tamakyi/tamabox-broadcast-mailsystem
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

退信核查记住扫描位置：UID 游标续扫，不再每次从头拉全量头部

Browse Source 
- broadcast_bounces_seen.json 扩展为 {seen, cursor}（兼容旧列表格式）：
  cursor 记录 INBOX 的 uidvalidity + last_uid
- 下次 --check-bounces 默认 UID n+1:* 续扫，只查新到的邮件；随邮箱
  邮件增多不再越扫越慢
- 回退全扫的情形：首次核查（无记录）、UIDVALIDITY 变化（换号/重建邮箱）、
  显式 --since、新增 --full-scan
- 扫描循环统一走 UID 检索/取信；UID n:* 在 n 超过邮箱最大 UID 时会返回
  最后一封，已防御过滤，游标不回退、不误报
- 发送中巡检不推进扫描位置（not_before 会过滤历史邮件，推进会把未处理
  的邮件标记成已扫导致漏报）；位置只由 --check-bounces 推进
- 退信报告新增 scan_mode / scan_cursor 字段
This commit is contained in:
tamakyi
2026-09-07 22:53:35 +08:00
parent 6dc33604d1
commit d4e5c611b0
2 changed files with 121 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files
broadcast.py
+110 -26
View File
@@ -1249,18 +1249,59 @@ def _parse_mail_date(value):
return dt
def _load_bounce_seen():
"""读取已核查记录文件（DEFAULT_BOUNCE_SEEN）。
旧格式：Message-ID 列表；新格式：{"seen": [...], "cursor": {...}}。
返回 (seen 集合, cursor dict 或 None)。
cursor = {"uidvalidity": int, "last_uid": int}，即上次扫描在 INBOX 里
扫到的位置；UIDVALIDITY 对不上（邮箱重建/换号）时游标作废回退全扫。
"""
if not os.path.isfile(DEFAULT_BOUNCE_SEEN):
return set(), None
try:
with open(DEFAULT_BOUNCE_SEEN, "r", encoding="utf-8") as f:
data = json.load(f)
except Exception: # noqa: BLE001
return set(), None
if isinstance(data, dict):
seen = set(data.get("seen") or [])
cur = data.get("cursor")
if (isinstance(cur, dict)
and isinstance(cur.get("uidvalidity"), int)
and isinstance(cur.get("last_uid"), int)):
return seen, cur
return seen, None
return set(data or []), None
def _save_bounce_seen(seen, cursor=None):
try:
with open(DEFAULT_BOUNCE_SEEN, "w", encoding="utf-8") as f:
json.dump({"seen": sorted(seen), "cursor": cursor}, f, ensure_ascii=False)
except Exception as e: # noqa: BLE001
print(f"[警告] 记录已核查邮件/扫描位置失败（不影响本次结果）：{e}")
def _scan_imap_bounces(cfg, args, start_dt, skip_ids=None, not_before=None,
start_desc="", verbose=True):
start_desc="", verbose=True, resume=None):
"""连 IMAP 扫描退信并逐收件人分类（check_bounces 与发送中巡检共用）。
skip_ids 已处理过的 Message-ID（seen 文件 / 本次运行内已见过），跳过不重复报
not_before 只统计 Date 头晚于该时刻的退信（发送中巡检用：只看本次运行之后
新到的退信，避免拿历史退信误触发中止）；Date 解析不了则放行
返回 dict：bounced / new_ids / scanned / touched_ids
resume 上次扫描位置 {"uidvalidity": int, "last_uid": int}；UIDVALIDITY
一致时从 last_uid 之后续扫（UID 检索），否则回退按日期窗口全扫
（全新 / 邮箱 UID 代际变更 / 显式 --since / --full-scan）。
只由 check_bounces 传入推进——发送中巡检的 not_before 会过滤掉
范围内的历史邮件，若推进游标会把没处理过的邮件标记成「已扫」。
返回 dict：bounced / new_ids / scanned / touched_ids / last_uid / uidvalidity
bounced {email: {"subject","date","category","reason","detail"}}
new_ids 本次计入的退信 Message-ID（是否落盘由调用方决定）
touched_ids 本次扫到的所有退信 Message-ID（含被 not_before 过滤掉的），
供巡检做运行期去重，避免每轮重复拉取同一封全文
last_uid 本次检查到的最大 UID（供调用方更新扫描位置）
uidvalidity 当前邮箱的 UIDVALIDITY（读不到为 None，调用方据此不落游标）
"""
smtp_host = cfg["mail_smtp"] or ""
imap_host = (args.imap_host or "").strip()
