weflow-cli 1.5.0

package/scripts/generate_review.py

@@ -0,0 +1,170 @@
+#!/usr/bin/env python3
+"""
+AI 日报生成 — 基于 git diff 分析今日知识变更，生成结构化学习日报。
+
+用法:
+  python scripts/generate_review.py --api-key <key> [--date 2026-05-15] [--output Reviews]
+"""
+import sys, os, json, re, time, subprocess
+from pathlib import Path
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+from _utils import create_engine, parse_frontmatter
+
+DAILY_PROMPT = """你是知识管理助手，基于今日文章内容生成学习日报。
+
+今日新增文章（标题 + 来源 + 摘要）：
+{articles_text}
+
+请按格式返回：
+【今日焦点】
+（2-3句话，今日最重要的知识收获）
+
+【按主题】
+AI:
+- 要点1
+- 要点2
+
+学术:
+- 要点1
+- 要点2
+
+投资:
+- 要点1
+
+【关键收获】
+（1句话，今日最值得记住的一点）
+
+要求：只基于提供的文章内容，不要编造。无对应主题则标注"（无）"。"""
+
+
+def scan_today_articles(target_date: str, source='output/biz-daily') -> list[dict]:
+    """扫描指定日期的文章目录，提取 frontmatter + 摘要。"""
+    date_dir = Path(source) / target_date
+    if not date_dir.exists():
+        return []
+
+    articles = []
+    for md_file in sorted(date_dir.rglob('*.md')):
+        if md_file.name == 'README.md' or md_file.name.startswith('.'):
+            continue
+        try:
+            with open(md_file, 'r', encoding='utf-8') as f:
+                content = f.read()
+        except:
+            continue
+
+        fm, body = parse_frontmatter(content)
+        if not fm:
+            continue
+
+        # Extract summary
+        summary = ''
+        m = re.search(r'## (?:AI 摘要|深度解析)\n\n(.+?)(?=\n\n(?:##|---)|\Z)', body, re.DOTALL)
+        if m:
+            summary = m.group(1).strip()[:300]
+
+        articles.append({
+            'title': fm.get('title', md_file.stem),
+            'source': fm.get('source', ''),
+            'topic': fm.get('topic', '学术'),
+            'tags': fm.get('tags', []),
+            'summary': summary,
+        })
+    return articles
+
+
+def main():
+    sys.stdout.reconfigure(encoding='utf-8', errors='replace')
+    import argparse
+    parser = argparse.ArgumentParser(description='AI 学习日报生成')
+    parser.add_argument('--api-key', help='DeepSeek API key（或环境变量 DEEPSEEK_API_KEY）')
+    parser.add_argument('--date', help='日期 YYYY-MM-DD')
+    parser.add_argument('--engine', default='deepseek', help='AI 引擎（deepseek/claude/ollama）')
+    parser.add_argument('--source', default='output/biz-daily', help='文章目录')
+    parser.add_argument('--output', default='output/reviews', help='输出目录')
+    args = parser.parse_args()
+
+    api_key = args.api_key or os.environ.get('DEEPSEEK_API_KEY', '')
+    if not api_key:
+        print('[ERROR] 需要 API key')
+        sys.exit(1)
+
+    # Date
+    from datetime import datetime, timezone, timedelta
+    if args.date:
+        target_date = args.date
+    else:
+        tz = timezone(timedelta(hours=8))
+        target_date = datetime.now(tz).strftime('%Y-%m-%d')
+
+    print(f'=== AI 日报 — {target_date} ===\n')
+
+    # Scan articles
+    articles = scan_today_articles(target_date, args.source)
+    if not articles:
+        print(f'未找到 {target_date} 的文章，请先运行 biz_daily.py')
+        return
+
+    print(f'找到 {len(articles)} 篇文章')
+
+    # Topic distribution
+    from collections import Counter
+    topics = Counter(a['topic'] for a in articles)
+    print(f'  主题: {dict(topics)}')
+
+    # Format articles for prompt
+    article_texts = []
+    for a in articles:
+        article_texts.append(f'- [{a["topic"]}] {a["title"]}（{a["source"]}）：{a["summary"][:150]}')
+    articles_text = '\n'.join(article_texts[: 30])  # max 30 articles
+
+    # Call AI
+    print(f'\n生成日报中...')
