(
+            "🔑 <b>添加关键词</b>\n\n请直接发送关键词文字（可空格分隔多个）：",
+            parse_mode=ParseMode.HTML,
+            reply_markup=InlineKeyboardMarkup([[
+                InlineKeyboardButton("取消", callback_data="keywords_menu")
+            ]]),
+        )
+
+    # ---------- 设置 ----------
+    elif data == "settings_menu":
+        await _show_settings_menu(query)
+
+    elif data == "toggle_push":
+        settings = storage.get_settings()
+        storage.update_settings({"push_enabled": not settings.get("push_enabled", True)})
+        await _show_settings_menu(query)
+
+    elif data == "score_min_up":
+        s = storage.get_settings()
+        storage.update_settings({"min_score": min(10, s.get("min_score", 6) + 1)})
+        await _show_settings_menu(query)
+
+    elif data == "score_min_down":
+        s = storage.get_settings()
+        storage.update_settings({"min_score": max(1, s.get("min_score", 6) - 1)})
+        await _show_settings_menu(query)
+
+    elif data == "score_instant_up":
+        s = storage.get_settings()
+        storage.update_settings({"instant_score": min(10, s.get("instant_score", 8) + 1)})
+        await _show_settings_menu(query)
+
+    elif data == "score_instant_down":
+        s = storage.get_settings()
+        storage.update_settings({"instant_score": max(1, s.get("instant_score", 8) - 1)})
+        await _show_settings_menu(query)
+
+    elif data == "interval_up":
+        s = storage.get_settings()
+        storage.update_settings({"batch_interval": min(120, s.get("batch_interval", 30) + 5)})
+        await _show_settings_menu(query)
+
+    elif data == "interval_down":
+        s = storage.get_settings()
+        storage.update_settings({"batch_interval": max(5, s.get("batch_interval", 30) - 5)})
+        await _show_settings_menu(query)
+
+
+async def handle_text(update: Update, context: ContextTypes.DEFAULT_TYPE):
+    """处理普通文本消息（用于添加关键词）"""
+    user_id = update.effective_user.id
+    if not _check_admin(user_id):
+        return
+    if user_id not in _waiting_kw_add:
+        return
+    _waiting_kw_add.discard(user_id)
+    text = update.message.text.strip()
+    added = []
+    for kw in text.split():
+        if storage.add_keyword(kw):
+            added.append(kw)
+    if added:
+        await update.message.reply_text(
+            f"✅ 已添加关键词: {', '.join(added)}",
+            reply_markup=InlineKeyboardMarkup([[
+                InlineKeyboardButton("🔑 关键词管理", callback_data="keywords_menu")
+            ]]),
+        )
+    else:
+        await update.message.reply_text("⚠️ 关键词已存在或无效")
+
+
+# ========== 后台抓取任务 ==========
+
+# 汇总队列（评分6-7的新闻暂存）
+_batch_queue: list = []
+_last_batch_push: float = 0.0
+_pushed_in_memory: set = set()  # 内存去重，防止并发重复推送
+
+
+async def _fetch_and_process(app: Application):
+    """核心抓取+处理循环"""
+    settings = storage.get_settings()
+    enabled = settings.get("sources", {})
+    push_on = settings.get("push_enabled", True)
+    min_score = settings.get("min_score", 6)
+    instant_score = settings.get("instant_score", 8)
+
+    logger.info("开始抓取新闻...")
+    try:
+        all_news = await sources.fetch_all(enabled)
+    except Exception as e:
+        logger.error(f"fetch_all 异常: {e}")
+        return
+
+    if not all_news:
+        return
+
+    # 评分
+    scorer.score_and_filter(all_news, settings.get("keywords", []))
+
+    # 去重（全局标题去重）
+    all_news = scorer.dedup_news(all_news)
+
+    # 过滤已推送（内存+文件双重去重）
+    pushed_ids = storage.get_pushed_ids() | _pushed_in_memory
+    new_items = [n for n in all_news if n["id"] not in pushed_ids and n.get("score", 0) >= min_score]
+
+    if not new_items:
+        return
+
+    # 保存到存储
+    storage.save_news(new_items)
+
+    if not push_on:
+        return
+
+    instant_items = [n for n in new_items if n.get("score", 0) >= instant_score]
+    batch_items = [n for n in new_items if min_score <= n.get("score", 0) < instant_score]
+
+    # 即时推送
+    for item in instant_items:
+        await _push_instant(app, item)
+
+    # 加入汇总队列（去重）
+    existing_ids = {n["id"] for n in _batch_queue}
+    for item in batch_items:
+        if item["id"] not in existing_ids:
+            _batch_queue.append(item)
+            existing_ids.add(item["id"])
+
+    # 标记已推送（即时+批量都标记，防止重复）
+    pushed = [n["id"] for n in instant_items + batch_items]
+    _pushed_in_memory.update(pushed)
+    storage.mark_pushed(pushed)
+
+
+async def _push_instant(app: Application, item: dict):
+    """立即推送单条重磅新闻"""
+    score = item.get("score", 0)
+    if score >= 9:
+        emoji = "🔥🔥🔥"
+    elif score >= 8:
+        emoji = "⚡ 重磅"