@@ -1280,33 +1321,58 @@ def _scan_imap_bounces(cfg, args, start_dt, skip_ids=None, not_before=None,
try:
conn.login(account, password)
conn.select("INBOX", readonly=True)
typ, data = conn.search(None, f'(SINCE "{since}")')
try:
uv = int(conn.response("UIDVALIDITY")[1][0])
except Exception: # noqa: BLE001
uv = None
resume_uid = None
if resume:
if uv is not None and uv == resume.get("uidvalidity"):
resume_uid = int(resume["last_uid"])
if verbose:
print(f" [扫描位置] 续扫：从 UID {resume_uid} 之后开始"
f"（uidvalidity {uv}）")
elif verbose:
print(" [扫描位置] 邮箱 UIDVALIDITY 已变化（或无法读取），"
"本次回退按日期窗口全量扫描")
if resume_uid is not None:
typ, data = conn.uid("SEARCH", None, f"UID {resume_uid + 1}:*")
else:
typ, data = conn.uid("SEARCH", None, f'(SINCE "{since}")')
if typ != "OK":
raise RuntimeError("IMAP search 失败")
ids = data[0].split()
if verbose:
print(f" [IMAP] 扫描范围内共 {len(ids)} 封待扫描")
print(f" [IMAP] 扫描范围内共 {len(ids)} 封"
f"（{'续扫' if resume_uid is not None else '日期窗口'}）")
skip_ids = skip_ids or set()
bounced, new_ids, touched, scanned = {}, [], [], 0
last_uid = resume_uid or 0
for i, mid in enumerate(ids, 1):
uid = mid.decode() if isinstance(mid, bytes) else str(mid)
uid = int(mid)
if uid <= last_uid:
# 防御：UID n:* 在 n 超过邮箱最大 UID 时会返回最后一封
continue
last_uid = uid
ustr = str(uid)
# 两段式：先取头部（省流量），命中退信特征再取全文
typ, hdata = conn.fetch(
mid, "(BODY.PEEK[HEADER.FIELDS (SUBJECT MESSAGE-ID FROM CONTENT-TYPE)])")
typ, hdata = conn.uid(
"fetch", ustr,
"(BODY.PEEK[HEADER.FIELDS (SUBJECT MESSAGE-ID FROM CONTENT-TYPE)])")
if typ != "OK" or not hdata or hdata[0] is None:
continue
try:
head = email.message_from_bytes(hdata[0][1])
except Exception: # noqa: BLE001
continue
msg_id = (head.get("Message-ID") or "").strip() or f"uid:{uid}"
msg_id = (head.get("Message-ID") or "").strip() or f"uid:{ustr}"
if msg_id in skip_ids:
continue
scanned += 1
if not is_bounce_message(head):
continue
typ, fdata = conn.fetch(mid, "(RFC822)")
typ, fdata = conn.uid("fetch", ustr, "(RFC822)")
if typ != "OK" or not fdata or fdata[0] is None:
continue
try:
@@ -1333,7 +1399,8 @@ def _scan_imap_bounces(cfg, args, start_dt, skip_ids=None, not_before=None,
for r in rcpts) if rcpts else "(未解析出收件人)"
print(f" [退信 {i}/{len(ids)}] {subject[:40]} → {desc}")
return {"bounced": bounced, "new_ids": new_ids,
"scanned": scanned, "touched_ids": touched}
"scanned": scanned, "touched_ids": touched,
"last_uid": last_uid, "uidvalidity": uv}
finally:
try:
conn.logout()
@@ -1342,29 +1409,43 @@ def _scan_imap_bounces(cfg, args, start_dt, skip_ids=None, not_before=None,
def check_bounces(cfg, args):
"""登录发件邮箱 IMAP，扫描近 N 天的退信并解析失败收件人。
"""登录发件邮箱 IMAP，扫描退信并解析失败收件人。
返回 {email: {"subject","date","category","reason","detail"}}；已处理过的
邮件（Message-ID）记录在 broadcast_bounces_seen.json，重复核查不会重复报告。
邮件（Message-ID）与上次扫描位置（UID 游标）都记录在
broadcast_bounces_seen.json：默认从上次位置续扫，只查新到的邮件，
游标不存在/作废、显式 --since 或 --full-scan 时才按日期窗口从头扫。