+    engine = create_engine(args.engine, api_key)
+    try:
+        response = engine.chat(DAILY_PROMPT.format(articles_text=articles_text), max_tokens=1000)
+    except Exception as e:
+        print(f'[ERR] AI 调用失败: {e}')
+        sys.exit(1)
+
+    # Parse response
+    focus_m = re.search(r'【今日焦点】\s*(.+?)(?=\n【|$)', response, re.DOTALL)
+    takeaway_m = re.search(r'【关键收获】\s*(.+?)(?=\n【|$)', response, re.DOTALL)
+
+    focus = focus_m.group(1).strip() if focus_m else ''
+    takeaway = takeaway_m.group(1).strip() if takeaway_m else ''
+
+    # Output
+    out_dir = Path(args.output) / 'Daily'
+    out_dir.mkdir(parents=True, exist_ok=True)
+    out_file = out_dir / f'Daily-{target_date}.md'
+
+    with open(out_file, 'w', encoding='utf-8') as f:
+        f.write(f'---\ntitle: "学习日报 — {target_date}"\ntype: daily-review\ndate: {target_date}\ncreated: {target_date}\n---\n\n')
+        f.write(f'# 学习日报 — {target_date}\n\n')
+        f.write(f'> 共 {len(articles)} 篇文章 | 主题: {", ".join(f"{t}({c})" for t, c in topics.most_common())}\n\n')
+        f.write(f'---\n\n')
+        if focus:
+            f.write(f'## 今日焦点\n\n{focus}\n\n')
+        # Topic breakdown
+        f.write(f'## 知识输入\n\n')
+        for topic, count in topics.most_common():
+            topic_articles = [a for a in articles if a['topic'] == topic]
+            f.write(f'### {topic}（{count}篇）\n\n')
+            for a in topic_articles[:8]:
+                f.write(f'- **{a["title"]}** — {a["source"]}\n')
+                if a['summary']:
+                    f.write(f'  {a["summary"][:200]}\n')
+            f.write('\n')
+        f.write(f'---\n\n')
+        if takeaway:
+            f.write(f'## 关键收获\n\n> {takeaway}\n\n')
+        f.write(f'*由 weflow-cli generate-review 自动生成*\n')
+
+    print(f'\n✓ 日报: {out_file}')
+    print(f'  焦点: {focus[:100] if focus else "(无)"}...')
+
+
+if __name__ == '__main__':
+    main()

package/scripts/mcp_bridge.py

@@ -0,0 +1,398 @@
+#!/usr/bin/env python3
+"""
+MCP Bridge — 为 MCP Server 提供微信数据库查询能力。
+
+用法（由 MCP Server 通过 child_process 调用）:
+  python scripts/mcp_bridge.py list_sessions [--limit 30]
+  python scripts/mcp_bridge.py search_messages --keyword "遥感" --talker "张三" --days 30 --limit 20
+  python scripts/mcp_bridge.py get_todos [--days 30]
+  python scripts/mcp_bridge.py get_action_suggestions [--date 2026-05-15]
+  python scripts/mcp_bridge.py get_chat_stats [--period week]
+
+输出: JSON 格式
+"""
+
+import sys, os, json, hashlib, argparse
+from datetime import datetime, timezone, timedelta
+from pathlib import Path
+
+try:
+    from sqlcipher3 import dbapi2 as sqlcipher
+except ImportError:
+    print(json.dumps({"error": "请安装: pip install sqlcipher3"}))
+    sys.exit(1)
+
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+from _utils import load_config, decrypt_lock, parse_frontmatter
+
+OUTPUT_ROOT = 'output'
+TZ = timezone(timedelta(hours=8))
+
+
+def json_output(data):
+    """Output as JSON."""