+    else:
+        emoji = "📌"
+    src = item.get("source_name", "")
+    title = item.get("title", "")
+    ts = item.get("time_str", "")
+    url = item.get("url", "")
+    import html as html_mod
+    title_safe = html_mod.escape(title)
+    if url:
+        title_display = f'<a href="{url}">{title_safe}</a>'
+    else:
+        title_display = f"<b>{title_safe}</b>"
+    text = f"{emoji}\n\n{title_display}\n\n📡 {src}  🕐 {ts}  评分:{score}"
+    try:
+        await app.bot.send_message(
+            chat_id=ADMIN_ID, text=text, parse_mode=ParseMode.HTML,
+        )
+    except Exception as e:
+        logger.error(f"即时推送失败: {e}")
+        try:
+            await app.bot.send_message(chat_id=ADMIN_ID, text=f"{emoji}\n\n{title}\n\n{src} {ts}")
+        except Exception as e2:
+            logger.error(f"即时推送备用失败: {e2}")
+
+
+async def _push_batch(app: Application):
+    """汇总推送（每N分钟一次）"""
+    global _batch_queue, _last_batch_push
+    if not _batch_queue:
+        return
+
+    settings = storage.get_settings()
+    interval = settings.get("batch_interval", 30) * 60
+    now = time.time()
+    if now - _last_batch_push < interval:
+        return
+
+    items = _batch_queue[:]
+    _batch_queue.clear()
+    _last_batch_push = now
+
+    # 标记已推送
+    storage.mark_pushed([n["id"] for n in items])
+
+    if not items:
+        return
+
+    import html as html_mod
+    from collections import defaultdict
+
+    # 按来源分组
+    source_map = defaultdict(list)
+    for item in sorted(items, key=lambda x: x.get("score", 0), reverse=True):
+        src = item.get("source_name", "未知")
+        source_map[src].append(item)
+
+    source_emoji = {
+        "金十数据": "💰", "华尔街见闻": "📈", "36氪": "🚀",
+        "新浪财经": "📊", "Google News": "🌐",
+        "Finviz": "📉", "TechCrunch": "💻",
+    }
+
+    lines = [f"🟡 <b>快讯汇总</b>（{len(items)}条）"]
+    lines.append("")
+    for src, src_items in source_map.items():
+        emoji = source_emoji.get(src, "📰")
+        lines.append(f"{emoji} <b>{src}</b>（{len(src_items)}）")
+        for item in src_items[:10]:
+            title = item.get("title", "")
+            url = item.get("url", "")
+            title_safe = html_mod.escape(title)
+            if url:
+                lines.append(f"  • <a href=\"{url}\">{title_safe}</a>")
+            else:
+                lines.append(f"  • {title_safe}")
+        if len(src_items) > 10:
+            lines.append(f"  ... 还有{len(src_items) - 10}条")
+        lines.append("")
+
+    lines.append(f"⏰ {datetime.now().strftime('%H:%M')}")
+
+    text = "\n".join(lines)
+    try:
+        await app.bot.send_message(
+            chat_id=ADMIN_ID, text=text, parse_mode=ParseMode.HTML,
+        )
+    except Exception as e:
+        logger.error(f"汇总推送失败: {e}")
+        try:
+            await app.bot.send_message(chat_id=ADMIN_ID, text="\n".join(lines[:15]))
+        except Exception as e2:
+            logger.error(f"汇总推送备用失败: {e2}")
+
+
+async def _send_scheduled_summary(app: Application, period: str):
+    """发送定时总结"""
+    settings = storage.get_settings()
+    start_ts, end_ts = summarizer.get_period_range(period)
+    news_list = storage.get_news(limit=500, since_ts=start_ts)
+    news_list = [n for n in news_list if n.get("timestamp", 0) <= end_ts]
+    scorer.score_and_filter(news_list, settings.get("keywords", []))
+    news_list = [n for n in news_list if n.get("score", 0) >= settings.get("min_score", 6)]
+    text = summarizer.generate_summary(news_list, period)
+    try:
+        await app.bot.send_message(
+            chat_id=ADMIN_ID, text=text, parse_mode=ParseMode.HTML,
+        )
+    except Exception as e:
+        logger.error(f"定时总结推送失败: {e}")
+        try:
+            await app.bot.send_message(chat_id=ADMIN_ID, text=text)
+        except Exception:
+            pass
+
+
+# ========== 调度器 ==========
+
+class Scheduler:
+    """轻量调度器，基于 asyncio"""
+
+    def __init__(self, app: Application):
+        self.app = app
+        self._tasks: list[asyncio.Task] = []
+
+    def start(self):
+        self._tasks.append(asyncio.create_task(self._fast_loop()))
+        self._tasks.append(asyncio.create_task(self._slow_loop()))
+        self._tasks.append(asyncio.create_task(self._batch_loop()))
+        self._tasks.append(asyncio.create_task(self._summary_loop()))
+        logger.info("调度器已启动")
+
+    async def _fast_loop(self):
+        """金十/华尔街见闻 每2分钟抓取"""
+        while True:
+            try:
+                await _fetch_and_process(self.app)
+            except Exception as e:
+                logger.error(f"fast_loop 异常: {e}")