category 见 CAT_RATE / CAT_HARD / CAT_UNKNOWN。"""
start_dt, start_desc = resolve_scan_start(args)
seen = set()
if os.path.isfile(DEFAULT_BOUNCE_SEEN):
try:
with open(DEFAULT_BOUNCE_SEEN, "r", encoding="utf-8") as f:
seen = set(json.load(f))
except Exception: # noqa: BLE001
seen = set()
seen, cursor = _load_bounce_seen()
since_explicit = ((getattr(args, "since", "auto") or "auto").strip().lower()
not in ("", "auto"))
full_scan = bool(getattr(args, "full_scan", False))
resume = None
if full_scan:
print("[扫描位置] --full-scan：忽略上次扫描位置，从头全量扫描")
elif since_explicit:
print("[扫描位置] 显式 --since 指定窗口，本次忽略上次扫描位置")
elif cursor is None:
print("[扫描位置] 全新扫描（尚无上次位置记录），按日期窗口全扫")
else:
resume = cursor
print(f"[扫描位置] 从上次扫到的 UID {cursor['last_uid']} 之后续扫"
f"（uidvalidity {cursor['uidvalidity']}）")
try:
res = _scan_imap_bounces(cfg, args, start_dt, skip_ids=seen, start_desc=start_desc)
res = _scan_imap_bounces(cfg, args, start_dt, skip_ids=seen,
start_desc=start_desc, resume=resume)
except Exception as e: # noqa: BLE001
sys.exit(f"[FATAL] IMAP 扫描失败（请确认邮箱已开启 IMAP，密码用邮箱登录密码）: {e}")
bounced = res["bounced"]
newly_seen, scanned = res["new_ids"], res["scanned"]
bounce_cnt = len(newly_seen)
scan_mode = "resume" if resume is not None else "full"
scan_cursor = ({"uidvalidity": res["uidvalidity"], "last_uid": res["last_uid"]}
if res.get("uidvalidity") is not None and res.get("last_uid") else None)
# ---- 按类别分组处理：只有「可重试」的才剔除补发 ----
groups = {CAT_RATE: [], CAT_HARD: [], CAT_UNKNOWN: []}
@@ -1428,6 +1509,8 @@ def check_bounces(cfg, args):
"script_version": SCRIPT_VERSION,
"scan_from": start_dt.strftime("%Y-%m-%d %H:%M:%S"),
"scan_from_desc": start_desc,
"scan_mode": scan_mode,
"scan_cursor": scan_cursor,
"scanned": scanned,
"bounce_messages": bounce_cnt,
"counts": {c: len(groups.get(c, [])) for c in (CAT_RATE, CAT_HARD, CAT_UNKNOWN)},
@@ -1437,12 +1520,10 @@ def check_bounces(cfg, args):
except Exception as e: # noqa: BLE001
print(f"[警告] 写入退信报告失败（不影响本次结果）：{e}")
if newly_seen:
try:
with open(DEFAULT_BOUNCE_SEEN, "w", encoding="utf-8") as f:
json.dump(sorted(set(newly_seen) | seen), f, ensure_ascii=False)
except Exception as e: # noqa: BLE001
print(f"[警告] 记录已核查邮件失败（不影响本次结果）：{e}")
# 落盘：已核查 Message-ID + 新的扫描位置
# （本次读不到 uidvalidity 时沿用旧游标：只可能往回多扫几封头部，不会漏）
if newly_seen or scan_cursor or cursor:
_save_bounce_seen(set(newly_seen) | seen, scan_cursor or cursor)
return bounced
@@ -1845,6 +1926,9 @@ def build_arg_parser():
parser.add_argument("--prune-unknown", action="store_true", default=None,
help="配合 --check-bounces：无法分类的退信也按「可重试」处理，一并剔除补发"
"（默认不剔除，避免给死信地址反复重发；也可在 params.ini [bounce] 设置）")
parser.add_argument("--full-scan", action="store_true",
help="退信核查忽略上次扫描位置，按日期窗口从头全量扫描"
"（默认记住上次扫到的位置只查新邮件；显式 --since 也会忽略位置）")
parser.add_argument("--invalid-file", default=None, metavar="FILE",
help=f"无效地址清单（硬退信地址）路径；默认脚本同目录 broadcast_invalid.json。"
f"清单里的地址每次运行都会跳过，即使 --reset-state")