+    print(json.dumps(data, ensure_ascii=False, indent=2))
+
+
+# ====== Database ======
+
+def open_db(db_path, key_hex, salt_hex):
+    raw_key = f"x'{key_hex}{salt_hex}'"
+    conn = sqlcipher.connect(db_path)
+    c = conn.cursor()
+    c.execute(f'PRAGMA key = "{raw_key}";')
+    c.execute("SELECT count(*) FROM sqlite_master")
+    return conn
+
+
+def get_name_map(contact_db, contact_key, contact_salt):
+    name_map = {}
+    if not contact_db or not contact_key or not os.path.exists(contact_db):
+        return name_map
+    try:
+        conn = open_db(contact_db, contact_key, contact_salt)
+        c = conn.cursor()
+        c.execute("SELECT username, COALESCE(NULLIF(remark,''), NULLIF(nick_name,''), username) FROM contact")
+        for r in c.fetchall():
+            name_map[r[0]] = r[1]
+        conn.close()
+    except:
+        pass
+    return name_map
+
+
+def get_biz_keys(config):
+    nt_db = config.get('ntDbPath', '')
+    msg_dir = os.path.dirname(nt_db.replace('\\', '/'))
+    wxid_dir = os.path.dirname(os.path.dirname(msg_dir))
+    contact_db = os.path.join(wxid_dir, 'db_storage', 'contact', 'contact.db')
+    contact_key_enc = config.get('contactKey', '')
+    contact_salt = config.get('contactSalt', '')
+    contact_key = decrypt_lock(contact_key_enc) if contact_key_enc else ''
+    biz_db = os.path.join(msg_dir, 'biz_message_0.db')
+    biz_key_enc = config.get('bizKey', '')
+    biz_salt = config.get('bizSalt', '')
+    if biz_key_enc and biz_salt:
+        biz_key = decrypt_lock(biz_key_enc)
+    else:
+        return {'error': '缺少 biz_message_0.db 密钥，请运行 python scripts/nt_decrypt.py scan --json'}
+    return {
+        'biz_db': biz_db, 'biz_key': biz_key, 'biz_salt': biz_salt,
+        'contact_db': contact_db, 'contact_key': contact_key, 'contact_salt': contact_salt,
+    }
+
+
+def load_dbs():
+    """Load all databases and return connections + metadata."""
+    config = load_config()
+    nt_db = config.get('ntDbPath', '')
+    if not nt_db:
+        return None, None, None, None, {"error": "未初始化，请先运行 weflow-cli init"}
+
+    nt_key = decrypt_lock(config.get('ntKey', ''))
+    nt_salt = config.get('ntSalt', '')
+    keys = get_biz_keys(config)
+    name_map = get_name_map(keys['contact_db'], keys['contact_key'], keys['contact_salt'])
+
+    try:
+        conn = open_db(nt_db, nt_key, nt_salt)
+    except Exception as e:
+        return None, None, None, None, {"error": f"数据库连接失败: {e}"}
+
+    return conn, keys, name_map, nt_db, None
+
+
+# ====== Commands ======
+
+def cmd_list_sessions(args):
+    """List recent chat sessions."""
+    conn, keys, name_map, nt_db, err = load_dbs()
+    if err:
+        return err
+
+    c = conn.cursor()
+    limit = args.limit or 30
+    now = datetime.now(TZ)
+    start_ts = int((now - timedelta(days=90)).timestamp())
+    end_ts = int(now.timestamp())
+
+    try:
+        c.execute("SELECT user_name FROM Name2Id WHERE is_session = 1")
+        sessions = [r[0] for r in c.fetchall()]
+    except:
+        c.execute("SELECT name FROM sqlite_master WHERE type='table' AND name LIKE 'Msg_%'")
+        sessions = [r[0] for r in c.fetchall()]
+
+    results = []
+    for talker in sessions:
+        tbl = f"Msg_{hashlib.md5(talker.encode()).hexdigest()}"
+        try:
+            c.execute(f'SELECT COUNT(*), MAX(create_time) FROM "{tbl}" WHERE create_time >= ? AND create_time < ?', (start_ts, end_ts))
+            count, last_ts = c.fetchone()
+            if count > 0:
+                name = name_map.get(talker, talker)
+                last_time = datetime.fromtimestamp(last_ts, tz=TZ).strftime('%m-%d %H:%M')
+                results.append({
+                    "name": name,
+                    "wxid": talker,
+                    "count": count,
+                    "last_active": last_time,
+                })
+        except:
+            pass
+
+    conn.close()
+    results.sort(key=lambda x: x['count'], reverse=True)
+    return {"sessions": results[:limit], "total": len(results)}
+
+
+def cmd_search_messages(args):
+    """Search messages by keyword."""
