123 lines
4.4 KiB
Python
123 lines
4.4 KiB
Python
"""
|
||
海运运输分类：只认 Port to Port、其他。
|
||
|
||
本文件职责：从原话收成标准词、生成补问清单。
|
||
禁止：发明第三项；把分类带去 TMS；改六态。
|
||
线程：纯函数，可在 Worker 槽调用。
|
||
"""
|
||
|
||
from __future__ import annotations
|
||
|
||
import re
|
||
|
||
SEA_CLASSES = ("Port to Port", "其他")
|
||
|
||
# 补问清单下一行「(如：Port to Port、其他）」只是举例，不能当成销售已选。
|
||
_CLARIFY_EXAMPLE_LINE = re.compile(r"^\s*[（(]如\s*[:：]")
|
||
|
||
# 长词优先，避免 P2P 误伤其它词；门到门不在名单里。
|
||
_ALIASES = (
|
||
("port to port", "Port to Port"),
|
||
("港到港", "Port to Port"),
|
||
("p2p", "Port to Port"),
|
||
("其它", "其他"),
|
||
("其他", "其他"),
|
||
)
|
||
|
||
|
||
def canonicalize_sea_class(raw: str) -> str:
|
||
"""
|
||
把销售原话收成 Port to Port 或 其他；对不上返回空串。
|
||
|
||
「门到门」不算。先匹配更长别名。副作用：无。
|
||
"""
|
||
text = (raw or "").strip()
|
||
if not text:
|
||
return ""
|
||
# 企微常把空格发成不换行空格，Port to Port 对不上普通空格就会整段丢掉。
|
||
folded = re.sub(r"\s+", " ", text).strip()
|
||
lowered = folded.lower()
|
||
squashed = re.sub(r"\s+", "", lowered)
|
||
if "port to port" in lowered or "porttoport" in squashed:
|
||
return "Port to Port"
|
||
for needle, canon in _ALIASES:
|
||
if needle == "port to port":
|
||
continue
|
||
if needle in lowered or needle in text or needle in squashed:
|
||
return canon
|
||
stripped = text.replace("运输分类", "").replace("：", "").replace(":", "").strip()
|
||
if stripped in SEA_CLASSES:
|
||
return stripped
|
||
return ""
|
||
|
||
|
||
def _without_clarify_examples(text: str) -> str:
|
||
"""丢掉补问举例行。销售只在冒号后作答，「(如：…)」整行不参与判断。"""
|
||
return "\n".join(
|
||
line for line in (text or "").splitlines() if not _CLARIFY_EXAMPLE_LINE.match(line)
|
||
)
|
||
|
||
|
||
def harvest_sea_class(text: str, current: dict[str, str] | None = None) -> dict[str, str]:
|
||
"""
|
||
当前句能认则覆盖；认不出则保留已有标准值。
|
||
|
||
参数：本轮原文、已有 facts。返回合并后的 facts 副本。
|
||
举例行「(如：Port to Port、其他）」先丢掉，避免示例里的 Port to Port 盖掉冒号后的「其它」。
|
||
「其它」与「其他」都收成「其他」。非法旧值会丢掉，避免「门到门」一直占着必填位。
|
||
"""
|
||
merged = dict(current or {})
|
||
found = canonicalize_sea_class(_without_clarify_examples(text or ""))
|
||
if found:
|
||
merged["运输分类"] = found
|
||
return merged
|
||
kept = canonicalize_sea_class(str(merged.get("运输分类") or ""))
|
||
if kept:
|
||
merged["运输分类"] = kept
|
||
return merged
|
||
from agent.schema.land_options import canonicalize_land_class
|
||
|
||
# 多段陆运补「急件」时，不能按海运名单把已收的陆运分类清掉。
|
||
land_kept = canonicalize_land_class(str(merged.get("运输分类") or ""))
|
||
if land_kept:
|
||
merged["运输分类"] = land_kept
|
||
return merged
|
||
merged.pop("运输分类", None)
|
||
return merged
|
||
|
||
|
||
# 运输分类用语，不能冒充贸易条款。销售原话写了「贸易条款」才保留。
|
||
_DOOR_TRADE_TERMS = frozenset(
|
||
{
|
||
"门到门",
|
||
"door to door",
|
||
"door-to-door",
|
||
"doortodoor",
|
||
"d2d",
|
||
}
|
||
)
|
||
|
||
|
||
def drop_invented_door_trade_terms(text: str, facts: dict[str, str] | None = None) -> dict[str, str]:
|
||
"""
|
||
丢掉冒充贸易条款的「门到门」。
|
||
|
||
测服 WO202609200013：原话没写贸易条款，确认卡却出现「贸易条款：门到门」。
|
||
门到门只是运输分类说法，不是 FOB/CIF 这类条款；原话没有「贸易条款」四字时不准收下。
|
||
销售明确写了「贸易条款：门到门」则保留原文，不改六态。
|
||
"""
|
||
merged = dict(facts or {})
|
||
trade = str(merged.get("贸易条款") or "").strip()
|
||
if not trade:
|
||
return merged
|
||
if "贸易条款" in (text or ""):
|
||
return merged
|
||
if trade.lower().replace(" ", "") in {x.replace(" ", "") for x in _DOOR_TRADE_TERMS} or trade in _DOOR_TRADE_TERMS:
|
||
merged.pop("贸易条款", None)
|
||
return merged
|
||
|
||
|
||
def sea_class_prompt() -> str:
|
||
"""缺运输分类时只列两项，不再追问「其他是什么」。"""
|
||
return "运输分类（必填）：请回复 Port to Port 或 其他。"
|