File size: 3,098 Bytes
6d35aff | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 | """Exercise the public one-image control path without APIs or model weights."""
import json
import sys
import tempfile
from pathlib import Path
from PIL import Image
if __package__ in {None, ""}:
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
from approach.run_ape import build_parser as build_ape_parser
from approach.run_ape import run as run_ape
from approach.run_vlm import build_parser as build_vlm_parser
from approach.run_vlm import run as run_vlm
from scripts.generate_questions import build_questions, write_jsonl
def main():
with tempfile.TemporaryDirectory(prefix="orienter-smoke-") as tmpdir:
root = Path(tmpdir)
images_dir = root / "images"
images_dir.mkdir()
Image.new("RGB", (32, 32), color=(32, 64, 96)).save(images_dir / "123_4.png")
questions_path = root / "questions.jsonl"
write_jsonl(questions_path, build_questions(images_dir, "Smoke test"))
candidates_path = root / "candidates.jsonl"
vlm_args = build_vlm_parser().parse_args(
[
"--questions",
str(questions_path),
"--images-dir",
str(images_dir),
"--output",
str(candidates_path),
]
)
def offline_processor(profile, question, image_path, ablation, key_index):
return {"objects": {"button": "synthetic blue square"}}
vlm_report = run_vlm(vlm_args, processor=offline_processor)
predictions_path = root / "predictions.json"
ape_args = build_ape_parser().parse_args(
[
"--questions",
str(questions_path),
"--candidates",
str(candidates_path),
"--images-dir",
str(images_dir),
"--output",
str(predictions_path),
]
)
def offline_inference(**kwargs):
return [
{
"category_name": "button",
"bbox": [4, 5, 12, 10],
"score": 0.9,
}
]
ape_report = run_ape(ape_args, inference=offline_inference)
predictions = json.loads(predictions_path.read_text(encoding="utf-8"))
expected = {
"image_id": 123004,
"category_id": "button",
"category_name": "button",
"bbox": [4, 5, 12, 10],
}
if len(predictions) != 1 or any(
predictions[0].get(key) != value for key, value in expected.items()
):
raise RuntimeError(f"Unexpected smoke prediction: {predictions}")
print(
json.dumps(
{
"status": "ok",
"vlm_records": vlm_report["completed"],
"ape_predictions": ape_report["predictions"],
"image_id": predictions[0]["image_id"],
},
indent=2,
)
)
return 0
if __name__ == "__main__":
sys.exit(main())
|