+    conn, keys, name_map, nt_db, err = load_dbs()
+    if err:
+        return err
+
+    keyword = args.keyword or ''
+    talker_query = args.talker or ''
+    days = args.days or 30
+    limit = args.limit or 20
+
+    now = datetime.now(TZ)
+    start_ts = int((now - timedelta(days=days)).timestamp())
+    end_ts = int(now.timestamp())
+
+    c = conn.cursor()
+    try:
+        c.execute("SELECT user_name FROM Name2Id WHERE is_session = 1")
+        sessions = [r[0] for r in c.fetchall()]
+    except:
+        c.execute("SELECT name FROM sqlite_master WHERE type='table' AND name LIKE 'Msg_%'")
+        sessions = [r[0] for r in c.fetchall()]
+
+    # Filter by talker if specified
+    if talker_query:
+        filtered = []
+        for s in sessions:
+            name = name_map.get(s, s)
+            if talker_query in name or talker_query in s:
+                filtered.append(s)
+        sessions = filtered
+        if not sessions:
+            conn.close()
+            return {"results": [], "message": f"未找到联系人: {talker_query}"}
+
+    results = []
+    for talker in sessions[:50]:  # Limit scan scope
+        tbl = f"Msg_{hashlib.md5(talker.encode()).hexdigest()}"
+        try:
+            c.execute(f'SELECT COUNT(*) FROM sqlite_master WHERE name="{tbl}"')
+            if c.fetchone()[0] == 0:
+                continue
+
+            c.execute(f'''
+                SELECT create_time, real_sender_id, message_content
+                FROM "{tbl}"
+                WHERE create_time >= ? AND create_time < ?
+                ORDER BY create_time DESC
+            ''', (start_ts, end_ts))
+
+            for ts, sender, content in c.fetchall():
+                if not content or not isinstance(content, str):
+                    continue
+                if keyword and keyword.lower() not in content.lower():
+                    continue
+                sender_name = name_map.get(sender, sender) if sender else name_map.get(talker, talker)
+                dt = datetime.fromtimestamp(ts, tz=TZ)
+                results.append({
+                    "talker": name_map.get(talker, talker),
+                    "sender": sender_name,
+                    "time": dt.strftime('%Y-%m-%d %H:%M'),
+                    "content": content[:500],
+                })
+                if len(results) >= limit:
+                    break
+        except:
+            pass
+        if len(results) >= limit:
+            break
+
+    conn.close()
+    results.sort(key=lambda x: x['time'], reverse=True)
+    return {"results": results, "total": len(results), "keyword": keyword, "days": days}
+
+
+def cmd_get_todos(args):
+    """Get existing todo files."""
+    days = args.days or 30
+    todo_dir = Path(OUTPUT_ROOT)
+    results = []
+
+    # Find all 待办_*.md files
+    for f in sorted(todo_dir.glob('待办_*.md'), reverse=True):
+        content = f.read_text(encoding='utf-8')
+        # Extract todo items
+        items = []
+        for line in content.split('\n'):
+            line = line.strip()
+            if line.startswith('- [ ]') or line.startswith('- [x]'):
+                items.append(line)
+        if items:
+            results.append({
+                "file": f.name,
+                "items": items,
+                "count": len(items),
+            })
+
+    if not results:
+        return {"todos": [], "message": "暂无待办事项文件，请先运行 extract_todos.py"}
+
+    return {"todos": results}
+
+
+def cmd_get_action_suggestions(args):
+    """Get action suggestion files."""
+    date = args.date or ''
+    daily_dir = Path(OUTPUT_ROOT) / 'biz-daily'
+
+    if not daily_dir.exists():
+        return {"suggestions": [], "message": "日报目录不存在"}
+
+    if date:
+        # Specific date
+        action_file = daily_dir / date / '行动建议.md'
+        if action_file.exists():
+            content = action_file.read_text(encoding='utf-8')
+            return {"date": date, "content": content}
+        return {"suggestions": [], "message": f"未找到 {date} 的行动建议"}
+    else:
+        # Latest
+        dates = sorted([d.name for d in daily_dir.iterdir() if d.is_dir()], reverse=True)
+        for d in dates[:7]:  # Check last 7 days
+            action_file = daily_dir / d / '行动建议.md'
+            if action_file.exists():
+                content = action_file.read_text(encoding='utf-8')
+                return {"date": d, "content": content}
+        return {"suggestions": [], "message": "最近 7 天无行动建议"}
+
+
+def cmd_get_chat_stats(args):
+    """Get chat statistics."""
