178 lines
5.9 KiB
Python
178 lines
5.9 KiB
Python
import imaplib
|
|
import email
|
|
import logging
|
|
from email.header import decode_header
|
|
from email.utils import parsedate_to_datetime, parseaddr
|
|
from typing import Optional
|
|
from src.config import EmailAccount
|
|
from src.retry import retry
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
class Email:
|
|
def __init__(self, uid: bytes, subject: str, sender: str, recipient: str,
|
|
body: str, date: str, reply_to: str = "", account_email: str = ""):
|
|
self.uid = uid
|
|
self.subject = subject
|
|
self.sender = sender
|
|
self.recipient = recipient
|
|
self.body = body
|
|
self.date = date
|
|
self.reply_to = reply_to
|
|
self.account_email = account_email
|
|
|
|
|
|
def _decode_str(s: str) -> str:
|
|
parts = decode_header(s)
|
|
result = []
|
|
for part, charset in parts:
|
|
if isinstance(part, bytes):
|
|
try:
|
|
result.append(part.decode(charset or "utf-8", errors="replace"))
|
|
except LookupError:
|
|
result.append(part.decode("utf-8", errors="replace"))
|
|
else:
|
|
result.append(part)
|
|
return "".join(result)
|
|
|
|
|
|
def _decode_address_header(header_value: str) -> str:
|
|
if not header_value:
|
|
return ""
|
|
decoded = _decode_str(header_value)
|
|
name, addr = parseaddr(decoded)
|
|
if not addr:
|
|
return name
|
|
if name:
|
|
return f"{name} <{addr}>"
|
|
return addr
|
|
|
|
|
|
def _get_text_from_msg(msg) -> str:
|
|
if msg.is_multipart():
|
|
for part in msg.walk():
|
|
content_type = part.get_content_type()
|
|
if content_type == "text/plain":
|
|
payload = part.get_payload(decode=True)
|
|
if payload:
|
|
return payload.decode("utf-8", errors="replace")
|
|
elif content_type == "text/html":
|
|
payload = part.get_payload(decode=True)
|
|
if payload:
|
|
return payload.decode("utf-8", errors="replace")
|
|
else:
|
|
payload = msg.get_payload(decode=True)
|
|
if payload:
|
|
return payload.decode("utf-8", errors="replace")
|
|
return ""
|
|
|
|
|
|
_PROVIDERS_NEED_ID = ("163.com", "126.com")
|
|
|
|
|
|
def _check_provider(email_addr: str, providers: tuple[str, ...]) -> bool:
|
|
return any(email_addr.endswith(p) for p in providers)
|
|
|
|
|
|
def _send_id_command(conn):
|
|
imaplib.Commands["ID"] = "AUTH"
|
|
args = ("name", "AI-Mail-Bot", "version", "1.0.0", "vendor", "personal")
|
|
conn._simple_command("ID", '("' + '" "'.join(args) + '")')
|
|
|
|
|
|
def _select_mailbox(conn, mailbox: str = "INBOX"):
|
|
status, data = conn.select(mailbox)
|
|
if status != "OK":
|
|
raise RuntimeError(f"无法选择邮箱 {mailbox}: {data}")
|
|
|
|
|
|
def _login_and_prepare(account: EmailAccount, timeout: int = 30):
|
|
logger.info("连接 %s:%d (超时%ds)", account.imap_server, account.imap_port, timeout)
|
|
if account.use_ssl:
|
|
conn = imaplib.IMAP4_SSL(account.imap_server, account.imap_port, timeout=timeout)
|
|
else:
|
|
conn = imaplib.IMAP4(account.imap_server, account.imap_port, timeout=timeout)
|
|
if account.use_starttls:
|
|
conn.starttls()
|
|
conn.login(account.username, account.password)
|
|
logger.info("登录成功: %s", account.username)
|
|
if _check_provider(account.username, _PROVIDERS_NEED_ID):
|
|
logger.info("发送 ID 命令 (126/163 兼容)")
|
|
_send_id_command(conn)
|
|
_select_mailbox(conn)
|
|
logger.info("已选择 INBOX")
|
|
return conn
|
|
|
|
|
|
def fetch_unseen_emails(account: EmailAccount) -> list[Email]:
|
|
conn = _login_and_prepare(account)
|
|
|
|
_, data = conn.uid("SEARCH", None, "UNSEEN")
|
|
uids = data[0].split() if data[0] else []
|
|
logger.info("UNSEEN 数量: %d", len(uids))
|
|
emails = []
|
|
|
|
for uid in uids:
|
|
try:
|
|
_, size_data = conn.uid("FETCH", uid, "(RFC822.SIZE)")
|
|
size = 0
|
|
if size_data[0]:
|
|
m = re.search(rb"RFC822\.SIZE (\d+)", size_data[0])
|
|
if m:
|
|
size = int(m.group(1))
|
|
large = size > 200_000
|
|
if large:
|
|
logger.info("UID %s 大邮件 (%d bytes),用 PEEK 拉取", uid, size)
|
|
|
|
fetch_cmd = "BODY.PEEK[]" if large else "RFC822"
|
|
_, msg_data = conn.uid("FETCH", uid, fetch_cmd)
|
|
if msg_data[0] is None:
|
|
logger.warning("UID %s FETCH 返回空,跳过", uid)
|
|
continue
|
|
raw_email = msg_data[0][1]
|
|
msg = email.message_from_bytes(raw_email)
|
|
|
|
subject = _decode_str(msg.get("Subject", ""))
|
|
sender = _decode_address_header(msg.get("From", ""))
|
|
recipient = _decode_address_header(msg.get("To", ""))
|
|
reply_to = _decode_address_header(msg.get("Reply-To", ""))
|
|
date_str = msg.get("Date", "")
|
|
body = _get_text_from_msg(msg)
|
|
body_len = len(body)
|
|
|
|
logger.info(" 邮件: [%s] from=%s to=%s reply-to=%s (%d 字符)", subject, sender, recipient, reply_to, body_len)
|
|
emails.append(Email(uid=uid, subject=subject, sender=sender, recipient=recipient,
|
|
body=body, date=date_str, reply_to=reply_to, account_email=account.username))
|
|
except Exception as e:
|
|
logger.error("UID %s FETCH 失败,跳过并重建连接: %s", uid, e)
|
|
try:
|
|
conn.logout()
|
|
except Exception:
|
|
pass
|
|
try:
|
|
conn = _login_and_prepare(account)
|
|
except Exception as e2:
|
|
logger.error("重建连接失败: %s", e2)
|
|
break
|
|
continue
|
|
|
|
try:
|
|
conn.logout()
|
|
except Exception:
|
|
pass
|
|
logger.info("共获取 %d 封新邮件", len(emails))
|
|
return emails
|
|
|
|
|
|
@retry()
|
|
def mark_as_seen(account: EmailAccount, uids: list[bytes]):
|
|
if not uids:
|
|
return
|
|
logger.info("标记 %d 封邮件为已读", len(uids))
|
|
conn = _login_and_prepare(account)
|
|
for uid in uids:
|
|
conn.uid("STORE", uid, "+FLAGS", "\\Seen")
|
|
conn.logout()
|
|
logger.info("标记完成")
|