+            await asyncio.sleep(120)
+
+    async def _slow_loop(self):
+        """36氪/新浪 每5分钟额外触发（fast_loop已包含全部，这里可扩展）"""
+        await asyncio.sleep(60)  # 错开启动时间
+        while True:
+            try:
+                await _fetch_and_process(self.app)
+            except Exception as e:
+                logger.error(f"slow_loop 异常: {e}")
+            await asyncio.sleep(300)
+
+    async def _batch_loop(self):
+        """每分钟检查是否需要汇总推送"""
+        await asyncio.sleep(30)
+        while True:
+            try:
+                await _push_batch(self.app)
+            except Exception as e:
+                logger.error(f"batch_loop 异常: {e}")
+            await asyncio.sleep(60)
+
+    async def _summary_loop(self):
+        """定时总结 08:00 / 11:30 / 20:00"""
+        schedule = [
+            (8, 0,  "昨晚到今早"),
+            (11, 30, "上午"),
+            (20, 0,  "全天"),
+        ]
+        triggered = set()
+        while True:
+            now = datetime.now()
+            key = f"{now.date()}"
+            for hour, minute, period in schedule:
+                tid = f"{key}_{hour}_{minute}"
+                if tid not in triggered:
+                    # 在目标时间 ±2 分钟内触发
+                    target = now.replace(hour=hour, minute=minute, second=0)
+                    diff = abs((now - target).total_seconds())
+                    if diff <= 120:
+                        triggered.add(tid)
+                        try:
+                            await _send_scheduled_summary(self.app, period)
+                        except Exception as e:
+                            logger.error(f"定时总结异常: {e}")
+            # 每分钟检查一次
+            await asyncio.sleep(60)
+
+    def stop(self):
+        for t in self._tasks:
+            t.cancel()
+
+
+# ========== 主函数 ==========
+
+def main():
+    if not BOT_TOKEN:
+        logger.error("请设置 BOT_TOKEN 环境变量")
+        return
+
+    logger.info(f"机器人启动，ADMIN_ID={ADMIN_ID}")
+
+    app = Application.builder().token(BOT_TOKEN).build()
+
+    # 注册命令
+    app.add_handler(CommandHandler("start",    cmd_start))
+    app.add_handler(CommandHandler("news",     cmd_news))
+    app.add_handler(CommandHandler("summary",  cmd_summary))
+    app.add_handler(CommandHandler("sources",  cmd_sources))
+    app.add_handler(CommandHandler("keywords", cmd_keywords))
+    app.add_handler(CommandHandler("settings", cmd_settings))
+
+    # 注册回调
+    app.add_handler(CallbackQueryHandler(handle_callback))
+
+    # 注册文本消息（用于添加关键词）
+    app.add_handler(MessageHandler(filters.TEXT & ~filters.COMMAND, handle_text))
+
+    # 启动调度器（在 post_init 中启动，确保 event loop 已就绪）
+    scheduler = Scheduler(app)
+
+    async def post_init(application: Application):
+        scheduler.start()
+        logger.info("调度器已在 post_init 中启动")
+
+    app.post_init = post_init
+
+    logger.info("开始 polling...")
+    app.run_polling(drop_pending_updates=True)
+
+
+if __name__ == "__main__":
+    main()

projects/news-bot/requirements.txt

@@ -0,0 +1,3 @@
+python-telegram-bot==21.3
+httpx==0.27.0
+apscheduler==3.10.4

projects/news-bot/scorer.py

@@ -0,0 +1,122 @@
+"""
+评分和过滤模块 - 基于关键词权重 + 规则引擎
+不调用外部 AI API，纯本地规则评分
+"""
+import re
+import logging
+
+logger = logging.getLogger(__name__)
+
+# 高权重关键词（出现即加分）
+HIGH_WEIGHT_KEYWORDS = {
+    # AI / 算力相关
+    "英伟达": 3, "NVIDIA": 3, "AMD": 2, "算力": 2, "AI": 2,
+    "光模块": 2, "存储": 1, "芯片": 2, "GPU": 2, "大模型": 2,
+    "ChatGPT": 2, "OpenAI": 2, "Anthropic": 2, "DeepSeek": 2,
+    # 中国科技公司
+    "华为": 2, "腾讯": 2, "阿里": 2, "字节": 2, "小米": 2,
+    "百度": 1, "京东": 1, "美团": 1, "拼多多": 1, "比亚迪": 2,
+    # 重磅事件
+    "收购": 2, "突破": 2, "泄露": 3, "内幕": 3, "传闻": 2,
+    "重磅": 3, "突发": 3, "暴涨": 2, "暴跌": 2, "崩盘": 3,
+    "熔断": 3, "制裁": 2, "禁令": 2, "封锁": 2,
+    # 股市相关
+    "A股": 1, "港股": 1, "美股": 1, "涨停": 2, "跌停": 2,
+    "IPO": 2, "上市": 1, "退市": 2, "回购": 1,
+    # 宏观
+    "降息": 2, "加息": 2, "美联储": 2, "央行": 1, "GDP": 1,
+    "CPI": 1, "非农": 2, "失业率": 1,
+}
+
+# 低价值关键词（出现则减分）
+LOW_VALUE_PATTERNS = [
+    r"广告", r"推广", r"优惠券", r"免费领",
+    r"点击.*查看", r"扫码", r"关注.*公众号",
+]
+
+
+def score_news(news_item: dict, user_keywords: list = None) -> int:
+    """
+    对单条新闻评分，返回 0-10 分
+    评分规则：
+    - 基础分 3 分
+    - 匹配高权重关键词加分
+    - 匹配用户自定义关键词加分
+    - 信息源标记为重要加分
+    - 匹配低价值模式减分
+    - 最终分数限制在 0-10
+    """
+    title = news_item.get("title", "")
+    if not title:
+        return 0
+
+    score = 3  # 基础分
+
+    # 1. 高权重关键词匹配
+    for keyword, weight in HIGH_WEIGHT_KEYWORDS.items():
+        if keyword.lower() in title.lower():
+            score += weight
+
+    # 2. 用户自定义关键词匹配
+    if user_keywords:
