feat: add chat group daily summarizer

- Telegram collector (Telethon MTProto) with checkpoint-based message fetching - Claude API summarization via proxy (haiku model) - Dashboard page with date picker, markdown summary, raw message drill-down - Separate container lifecycle from dashboard for stable Telethon sessions - Shared SQLite DB mounted read-only into dashboard
2026-02-27 02:51:52 +08:00
parent 7f94013bf9
commit c676c84bc1
14 changed files with 441 additions and 1 deletions
@@ -0,0 +1,80 @@
+import os
+import logging
+import aiosqlite
+from telethon import TelegramClient
+from db import DB_PATH, init_db
+
+log = logging.getLogger(__name__)
+
+API_ID = int(os.environ.get("TG_API_ID", 0))
+API_HASH = os.environ.get("TG_API_HASH", "")
+SESSION_NAME = os.environ.get("TG_SESSION_NAME", "chat_summarizer")
+GROUP_IDS = [int(g) for g in os.environ.get("TG_GROUP_IDS", "").split(",") if g.strip()]
+
+def get_client():
+    return TelegramClient(f"/app/data/{SESSION_NAME}", API_ID, API_HASH)
+
+async def auth():
+    """Interactive one-time auth. Run: docker exec -it chat-summarizer python3 -c 'import asyncio; from collector import auth; asyncio.run(auth())'"""
+    client = get_client()
+    await client.start()
+    me = await client.get_me()
+    print(f"Logged in as {me.first_name} (id={me.id})")
+    await client.disconnect()
+
+async def collect_messages():
+    if not GROUP_IDS:
+        log.warning("No TG_GROUP_IDS configured, skipping collection")
+        return
+
+    client = get_client()
+    await client.connect()
+    if not await client.is_user_authorized():
+        log.error("Not authorized — run interactive auth first")
+        return
+
+    try:
+        async with aiosqlite.connect(DB_PATH) as db:
+            for gid in GROUP_IDS:
+                try:
+                    # Get checkpoint
+                    cursor = await db.execute("SELECT last_msg_id FROM checkpoints WHERE group_id=?", (gid,))
+                    row = await cursor.fetchone()
+                    min_id = row[0] if row else 0
+
+                    entity = await client.get_entity(gid)
+                    group_name = getattr(entity, 'title', str(gid))
+                    count = 0
+
+                    async for msg in client.iter_messages(entity, min_id=min_id, limit=1000):
+                        if not msg.text:
+                            continue
+                        sender = ""
+                        if msg.sender:
+                            sender = getattr(msg.sender, 'first_name', '') or getattr(msg.sender, 'title', '') or ''
+                            last = getattr(msg.sender, 'last_name', '')
+                            if last:
+                                sender += f" {last}"
+
+                        await db.execute(
+                            "INSERT INTO messages (group_id, group_name, sender_name, text, msg_id, timestamp) VALUES (?,?,?,?,?,?)",
+                            (gid, group_name, sender, msg.text, msg.id, msg.date.isoformat())
+                        )
+                        count += 1
+
+                    if count > 0:
+                        # Update checkpoint to max msg_id
+                        cursor = await db.execute("SELECT MAX(msg_id) FROM messages WHERE group_id=?", (gid,))
+                        max_row = await cursor.fetchone()
+                        if max_row and max_row[0]:
+                            await db.execute(
+                                "INSERT OR REPLACE INTO checkpoints (group_id, last_msg_id) VALUES (?,?)",
+                                (gid, max_row[0])
+                            )
+                        log.info("Collected %d messages from %s", count, group_name)
+
+                    await db.commit()
+                except Exception as e:
+                    log.error("Error collecting from group %s: %s", gid, e)
+    finally:
+        await client.disconnect()