Files
tw2/qingjian_cs/scripts/replay_golden.py
jimi a842f1861f
Some checks failed
Pre-commit / run (ubuntu-latest) (push) Has been cancelled
Deploy Sphinx documentation to Pages / build_en (ubuntu-latest, 3.10) (push) Has been cancelled
Deploy Sphinx documentation to Pages / build_zh (ubuntu-latest, 3.10) (push) Has been cancelled
Python Unittest Coverage / test (macos-15, 3.10) (push) Has been cancelled
Python Unittest Coverage / test (macos-15, 3.11) (push) Has been cancelled
Python Unittest Coverage / test (macos-15, 3.12) (push) Has been cancelled
Python Unittest Coverage / test (ubuntu-latest, 3.10) (push) Has been cancelled
Python Unittest Coverage / test (ubuntu-latest, 3.11) (push) Has been cancelled
Python Unittest Coverage / test (ubuntu-latest, 3.12) (push) Has been cancelled
Python Unittest Coverage / test (windows-latest, 3.10) (push) Has been cancelled
Python Unittest Coverage / test (windows-latest, 3.11) (push) Has been cancelled
Python Unittest Coverage / test (windows-latest, 3.12) (push) Has been cancelled
chore: initial import of standalone agentscope project
2026-03-02 18:21:40 +08:00

103 lines
3.5 KiB
Python

from __future__ import annotations
import argparse
import asyncio
import json
import sys
from pathlib import Path
ROOT = Path(__file__).resolve().parents[1]
if str(ROOT) not in sys.path:
sys.path.insert(0, str(ROOT))
from app.orchestrator import Orchestrator
from app.rules import detect_intent, detect_order_status, has_map_or_political_risk, requests_external_contact
from app.state_machine import evolve_after_sales_state, migrate_state_schema
CASES_PATH = ROOT / "golden" / "golden_cases.jsonl"
def heuristic_decide(msg: str, goods_order: str) -> tuple[str, str, dict]:
m = (msg or "")
intent = detect_intent(m)
order_status = detect_order_status(goods_order or "")
if has_map_or_political_risk(m) or requests_external_contact(m):
route, action = "risk", "transfer"
elif "退款" in m:
route, action = "after_sales", "reply"
elif intent == "image":
route, action = "quote", "quote"
elif any(k in m for k in ["多少钱", "报价", "价格", "怎么收费"]):
route, action = "quote", "reply"
elif order_status == "paid":
route, action = "after_sales", "reply"
else:
route, action = "pre_sales", "reply"
state = migrate_state_schema({})
state = evolve_after_sales_state(state, route=route, action=action, intent=intent, order_status=order_status, msg=m)
return route, action, state
async def full_decide(orch: Orchestrator, cid: str, msg: str, goods_order: str) -> tuple[str, str, dict]:
context = {
"customer_key": f"demo_acc:{cid}",
"acc_id": "demo_acc",
"customer_id": cid,
"goods_name": "demo_goods",
"goods_order": goods_order,
"msg": msg,
"pending_images": 0,
"auto_quote_trigger": False,
"last_reply": "",
}
route, decision, state = await orch.decide(context)
return route, decision.action, state
async def main() -> int:
parser = argparse.ArgumentParser()
parser.add_argument("--mode", choices=["heuristic", "full"], default="heuristic")
parser.add_argument("--cases", default=str(CASES_PATH))
args = parser.parse_args()
p = Path(args.cases)
if not p.exists():
print(f"cases not found: {p}")
return 2
cases = [json.loads(line) for line in p.read_text(encoding="utf-8").splitlines() if line.strip()]
orch = Orchestrator() if args.mode == "full" else None
passed = 0
failed = 0
for i, c in enumerate(cases, 1):
cid = f"case_{i}"
if args.mode == "full":
route, action, state = await full_decide(orch, cid, c.get("msg", ""), c.get("goods_order", ""))
else:
route, action, state = heuristic_decide(c.get("msg", ""), c.get("goods_order", ""))
ok = route == c.get("expected_route") and action == c.get("expected_action")
if ok and c.get("expected_stage"):
ok = state.get("after_sales_stage") == c.get("expected_stage")
if ok:
passed += 1
print(f"[PASS] {c.get('id')} route={route} action={action} stage={state.get('after_sales_stage')}")
else:
failed += 1
print(
f"[FAIL] {c.get('id')} expect=({c.get('expected_route')},{c.get('expected_action')},{c.get('expected_stage','-')}) "
f"got=({route},{action},{state.get('after_sales_stage')})"
)
print(f"\nSummary: passed={passed} failed={failed} total={len(cases)} mode={args.mode}")
return 1 if failed else 0
if __name__ == "__main__":
raise SystemExit(asyncio.run(main()))