+        for kw in user_keywords:
+            if kw.lower() in title.lower():
+                score += 1
+
+    # 3. 信息源标记为重要
+    if news_item.get("important"):
+        score += 2
+
+    # 4. 低价值内容减分
+    for pattern in LOW_VALUE_PATTERNS:
+        if re.search(pattern, title):
+            score -= 3
+
+    # 5. 标题太短可能是无效内容
+    if len(title) < 10:
+        score -= 2
+
+    return max(0, min(10, score))
+
+
+def is_similar(title1: str, title2: str, threshold: float = 0.7) -> bool:
+    """
+    简单的标题相似度判断
+    使用字符级别的 Jaccard 相似度
+    """
+    if not title1 or not title2:
+        return False
+    # 去除标点和空格
+    clean1 = re.sub(r"[^\w]", "", title1)
+    clean2 = re.sub(r"[^\w]", "", title2)
+    if not clean1 or not clean2:
+        return False
+    # 2-gram 集合
+    set1 = set(clean1[i:i+2] for i in range(len(clean1)-1))
+    set2 = set(clean2[i:i+2] for i in range(len(clean2)-1))
+    if not set1 or not set2:
+        return clean1 == clean2
+    intersection = set1 & set2
+    union = set1 | set2
+    return len(intersection) / len(union) >= threshold
+
+
+def dedup_news(news_list: list) -> list:
+    """去重：基于标题相似度，保留最早的一条"""
+    result = []
+    for item in news_list:
+        is_dup = False
+        for existing in result:
+            if is_similar(item.get("title", ""), existing.get("title", "")):
+                is_dup = True
+                break
+        if not is_dup:
+            result.append(item)
+    return result
+
+
+def score_and_filter(news_list: list, user_keywords: list = None) -> list:
+    """批量评分并过滤，返回带评分的新闻列表"""
+    for item in news_list:
+        item["score"] = score_news(item, user_keywords)
+    return news_list

projects/news-bot/sources.py

@@ -0,0 +1,374 @@
+"""
+信息源抓取模块 - 支持金十、华尔街见闻、36氪、新浪财经、Google News、Finviz、TechCrunch
+"""
+import re
+import time
+import hashlib
+import logging
+import httpx
+import xml.etree.ElementTree as ET
+from datetime import datetime
+from email.utils import parsedate_to_datetime
+
+logger = logging.getLogger(__name__)
+
+# 请求超时设置
+TIMEOUT = 10
+
+# 翻译缓存，避免重复翻译
+_translate_cache = {}
+
+
+async def translate_to_zh(text: str) -> str:
+    """用 Google Translate 免费接口将英文翻译为中文"""
+    if not text:
+        return text
+    # 检测是否主要是中文，是则跳过
+    zh_count = sum(1 for c in text if '\u4e00' <= c <= '\u9fff')
+    if zh_count > len(text) * 0.3:
+        return text
+    # 查缓存
+    cache_key = text[:100]
+    if cache_key in _translate_cache:
+        return _translate_cache[cache_key]
+    try:
+        async with httpx.AsyncClient(timeout=8) as client:
+            resp = await client.get(
+                "https://translate.googleapis.com/translate_a/single",
+                params={"client": "gtx", "sl": "en", "tl": "zh-CN", "dt": "t", "q": text[:500]},
+            )
+            result = resp.json()
+            translated = "".join(seg[0] for seg in result[0] if seg[0])
+            _translate_cache[cache_key] = translated
+            # 限制缓存大小
+            if len(_translate_cache) > 500:
+                keys = list(_translate_cache.keys())[:200]
+                for k in keys:
+                    del _translate_cache[k]
+            return translated
+    except Exception as e:
+        logger.error(f"翻译失败: {e}")
+        return text
+
+
+def _make_id(source: str, title: str) -> str:
+    """生成新闻唯一 ID"""
+    raw = f"{source}:{title}"
+    return hashlib.md5(raw.encode()).hexdigest()[:16]
+
+
+def _safe_ts(val, default=0) -> int:
+    """安全转换时间戳"""
+    try:
+        if isinstance(val, (int, float)):
+            # 如果是毫秒级时间戳，转为秒
+            return int(val) if val < 2000000000 else int(val / 1000)
+        if isinstance(val, str):
+            # 尝试解析常见格式
+            for fmt in ["%Y-%m-%d %H:%M:%S", "%Y-%m-%dT%H:%M:%S", "%Y-%m-%dT%H:%M:%S.%f"]:
+                try:
+                    return int(datetime.strptime(val[:19], fmt[:len(val)+2]).timestamp())
+                except ValueError:
+                    continue
+    except Exception:
+        pass
+    return default or int(time.time())
+
+
+async def fetch_jin10() -> list:
+    """抓取金十数据快讯"""
+    url = "https://flash-api.jin10.com/get_flash_list"
+    params = {"channel": "-8200", "vip": "1", "max_time": "", "t": "1"}
+    headers = {"x-app-id": "bVBF4FyRTn5NJF5n", "x-version": "1.0.0"}
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, params=params, headers=headers)
+            data = resp.json().get("data", [])
+            for item in data:
+                # 金十的 data 字段可能是 dict 或直接文本
+                content = ""
+                if isinstance(item.get("data"), dict):
