File size: 3,365 Bytes
3726258 3f3fbe2 3726258 3f3fbe2 3726258 3f3fbe2 3726258 3f3fbe2 | 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 | """CPU tests of the mask export (no GPU, no SAM 3).
"""
import json
import numpy as np
import pytest
from strangebench.masks import export_sam3_masks as E
def test_rle_round_trip_and_box_convention():
m = np.zeros((40, 60), dtype=np.uint8)
m[5:15, 10:30] = 255
rle = E.rle_from_mask(m)
assert rle["size"] == [40, 60] and isinstance(rle["counts"], str)
assert E.rle_bbox(rle) == [10.0, 5.0, 20.0, 10.0]
assert E.rle_bbox(E.rle_from_mask(np.zeros((4, 4), bool))) is None
def test_box_rle_covers_the_box():
rle = E.box_rle([10, 5, 20, 10], 40, 60)
assert E.rle_bbox(rle) == pytest.approx([10.0, 5.0, 20.0, 10.0], abs=1.0)
def test_match_raw_ann_picks_same_class_best_box():
raw = [{"category_name": "dog", "bbox": [0, 0, 10, 10]}, {"category_name": "cat", "bbox": [50, 50, 10, 10]},
{"category_name": "cat", "bbox": [0, 0, 10, 10]}]
assert E.match_raw_ann(raw, "cat", [1, 1, 10, 10])["bbox"] == [0, 0, 10, 10]
assert E.match_raw_ann(raw, "bird", [0, 0, 5, 5]) is None
def _tree(root, tier="tier2_1typical"):
(root / tier / "images").mkdir(parents=True)
d = {"images": [{"id": 1, "scene_id": "s1", "file_name": "s1.png", "height": 40, "width": 60}],
"categories": [{"id": 1, "name": "dog"}, {"id": 2, "name": "cat"}],
"annotations": [{"id": 7, "image_id": 1, "category_id": 1, "bbox": [10, 5, 20, 10], "area": 200, "iscrowd": 0},
{"id": 8, "image_id": 1, "category_id": 2, "bbox": [40, 20, 10, 10], "area": 100, "iscrowd": 1}]}
(root / tier / "annotations.json").write_text(json.dumps(d))
return d
def test_attach_writes_new_tree_and_keeps_boxes(tmp_path):
tree, out = tmp_path / "shipped", tmp_path / "out"
shipped = _tree(tree)
m = np.zeros((40, 60), dtype=np.uint8)
m[5:15, 10:30] = 255
rle = E.rle_from_mask(m)
E.write_json_atomic(out / "fragments" / "tier2_1typical" / "s1.json", {"scene_id": "s1", "anns": [
{"ann_id": 7, "category": "dog", "bbox_shipped": [10, 5, 20, 10], "segmentation": rle,
"mask_bbox": E.rle_bbox(rle), "box_iou": 1.0}]})
E.main(["attach", "--out-root", str(out), "--variant", f"balanced={tree}"])
got = json.load(open(out / "balanced" / "tier2_1typical" / "annotations.json"))
a7, a8 = got["annotations"]
assert a7["segmentation_source"] == "sam3" and a7["segmentation"] == rle
assert a8["segmentation_source"] == "box"
assert [a["bbox"] for a in got["annotations"]] == [a["bbox"] for a in shipped["annotations"]]
assert json.load(open(tree / "tier2_1typical" / "annotations.json")) == shipped
assert (out / "mask_report.csv").exists()
def test_attach_falls_back_to_box_and_refuses_missing(tmp_path):
tree, out = tmp_path / "shipped", tmp_path / "out"
_tree(tree)
with pytest.raises(SystemExit):
E.main(["attach", "--out-root", str(out), "--variant", f"balanced={tree}"])
E.write_json_atomic(out / "fragments" / "tier2_1typical" / "s1.json", {"scene_id": "s1", "anns": [
{"ann_id": 7, "category": "dog", "bbox_shipped": [10, 5, 20, 10], "segmentation": None, "box_iou": 0.0}]})
E.main(["attach", "--out-root", str(out), "--variant", f"balanced={tree}"])
got = json.load(open(out / "balanced" / "tier2_1typical" / "annotations.json"))
assert got["annotations"][0]["segmentation_source"] == "box"
|