inquiry_robot/inquiry-agent/agent/channel/wshoto/parse.py

"""
微盛群消息 → InboundMessage。

本文件职责：从管家 records 抽出 sender_id / chat_id / 正文 / 附件名。
禁止：把存档席「询价机器人」当成发言人入业务；禁止再向企微换 ID。
"""

from __future__ import annotations

import json
from typing import Any, Mapping, Optional, Sequence

from agent.channel.wecom.models import InboundMessage

# 存档席在群里的显示名；from 偶发不是 userid 时也要丢掉
SEAT_DISPLAY_NAMES = frozenset({"询价机器人"})
_SKIP_TYPES = frozenset({"revoke", "recall", "event", "unknown"})
_FILE_TYPES = frozenset({"file", "image", "video", "emotion", "voice"})


def _as_mapping(raw: Any) -> dict[str, Any]:
    if isinstance(raw, Mapping):
        return dict(raw)
    if isinstance(raw, str) and raw.strip().startswith("{"):
        try:
            parsed = json.loads(raw)
        except json.JSONDecodeError:
            return {}
        return dict(parsed) if isinstance(parsed, Mapping) else {}
    return {}


def _create_time(msgtime: Any) -> int:
    """微盛 msgtime 常是毫秒；InboundMessage.create_time 用秒。"""
    try:
        n = int(msgtime or 0)
    except (TypeError, ValueError):
        return 0
    if n > 10**12:
        return n // 1000
    return n


def _unwrap_text(raw: Any) -> str:
    """
    从字符串或嵌套对象取出正文。

    管家 content 经常是摘要；企微官方结构是 text.content。空、null 不当正文。
    """
    if raw is None:
        return ""
    if isinstance(raw, str):
        text = raw.strip()
        if not text or text.lower() in {"null", "none"}:
            return ""
        if text.startswith("{") or text.startswith("["):
            try:
                return _unwrap_text(json.loads(text))
            except json.JSONDecodeError:
                return text
        return text
    if isinstance(raw, Mapping):
        for key in ("content", "text", "msg"):
            got = _unwrap_text(raw.get(key))
            if got:
                return got
        return ""
    return str(raw).strip()


def _best_text(record: Mapping[str, Any], payload: Mapping[str, Any]) -> str:
    """多处正文里取最长的一份，避免摘要盖掉带金额的原文。"""
    candidates = [
        _unwrap_text(record.get("content")),
        _unwrap_text(payload.get("content")),
        _unwrap_text(payload.get("text")),
        _unwrap_text(payload.get("msg")),
    ]
    texts = [x for x in candidates if x]
    if not texts:
        return ""
    return max(texts, key=len)


def _filename(msg_type: str, content: str, payload: Mapping[str, Any]) -> str:
    for key in ("filename", "file_name", "FileName", "title", "name"):
        val = str(payload.get(key) or "").strip()
        if val:
            return val
    if msg_type in _FILE_TYPES:
        return str(content or "").strip()
    return ""


def archive_record_to_inbound(
    record: Mapping[str, Any],
    *,
    seat_userids: Sequence[str] = (),
) -> Optional[InboundMessage]:
    """
    一条管家记录变成群入站。

    返回 None：不是群、没有发言人、发言人是存档席、或撤回类消息。
    message_id 带 wshoto: 前缀，避免和回调 MsgId 撞车。
    """
    room_id = str(record.get("room_id") or record.get("roomid") or "").strip()
    sender_id = str(record.get("from") or record.get("from_id") or "").strip()
    if not room_id or not sender_id:
        return None
    seats = {str(x).strip() for x in seat_userids if str(x).strip()}
    if sender_id in seats or sender_id in SEAT_DISPLAY_NAMES:
        return None

    msg_type = str(record.get("msgtype") or record.get("msg_type") or "text").strip().lower()
    if msg_type in _SKIP_TYPES:
        return None

    payload = _as_mapping(record.get("msg_content") or record.get("msgcontent"))
    content = _best_text(record, payload)
    msgid = str(record.get("msgid") or record.get("msg_id") or "").strip()
    seq = record.get("seq")
    if not msgid:
        msgid = f"{room_id}:{seq or 0}:{sender_id}"

    media: dict[str, Any] = {}
    inbound_type = "text"
    if msg_type in _FILE_TYPES:
        inbound_type = "file" if msg_type == "file" else msg_type
        media = {
            "filename": _filename(msg_type, content, payload),
            "media_id": str(payload.get("sdkfileid") or payload.get("sdk_file_id") or ""),
            "file_ext": str(payload.get("fileext") or payload.get("file_ext") or ""),
            "content_type": "",
            "filesize": payload.get("filesize") or payload.get("file_size") or 0,
        }
        if not content:
            content = str(media.get("filename") or "")

    return InboundMessage(
        sender_id=sender_id,
        message_id=f"wshoto:{msgid}",
        content=content,
        msg_type=inbound_type,
        chat_id=room_id,
        chat_type="group",
        create_time=_create_time(record.get("msgtime") or record.get("msg_time")),
        media=media,
        raw={
            "source": "wshoto_archive",
            "seq": seq,
            "msgtype": msg_type,
            "direction": record.get("direction") or "",
        },
    )