+                    content = item["data"].get("content", "") or item["data"].get("title", "")
+                elif isinstance(item.get("data"), str):
+                    content = item["data"]
+                # 备用：用 content 字段
+                if not content:
+                    content = item.get("content", "")
+                if not content:
+                    continue
+                # 清理 HTML 标签
+                import re
+                content = re.sub(r"<[^>]+>", "", content).strip()
+                if not content:
+                    continue
+                ts = _safe_ts(item.get("time", ""))
+                news_id = item.get("id", "")
+                news_url = f"https://www.jin10.com/flash_detail/{news_id}.html" if news_id else ""
+                results.append({
+                    "id": _make_id("jin10", content[:80]),
+                    "source": "jin10",
+                    "source_name": "金十数据",
+                    "title": content[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": item.get("time", ""),
+                    "important": item.get("important", 0) == 1,
+                })
+    except Exception as e:
+        logger.error(f"金十数据抓取失败: {e}")
+    return results
+
+
+async def fetch_wallstreet() -> list:
+    """抓取华尔街见闻快讯"""
+    url = "https://api-one.wallstcn.com/apiv1/content/lives"
+    params = {"channel": "global-channel", "limit": "20"}
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, params=params)
+            items = resp.json().get("data", {}).get("items", [])
+            for item in items:
+                title = item.get("content_text", "") or item.get("title", "")
+                if not title:
+                    continue
+                import re
+                title = re.sub(r"<[^>]+>", "", title).strip()
+                if not title:
+                    continue
+                ts = _safe_ts(item.get("display_time", 0))
+                news_url = item.get("uri", "") or ""
+                results.append({
+                    "id": _make_id("wallstreet", title[:80]),
+                    "source": "wallstreet",
+                    "source_name": "华尔街见闻",
+                    "title": title[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": datetime.fromtimestamp(ts).strftime("%H:%M:%S") if ts else "",
+                    "important": item.get("is_important", False),
+                })
+    except Exception as e:
+        logger.error(f"华尔街见闻抓取失败: {e}")
+    return results
+
+
+async def fetch_kr36() -> list:
+    """抓取36氪快讯"""
+    url = "https://36kr.com/api/newsflash"
+    params = {"per_page": "20"}
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, params=params)
+            items = resp.json().get("data", {}).get("items", [])
+            for item in items:
+                title = item.get("title", "") or item.get("entity_name", "")
+                if not title:
+                    continue
+                ts = _safe_ts(item.get("published_at", ""))
+                news_id = item.get("id", "")
+                news_url = f"https://36kr.com/newsflashes/{news_id}" if news_id else ""
+                results.append({
+                    "id": _make_id("kr36", title[:80]),
+                    "source": "kr36",
+                    "source_name": "36氪",
+                    "title": title[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": datetime.fromtimestamp(ts).strftime("%H:%M:%S") if ts else "",
+                    "important": False,
+                })
+    except Exception as e:
+        logger.error(f"36氪抓取失败: {e}")
+    return results
+
+
+async def fetch_sina() -> list:
+    """抓取新浪财经快讯"""
+    url = "https://feed.mix.sina.com.cn/api/roll/get"
+    params = {"pageid": "153", "lid": "2516", "k": "", "num": "20", "page": "1"}
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, params=params)
+            data = resp.json().get("result", {}).get("data", [])
+            for item in data:
+                title = item.get("title", "")
+                if not title:
+                    continue
+                import re
+                title = re.sub(r"<[^>]+>", "", title).strip()
+                if not title:
+                    continue
+                ts = _safe_ts(item.get("ctime", item.get("createtime", 0)))
+                news_url = item.get("url", "") or item.get("link", "") or ""
+                results.append({
+                    "id": _make_id("sina", title[:80]),
+                    "source": "sina",
+                    "source_name": "新浪财经",
