scripts/register_amul_target.py

"""Register Amul AI as a target in CeRAI's database.

Run this once after `python AIEvaluationTool/src/app/importer/main.py --config config.json`.
Idempotent: CeRAI's `add_or_get_target` looks up by name.

Usage:
    python scripts/register_amul_target.py
"""
from __future__ import annotations

import json
import os
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
CERAI = ROOT / "AIEvaluationTool"
sys.path.insert(0, str(CERAI / "src"))

from lib.orm.DB import DB  # noqa: E402
from lib.data import Target  # noqa: E402


def main() -> None:
    cfg_path = CERAI / "config.json"
    cfg = json.loads(cfg_path.read_text())
    engine = cfg["db"].get("engine", "sqlite").lower()

    if engine == "sqlite":
        # Match the importer/executor: DB lives under AIEvaluationTool/data/<file>.
        db_path = CERAI / "data" / cfg["db"].get("file", "AIEvaluationData.db")
        db_url = f"sqlite:///{db_path}"
    elif engine == "mariadb":
        c = cfg["db"]
        db_url = (
            f"mariadb+mariadbconnector://{c['user']}:{c['password']}"
            f"@{c['host']}:{c['port']}/{c['database']}"
        )
    else:
        raise SystemExit(f"Unsupported db engine: {engine}")

    db = DB(db_url=db_url, debug=False)

    tgt = Target(
        target_name="Amul AI",
        target_type="API",
        target_url="http://localhost:8088",
        target_description=(
            "Public dairy/animal-husbandry chat assistant hosted at amulai.in/chat. "
            "The upstream is a GET-SSE endpoint on api.prod.amulai.in; "
            "the URL above points to amul_proxy/server.py — an OpenAI-compatible "
            "shim that forwards POST /v1/chat/completions to the upstream."
        ),
        target_domain="agriculture",
        target_languages=["english"],
    )
    tid = db.add_or_get_target(target=tgt)
    print(f"Registered target 'Amul AI' (id={tid}) in {db_url}")


if __name__ == "__main__":
    main()