refactor: replace all in-memory state with SQLite database

main.py

@@ -9,6 +9,7 @@ from concurrent.futures import ThreadPoolExecutor
 from src.config import load_config
 from src.summarizer import poll_accounts, ai_process, tg_send_and_mark
 from src.tg_bot import poll_telegram, send_thinking
+from src.database import init_db, is_uids_processed, mark_fetched, mark_processing, mark_sent, mark_failed
 
 # Thread-safe logging: all log records go through a single QueueListener thread
 _log_queue = queue_module.Queue(-1)
@@ -29,9 +30,6 @@ logger = logging.getLogger("main")
 
 _running = True
 _shutdown_event = threading.Event()
-_pending_lock = threading.Lock()
-_pending_uids: set[bytes] = set()
-_processing_uids: set[bytes] = set()  # 正在处理（含等待TG发送）的邮件UID
 
 # Queue 1: raw emails → AI processor
 _email_queue: queue_module.Queue = queue_module.Queue()
@@ -50,19 +48,18 @@ def _email_poller(cfg):
     logger.info("邮件轮询线程已启动")
     while _running:
         try:
-            for acct_idx, mail in poll_accounts(cfg, skip_uids=_processing_uids):
+            for acct_idx, mail in poll_accounts(cfg):
                 if not _running:
                     return
-                with _pending_lock:
-                    if mail.uid not in _processing_uids:
-                        _processing_uids.add(mail.uid)
-                        _pending_uids.add(mail.uid)
-                        # 发送"获取邮件中"提示
-                        try:
-                            tg_msg_id = send_thinking(cfg.telegram, cfg.telegram.chat_id, "📥 获取邮件中...")
-                        except Exception:
-                            tg_msg_id = 0
-                        _email_queue.put((acct_idx, mail, tg_msg_id))
+                uid_str = mail.uid.decode("utf-8", errors="replace") if isinstance(mail.uid, bytes) else str(mail.uid)
+                if not is_uid_processed(uid_str):
+                    mark_fetched(uid_str, mail.account_email, mail.folder,
+                                 mail.subject, mail.sender)
+                    try:
+                        tg_msg_id = send_thinking(cfg.telegram, cfg.telegram.chat_id, "📥 获取邮件中…")
+                    except Exception:
+                        tg_msg_id = 0
+                    _email_queue.put((acct_idx, mail, tg_msg_id))
         except Exception as e:
             logger.error(f"邮件轮询线程异常: {e}", exc_info=True)
             time.sleep(5)
@@ -81,6 +78,8 @@ def _ai_processor(cfg):
                     acct_idx, mail, tg_msg_id = _email_queue.get_nowait()
                 except queue_module.Empty:
                     break
+                uid_str = mail.uid.decode("utf-8", errors="replace") if isinstance(mail.uid, bytes) else str(mail.uid)
+                mark_processing(uid_str)
                 future = executor.submit(ai_process, cfg, acct_idx, mail, tg_msg_id)
                 pending_futures[future] = (acct_idx, mail)
 
@@ -93,10 +92,8 @@ def _ai_processor(cfg):
                         _tg_queue.put(info)
                 except Exception as e:
                     logger.error(f"AI 处理邮件失败: {e}", exc_info=True)
-                    # AI 失败也要移除 pending，否则永远卡住
-                    with _pending_lock:
-                        _pending_uids.discard(mail.uid)
-                        _processing_uids.discard(mail.uid)
+                    uid_str = mail.uid.decode("utf-8", errors="replace") if isinstance(mail.uid, bytes) else str(mail.uid)
+                    mark_failed(uid_str)
 
             time.sleep(0.1 if pending_futures else 0.5)
 
@@ -114,16 +111,15 @@ def _tg_worker(cfg):
             try:
                 info = _tg_queue.get_nowait()
                 tg_send_and_mark(cfg, info)
-                # 发送并标记已读成功后移除 pending（_processing_uids 保留直到重启）
-                with _pending_lock:
-                    _pending_uids.discard(info["uid"])
+                uid_str = info["uid"].decode("utf-8", errors="replace") if isinstance(info["uid"], bytes) else str(info["uid"])
+                mark_sent(uid_str)
             except queue_module.Empty:
                 break
             except Exception as e:
                 logger.error(f"TG 发送失败: {e}", exc_info=True)
                 try:
-                    with _pending_lock:
-                        _pending_uids.discard(info["uid"])
+                    uid_str = info["uid"].decode("utf-8", errors="replace") if isinstance(info["uid"], bytes) else str(info["uid"])
+                    mark_failed(uid_str)
                 except Exception:
                     pass
 
@@ -138,6 +134,7 @@ def main():
 
     cfg_path = sys.argv[1] if len(sys.argv) > 1 else "config.yaml"
     cfg = load_config(cfg_path)
+    init_db()
     logger.info(f"AI邮件摘要机器人已启动，轮询间隔: {cfg.polling.interval_seconds}s")
 
     threads = [
@@ -155,8 +152,8 @@ def main():
         while _running:
             tick += 1
             if tick % 30 == 0:
-                logger.info("队列状态: email_queue=%d tg_queue=%d pending_uids=%d",
-                            _email_queue.qsize(), _tg_queue.qsize(), len(_pending_uids))
+                logger.info("队列状态: email_queue=%d tg_queue=%d",
+                            _email_queue.qsize(), _tg_queue.qsize())
             _shutdown_event.wait(1)
     except KeyboardInterrupt:
         _running = False