+                    "title": title[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": datetime.fromtimestamp(ts).strftime("%H:%M:%S") if ts else "",
+                    "important": False,
+                })
+    except Exception as e:
+        logger.error(f"新浪财经抓取失败: {e}")
+    return results
+
+
+async def fetch_google_news() -> list:
+    """抓取 Google News 科技频道 RSS（聚合路透社/彭博社等）"""
+    url = "https://news.google.com/rss/topics/CAAqJggKIiBDQkFTRWdvSUwyMHZNRGRqTVhZU0FtVnVHZ0pWVXlnQVAB"
+    params = {"hl": "en-US", "gl": "US", "ceid": "US:en"}
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, params=params, headers={"User-Agent": "Mozilla/5.0"})
+            root = ET.fromstring(resp.text)
+            for item in root.findall(".//item")[:20]:
+                title = item.findtext("title", "").strip()
+                if not title:
+                    continue
+                # 去掉来源后缀 " - Bloomberg.com" 等
+                source_tag = title.rsplit(" - ", 1)[-1] if " - " in title else ""
+                news_url = item.findtext("link", "").strip()
+                ts = 0
+                pub = item.findtext("pubDate", "")
+                if pub:
+                    try:
+                        ts = int(parsedate_to_datetime(pub).timestamp())
+                    except Exception:
+                        ts = int(time.time())
+                title_zh = await translate_to_zh(title)
+                results.append({
+                    "id": _make_id("google", title[:80]),
+                    "source": "google",
+                    "source_name": f"Google News ({source_tag})" if source_tag else "Google News",
+                    "title": title_zh[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": datetime.fromtimestamp(ts).strftime("%H:%M:%S") if ts else "",
+                    "important": False,
+                })
+    except Exception as e:
+        logger.error(f"Google News 抓取失败: {e}")
+    return results
+
+
+async def fetch_finviz() -> list:
+    """抓取 Finviz 美股财经新闻"""
+    url = "https://finviz.com/news.ashx"
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, headers={"User-Agent": "Mozilla/5.0"})
+            # 解析 HTML 中的新闻标题和链接
+            matches = re.findall(r'class="nn-tab-link"[^>]*href="([^"]*)"[^>]*>([^<]+)', resp.text)
+            if not matches:
+                # 备用：只提取标题
+                titles = re.findall(r'class="nn-tab-link"[^>]*>([^<]+)', resp.text)
+                matches = [("", t) for t in titles]
+            for link, title in matches[:20]:
+                title = title.strip()
+                if not title:
+                    continue
+                title_zh = await translate_to_zh(title)
+                results.append({
+                    "id": _make_id("finviz", title[:80]),
+                    "source": "finviz",
+                    "source_name": "Finviz",
+                    "title": title_zh[:200],
+                    "url": link or "",
+                    "timestamp": int(time.time()),
+                    "time_str": datetime.now().strftime("%H:%M:%S"),
+                    "important": False,
+                })
+    except Exception as e:
+        logger.error(f"Finviz 抓取失败: {e}")
+    return results
+
+
+async def fetch_techcrunch() -> list:
+    """抓取 TechCrunch RSS 科技新闻"""
+    url = "https://techcrunch.com/feed/"
+    results = []
+    try:
+        async with httpx.AsyncClient(timeout=TIMEOUT) as client:
+            resp = await client.get(url, headers={"User-Agent": "Mozilla/5.0"})
+            root = ET.fromstring(resp.text)
+            for item in root.findall(".//item")[:20]:
+                title = item.findtext("title", "").strip()
+                if not title:
+                    continue
+                news_url = item.findtext("link", "").strip()
+                ts = 0
+                pub = item.findtext("pubDate", "")
+                if pub:
+                    try:
+                        ts = int(parsedate_to_datetime(pub).timestamp())
+                    except Exception:
+                        ts = int(time.time())
+                title_zh = await translate_to_zh(title)
+                results.append({
+                    "id": _make_id("techcrunch", title[:80]),
+                    "source": "techcrunch",
+                    "source_name": "TechCrunch",
+                    "title": title_zh[:200],
+                    "url": news_url,
+                    "timestamp": ts,
+                    "time_str": datetime.fromtimestamp(ts).strftime("%H:%M:%S") if ts else "",
