tests/test_scan_script.py

from __future__ import annotations

import importlib.util
import json
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
SCAN_PATH = ROOT / "skills" / "keys-receipt-scanner" / "scripts" / "scan.py"


def _load_scan():
    spec = importlib.util.spec_from_file_location("lamp_scan", SCAN_PATH)
    assert spec and spec.loader
    mod = importlib.util.module_from_spec(spec)
    spec.loader.exec_module(mod)
    return mod


def test_post_gemma_sends_image_url(tmp_path: Path, monkeypatch) -> None:
    scan = _load_scan()
    jpeg = tmp_path / "r.jpg"
    jpeg.write_bytes(b"fake-jpeg")
    captured: dict = {}

    def fake_json_req(url, *, data=None, headers=None, timeout=30.0):
        captured["url"] = url
        captured["body"] = json.loads(data.decode())
        captured["headers"] = headers
        return {
            "choices": [
                {
                    "message": {
                        "content": json.dumps(
                            {"doc_kind": "receipt", "vendor": "Cafe", "total": 3.5}
                        )
                    }
                }
            ]
        }

    monkeypatch.setattr(scan, "_json_req", fake_json_req)
    out = scan.post_gemma("http://gpu:8080/v1", "google/gemma-4-12B-it", jpeg)
    assert out["vendor"] == "Cafe"
    assert captured["url"].endswith("/v1/chat/completions")
    content = captured["body"]["messages"][1]["content"]
    kinds = {p["type"] for p in content}
    assert "image_url" in kinds
    url = next(p["image_url"]["url"] for p in content if p["type"] == "image_url")
    assert url.startswith("data:image/jpeg;base64,")


def test_post_studio_polls_job(tmp_path: Path, monkeypatch) -> None:
    scan = _load_scan()
    jpeg = tmp_path / "r.jpg"
    jpeg.write_bytes(b"fake-jpeg")
    calls: list[str] = []

    def fake_json_req(url, *, data=None, headers=None, timeout=30.0):
        calls.append(url)
        if url.endswith("/api/inbox"):
            return {"ok": True, "job_id": "abc"}
        if url.endswith("/api/jobs/abc"):
            return {"state": "done", "vendor": "Whole Foods", "total": "12.00"}
        raise AssertionError(url)

    monkeypatch.setattr(scan, "_json_req", fake_json_req)
    out = scan.post_studio("http://gpu:7860", jpeg)
    assert out["vendor"] == "Whole Foods"
    assert any(u.endswith("/api/inbox") for u in calls)
    assert any("/api/jobs/abc" in u for u in calls)