inquiry_robot/inquiry-agent/agent/schema/tms_land_query.py

"""
陆运查价参数：对照 TMS 出站是否带全。

本文件职责：核对卡通过后、建单查价前，确认当前运输类型的货物字段能拼出查价正文。
带去：运输类型、线路类别、始发/目的/品名/报价日期，以及该类型的体积重量或车型/数量；通关口岸有值才带。
出站类型/线路必须译成 TMS 6.1.1 英文枚举，线路键名 routeCategory。工单与卡片仍用中文。
中港整车不带车型/数量。车型/数量是一个字段，出站再拆。
禁止：把运输分类、中港装载方式放进出站；直连 TMS；改六态。
线程：纯函数。
"""

from __future__ import annotations

import logging
import re
from typing import Any

from agent.schema.field_validate import apply_quote_date_default, normalize_facts
from agent.schema.land_options import canonicalize_hk_load

logger = logging.getLogger(__name__)

LCL_TYPES = frozenset({"国内运输拼车", "跨境集拼"})
FTL_TYPES = frozenset({"国内运输整车", "跨境整车"})
CUSTOMS_TYPES = frozenset({"跨境集拼", "跨境整车", "中港"})

_TYPE_ENUM = {
    "国内运输拼车": "DOMESTIC_LTL",
    "国内运输整车": "DOMESTIC_FTL",
    "跨境集拼": "CROSS_BORDER_LTL",
    "跨境整车": "CROSS_BORDER_FTL",
    "中港": "CHINA_HK",
}
_ROUTE_ENUM = {
    "国内长途/零担": "DOMESTIC_LONG_LTL",
    "国内城配/拖车/打包": "DOMESTIC_CITY_DIST",
    "中港/中亚/中欧": "CN_HK_CA_EU",
    "东南亚": "SOUTHEAST_ASIA",
}

_NUMBER = re.compile(r"(\d+(?:\.\d+)?)")
_VEHICLE_SPLIT = re.compile(r"\s*[/xX×*＊]\s*")
_TAIL_QTY = re.compile(r"^(.*?)[\s,，]+(\d+)\s*(台|辆)?$")


def tms_land_type_code(raw: str) -> str:
    """销售中文运输类型 → TMS 6.1.1 landTransportType。对不上空串。"""
    text = (raw or "").strip()
    if text in _TYPE_ENUM:
        return _TYPE_ENUM[text]
    if text in set(_TYPE_ENUM.values()):
        return text
    return ""


def tms_land_route_code(raw: str) -> str:
    """销售中文线路类别 → TMS 6.1.1 routeCategory。对不上空串。"""
    text = (raw or "").strip()
    if text in _ROUTE_ENUM:
        return _ROUTE_ENUM[text]
    if text in set(_ROUTE_ENUM.values()):
        return text
    return ""


def _first_number(raw: str) -> str:
    hit = _NUMBER.search((raw or "").strip())
    return hit.group(1) if hit else ""


def split_vehicle_qty(raw: str) -> tuple[str, str]:
    """
    把「车型/数量」拆成出站车型与台数。

    例：9米6/1、9米6*2、9米6 1台。拆不出台数时台数默认 1，不回头补问。
    不修改入参。
    """
    text = (raw or "").strip()
    if not text:
        return "", ""
    parts = _VEHICLE_SPLIT.split(text, maxsplit=1)
    if len(parts) == 2 and parts[0].strip() and parts[1].strip():
        return parts[0].strip(), _first_number(parts[1]) or parts[1].strip()
    hit = _TAIL_QTY.match(text)
    if hit and hit.group(1).strip():
        return hit.group(1).strip(), hit.group(2)
    return text, "1"


def assemble_land_query(facts: dict[str, Any] | None) -> dict[str, Any]:
    """
    把已采集陆运字段拼成 TMS 查价参数。

    不齐则 ok=False，missing 给销售补问。齐了也不带分类/中港装载。
    改了运输类型后，outbound 只带当前类型需要的键。
    """
    src = apply_quote_date_default(
        normalize_facts({str(k): v for k, v in dict(facts or {}).items() if v is not None})
    )
    # 选项键不走港口别名折叠，再从原文补回
    raw = {str(k): str(v).strip() for k, v in dict(facts or {}).items() if v is not None}
    land_type = (raw.get("运输类型") or src.get("运输类型") or "").strip()
    route = (raw.get("线路类别") or src.get("线路类别") or "").strip()
    origin = (src.get("起运港") or "").strip()
    dest = (src.get("目的港") or "").strip()
    cargo = (src.get("品名") or "").strip()
    quote_date = (src.get("报价日期") or "").strip()
    customs = (raw.get("通关口岸") or src.get("通关口岸") or "").strip()
    hk_load = canonicalize_hk_load(raw.get("中港装载方式") or src.get("中港装载方式") or "")

    missing: list[str] = []
    if not land_type:
        missing.append("运输类型")
    if not route:
        missing.append("线路类别")
    if not origin:
        missing.append("起运港")
    if not dest:
        missing.append("目的港")
    if not cargo:
        missing.append("品名")

    outbound: dict[str, Any] = {
        "transportMode": "LAND",
        "landTransportType": tms_land_type_code(land_type),
        "routeCategory": tms_land_route_code(route),
        "origin": origin,
        "destination": dest,
        "cargoName": cargo,
        "quoteDate": quote_date,
    }

    if land_type in LCL_TYPES:
        vol = _first_number(src.get("体积") or "")
        wgt = _first_number(src.get("毛重") or "")
        if not vol:
            missing.append("体积")
        else:
            outbound["volumeCbm"] = vol
        if not wgt:
            missing.append("毛重")
        else:
            outbound["weightKg"] = wgt
    elif land_type == "中港" and hk_load != "零担/集拼":
        # 中港整车（或尚未选定装载）不采集、不带出车型/数量
        pass
    elif land_type in FTL_TYPES or land_type == "中港":
        combo = (
            src.get("车型/数量")
            or raw.get("车型/数量")
            or src.get("车型")
            or raw.get("车型")
            or ""
        ).strip()
        if not combo:
            missing.append("车型/数量")
        else:
            vehicle, qty = split_vehicle_qty(combo)
            leftover_qty = (src.get("数量") or raw.get("数量") or "").strip()
            if leftover_qty and leftover_qty not in combo:
                qty = leftover_qty
            outbound["vehicleType"] = vehicle
            outbound["quantity"] = qty or "1"
    elif land_type:
        missing.append("运输类型")

    if land_type in CUSTOMS_TYPES and customs:
        outbound["customsPort"] = customs

    ok = len(missing) == 0
    logger.info("tms.land.assemble ok=%s missing=%s type=%s", ok, missing, land_type or "-")
    return {
        "ok": ok,
        "missing": missing,
        "facts": src,
        "outbound": outbound,
    }