+                    "important": False,
+                })
+    except Exception as e:
+        logger.error(f"TechCrunch 抓取失败: {e}")
+    return results
+
+
+# 源名称 → 抓取函数映射
+SOURCE_FETCHERS = {
+    "jin10": fetch_jin10,
+    "wallstreet": fetch_wallstreet,
+    "kr36": fetch_kr36,
+    "sina": fetch_sina,
+    "google": fetch_google_news,
+    "finviz": fetch_finviz,
+    "techcrunch": fetch_techcrunch,
+}
+
+SOURCE_NAMES = {
+    "jin10": "金十数据",
+    "wallstreet": "华尔街见闻",
+    "kr36": "36氪",
+    "sina": "新浪财经",
+    "google": "Google News",
+    "finviz": "Finviz",
+    "techcrunch": "TechCrunch",
+}
+
+
+async def fetch_all(enabled_sources: dict = None) -> list:
+    """抓取所有启用的信息源，返回合并后的新闻列表"""
+    import asyncio
+    if enabled_sources is None:
+        enabled_sources = {k: True for k in SOURCE_FETCHERS}
+
+    tasks = []
+    for name, fetcher in SOURCE_FETCHERS.items():
+        if enabled_sources.get(name, True):
+            tasks.append(fetcher())
+
+    all_news = []
+    results = await asyncio.gather(*tasks, return_exceptions=True)
+    for result in results:
+        if isinstance(result, Exception):
+            logger.error(f"抓取异常: {result}")
+            continue
+        if isinstance(result, list):
+            all_news.extend(result)
+
+    # 按时间排序，最新的在前
+    all_news.sort(key=lambda x: x.get("timestamp", 0), reverse=True)
+    return all_news

projects/news-bot/storage.py

@@ -0,0 +1,187 @@
+"""
+数据存储模块 - 使用 JSON 文件持久化
+"""
+import json
+import os
+import time
+import logging
+from pathlib import Path
+from threading import Lock
+
+logger = logging.getLogger(__name__)
+
+DATA_DIR = os.environ.get("DATA_DIR", "/opt/news-bot/data")
+
+# 文件锁，防止并发写入
+_locks = {}
+
+
+def _get_lock(name: str) -> Lock:
+    if name not in _locks:
+        _locks[name] = Lock()
+    return _locks[name]
+
+
+def _ensure_dir():
+    """确保数据目录存在"""
+    Path(DATA_DIR).mkdir(parents=True, exist_ok=True)
+
+
+def _read_json(filename: str, default=None):
+    """读取 JSON 文件"""
+    filepath = os.path.join(DATA_DIR, filename)
+    try:
+        if os.path.exists(filepath):
+            with open(filepath, "r", encoding="utf-8") as f:
+                return json.load(f)
+    except Exception as e:
+        logger.error(f"读取 {filename} 失败: {e}")
+    return default if default is not None else {}
+
+
+def _write_json(filename: str, data):
+    """写入 JSON 文件"""
+    _ensure_dir()
+    filepath = os.path.join(DATA_DIR, filename)
+    try:
+        with open(filepath, "w", encoding="utf-8") as f:
+            json.dump(data, f, ensure_ascii=False, indent=2)
+    except Exception as e:
+        logger.error(f"写入 {filename} 失败: {e}")
+
+
+# ========== 新闻存储 ==========
+
+def save_news(news_list: list):
+    """保存新闻列表（追加，自动去重）"""
+    with _get_lock("news"):
+        existing = _read_json("news.json", [])
+        existing_ids = {n.get("id") for n in existing}
+        added = 0
+        for item in news_list:
+            if item.get("id") not in existing_ids:
+                existing.append(item)
+                existing_ids.add(item["id"])
+                added += 1
+        # 只保留最近 2000 条
+        if len(existing) > 2000:
+            existing = existing[-2000:]
+        _write_json("news.json", existing)
+        return added
+
+
+def get_news(limit=50, since_ts=0) -> list:
+    """获取新闻，可按时间过滤"""
+    news = _read_json("news.json", [])
+    if since_ts:
+        news = [n for n in news if n.get("timestamp", 0) >= since_ts]
+    return news[-limit:]
+
+
+def get_news_by_score(min_score=6, since_ts=0) -> list:
+    """按评分获取新闻"""
+    news = get_news(limit=500, since_ts=since_ts)
+    return [n for n in news if n.get("score", 0) >= min_score]
+
+
+# ========== 已推送记录 ==========
+
+def mark_pushed(news_ids: list):
+    """标记新闻为已推送"""
+    with _get_lock("pushed"):
+        pushed = set(_read_json("pushed.json", []))
+        pushed.update(news_ids)
+        # 只保留最近 5000 条
+        pushed_list = list(pushed)[-5000:]
+        _write_json("pushed.json", pushed_list)
+
+
+def is_pushed(news_id: str) -> bool:
+    """检查新闻是否已推送"""
+    pushed = set(_read_json("pushed.json", []))
+    return news_id in pushed
+
+
+def get_pushed_ids() -> set:
+    """获取所有已推送 ID"""
+    return set(_read_json("pushed.json", []))
+
+
+# ========== 用户设置 ==========
+
+DEFAULT_SETTINGS = {
+    "push_enabled": True,
+    "batch_interval": 30,  # 汇总推送间隔（分钟）
+    "min_score": 6,  # 最低推送评分
+    "instant_score": 8,  # 即时推送评分阈值
+    "sources": {
+        "jin10": True,
+        "wallstreet": True,
+        "kr36": True,
+        "sina": True,
+    },