+    conn, keys, name_map, nt_db, err = load_dbs()
+    if err:
+        return err
+
+    period = args.period or 'week'
+    now = datetime.now(TZ)
+
+    if period == 'week':
+        days_since_monday = now.weekday()
+        start_date = (now - timedelta(days=days_since_monday)).replace(hour=0, minute=0, second=0, microsecond=0)
+    else:
+        start_date = now.replace(day=1, hour=0, minute=0, second=0, microsecond=0)
+
+    start_ts = int(start_date.timestamp())
+    end_ts = int(now.timestamp())
+
+    c = conn.cursor()
+    try:
+        c.execute("SELECT user_name FROM Name2Id WHERE is_session = 1")
+        sessions = [r[0] for r in c.fetchall()]
+    except:
+        c.execute("SELECT name FROM sqlite_master WHERE type='table' AND name LIKE 'Msg_%'")
+        sessions = [r[0] for r in c.fetchall()]
+
+    total_sent = 0
+    total_recv = 0
+    talker_counts = {}
+    hour_dist = [0] * 24
+
+    for talker in sessions:
+        tbl = f"Msg_{hashlib.md5(talker.encode()).hexdigest()}"
+        try:
+            c.execute(f'SELECT COUNT(*) FROM sqlite_master WHERE name="{tbl}"')
+            if c.fetchone()[0] == 0:
+                continue
+            c.execute(f'SELECT create_time FROM "{tbl}" WHERE create_time >= ? AND create_time < ?', (start_ts, end_ts))
+            count = 0
+            for (ts,) in c.fetchall():
+                count += 1
+                dt = datetime.fromtimestamp(ts, tz=TZ)
+                hour_dist[dt.hour] += 1
+            if count > 0:
+                name = name_map.get(talker, talker)
+                talker_counts[name] = talker_counts.get(name, 0) + count
+                total_recv += count  # Approximate
+        except:
+            pass
+
+    conn.close()
+
+    # Top talkers
+    top_talkers = sorted(talker_counts.items(), key=lambda x: x[1], reverse=True)[:10]
+
+    # Most active hour
+    most_active_hour = max(range(24), key=lambda h: hour_dist[h])
+
+    return {
+        "period": period,
+        "start_date": start_date.strftime('%Y-%m-%d'),
+        "end_date": now.strftime('%Y-%m-%d'),
+        "total_messages": sum(talker_counts.values()),
+        "active_contacts": len(talker_counts),
+        "top_talkers": [{"name": n, "count": c} for n, c in top_talkers],
+        "most_active_hour": f"{most_active_hour:02d}:00",
+        "hour_distribution": {f"{h:02d}:00": hour_dist[h] for h in range(24)},
+    }
+
+
+# ====== Main ======
+
+def main():
+    sys.stdout.reconfigure(encoding='utf-8', errors='replace')
+    parser = argparse.ArgumentParser()
+    subparsers = parser.add_subparsers(dest='command')
+
+    # list_sessions
+    p = subparsers.add_parser('list_sessions')
+    p.add_argument('--limit', type=int, default=30)
+
+    # search_messages
+    p = subparsers.add_parser('search_messages')
+    p.add_argument('--keyword', default='')
+    p.add_argument('--talker', default='')
+    p.add_argument('--days', type=int, default=30)
+    p.add_argument('--limit', type=int, default=20)
+
+    # get_todos
+    p = subparsers.add_parser('get_todos')
+    p.add_argument('--days', type=int, default=30)
+
+    # get_action_suggestions
+    p = subparsers.add_parser('get_action_suggestions')
+    p.add_argument('--date', default='')
+
+    # get_chat_stats
+    p = subparsers.add_parser('get_chat_stats')
+    p.add_argument('--period', choices=['week', 'month'], default='week')
+
+    args = parser.parse_args()
+
+    handlers = {
+        'list_sessions': cmd_list_sessions,
+        'search_messages': cmd_search_messages,
+        'get_todos': cmd_get_todos,
+        'get_action_suggestions': cmd_get_action_suggestions,
+        'get_chat_stats': cmd_get_chat_stats,
+    }
+
+    handler = handlers.get(args.command)
+    if not handler:
+        json_output({"error": f"未知命令: {args.command}", "available": list(handlers.keys())})
+        sys.exit(1)
+
+    result = handler(args)
+    json_output(result)
+
+
+if __name__ == '__main__':
+    main()