"""Time the endpoint through the official SDK: one cold request, then warm ones.

just smoke [app]
"""

import json
import os
import statistics
import time

from typesafe_sdk import TypeSafeClient

STATE = {
    "from": "user@acme.com",
    "subject": "Duplicate charge on invoice #4411",
    "body": "Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan.",
}
QUESTIONS = {
    "department": {
        "type": "choice",
        "instructions": "Which department should handle this request?",
        "criteria": {
            "billing": "invoices, payments, refunds",
            "technical": "bugs, outages, system errors",
            "sales": "pricing, new contracts",
            "other": "everything else",
        },
    },
    "urgency": {
        "type": "score",
        "instructions": "How urgent is this request?",
        "criteria": ["not urgent", "soon", "critical deadline or blocking issue"],
    },
    "churn_risk": {
        "type": "noul",
        "instructions": "Does the user threaten to cancel or leave?",
    },
}


def main() -> None:
    client = TypeSafeClient(
        api_key=os.environ["SYSTEMONE_API_KEY"],
        base_url=os.environ["SYSTEMONE_URL"],
        model="kev-latest",
        timeout=600,
    )
    with client:
        t = time.perf_counter()
        first = client.system_one(state=STATE, questions=QUESTIONS)
        cold = time.perf_counter() - t
        warm = []
        for _ in range(5):
            t = time.perf_counter()
            client.system_one(state=STATE, questions=QUESTIONS)
            warm.append(time.perf_counter() - t)
    print(json.dumps(first.model_dump(), indent=1, default=str))
    print(
        f"first request {cold:.1f}s, warm median {statistics.median(warm) * 1000:.0f}ms, warm max {max(warm) * 1000:.0f}ms"
    )


if __name__ == "__main__":
    main()