+    "keywords": [
+        "AI", "英伟达", "NVIDIA", "AMD", "存储", "光模块", "算力",
+        "华为", "腾讯", "阿里", "字节", "小米",
+        "收购", "突破", "泄露", "内幕", "传闻", "重磅",
+        "A股", "港股", "美股",
+    ],
+}
+
+
+def get_settings() -> dict:
+    """获取用户设置"""
+    settings = _read_json("settings.json", None)
+    if settings is None:
+        settings = DEFAULT_SETTINGS.copy()
+        _write_json("settings.json", settings)
+    # 补全缺失字段
+    for k, v in DEFAULT_SETTINGS.items():
+        if k not in settings:
+            settings[k] = v
+    return settings
+
+
+def update_settings(updates: dict):
+    """更新用户设置"""
+    with _get_lock("settings"):
+        settings = get_settings()
+        settings.update(updates)
+        _write_json("settings.json", settings)
+    return settings
+
+
+def toggle_source(source_name: str) -> bool:
+    """切换信息源开关，返回新状态"""
+    with _get_lock("settings"):
+        settings = get_settings()
+        current = settings.get("sources", {}).get(source_name, True)
+        settings.setdefault("sources", {})[source_name] = not current
+        _write_json("settings.json", settings)
+        return not current
+
+
+def add_keyword(keyword: str) -> bool:
+    """添加关键词，返回是否成功"""
+    with _get_lock("settings"):
+        settings = get_settings()
+        kw_list = settings.get("keywords", [])
+        if keyword in kw_list:
+            return False
+        kw_list.append(keyword)
+        settings["keywords"] = kw_list
+        _write_json("settings.json", settings)
+        return True
+
+
+def remove_keyword(keyword: str) -> bool:
+    """删除关键词，返回是否成功"""
+    with _get_lock("settings"):
+        settings = get_settings()
+        kw_list = settings.get("keywords", [])
+        if keyword not in kw_list:
+            return False
+        kw_list.remove(keyword)
+        settings["keywords"] = kw_list
+        _write_json("settings.json", settings)
+        return True

projects/news-bot/summarizer.py

@@ -0,0 +1,97 @@
+"""
+新闻总结生成模块
+使用规则引擎生成中文总结（不调用外部 AI API）
+"""
+import logging
+from datetime import datetime, timedelta
+from collections import defaultdict
+
+logger = logging.getLogger(__name__)
+
+
+def generate_summary(news_list: list, period: str = "全天") -> str:
+    """
+    生成新闻总结
+    period: "最近1小时" / "上午" / "全天" / "昨晚到今早"
+    """
+    if not news_list:
+        return f"📭 {period}暂无重要新闻"
+
+    # 按评分排序，高分在前
+    sorted_news = sorted(news_list, key=lambda x: x.get("score", 0), reverse=True)
+
+    # 按来源分组统计
+    source_count = defaultdict(int)
+    for n in sorted_news:
+        source_count[n.get("source_name", "未知")] += 1
+
+    # 构建总结
+    lines = []
+    lines.append(f"📊 *{period}新闻总结*")
+    lines.append(f"━━━━━━━━━━━━━━━")
+    lines.append(f"📰 共 {len(sorted_news)} 条快讯")
+    lines.append("")
+
+    # 🔴 重磅新闻（评分 >= 8）
+    hot_news = [n for n in sorted_news if n.get("score", 0) >= 8]
+    if hot_news:
+        lines.append("🔴 *重磅快讯*")
+        for n in hot_news[:5]:
+            emoji = _score_emoji(n.get("score", 0))
+            lines.append(f"  {emoji} {n['title']}")
+        lines.append("")
+
+    # 🟡 值得关注（评分 6-7）
+    mid_news = [n for n in sorted_news if 6 <= n.get("score", 0) <= 7]
+    if mid_news:
+        lines.append("🟡 *值得关注*")
+        for n in mid_news[:8]:
+            lines.append(f"  • {n['title']}")
+        lines.append("")
+
+    # 来源统计
+    lines.append("📡 *来源分布*")
+    for src, cnt in sorted(source_count.items(), key=lambda x: x[1], reverse=True):
+        lines.append(f"  {src}: {cnt}条")
+
+    lines.append("")
+    lines.append(f"⏰ 生成时间: {datetime.now().strftime('%H:%M')}")
+
+    return "\n".join(lines)
+
+
+def _score_emoji(score: int) -> str:
+    """根据评分返回 emoji"""
+    if score >= 9:
+        return "🔥"
+    elif score >= 8:
+        return "⚡"
+    elif score >= 7:
+        return "📌"
+    elif score >= 6:
+        return "📎"
+    else:
+        return "•"
+
+
+def get_period_range(period: str) -> tuple:
+    """
+    根据时间段名称返回 (start_ts, end_ts)
+    """
+    now = datetime.now()
+    today_start = now.replace(hour=0, minute=0, second=0, microsecond=0)
+
+    if period == "最近1小时":
+        start = now - timedelta(hours=1)
+    elif period == "上午":
+        start = today_start.replace(hour=6)
+        now = min(now, today_start.replace(hour=12))
+    elif period == "全天":
+        start = today_start
+    elif period == "昨晚到今早":
+        start = (today_start - timedelta(days=1)).replace(hour=20)
+        now = today_start.replace(hour=8)
+    else:
+        start = now - timedelta(hours=3)
+
+    return int(start.timestamp()), int(now.timestamp())