Instructions to use Emreuludasdemir/teknofest2026-task3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LightGlue
How to use Emreuludasdemir/teknofest2026-task3 with LightGlue:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 12,209 Bytes
699f3cd | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 | from __future__ import annotations
import json
import shutil
import tempfile
import unittest
from pathlib import Path
from unittest.mock import patch
import numpy as np
from PIL import Image
from src.config.settings import MvpRuntimeSettings
from src.core.frame_state import CanonicalUndefinedObject, DecodedFrame, FrameEnvelope
from src.task3.matcher import Task3Matcher
from src.task3.reference_cache import ReferenceCache
OFFICIAL_REFERENCE_FILES = {
"ref_01": "Referans_Nesne_01.JPG",
"ref_02": "Referans_Nesne_02.JPG",
"ref_03": "Referans_Nesne_03.JPG",
"ref_04": "Referans_Nesne_04.JPG",
"ref_05": "Referans_Nesne_05.jpg",
"ref_06": "Referans_Nesne_06.jpg",
"ref_07": "Referans_Nesne_07.png",
"ref_08": "Referans_Nesne_08.png",
"ref_09": "Referans_Nesne_09.png",
"ref_10": "Referans_Nesne_10.png",
"ref_11": "Referans_Nesne_11.png",
"ref_12": "Referans_Nesne_12.png",
}
def _candidate(object_id: str, score: float, source: str) -> CanonicalUndefinedObject:
return CanonicalUndefinedObject(
object_id=object_id,
top_left_x=0.0,
top_left_y=0.0,
bottom_right_x=10.0,
bottom_right_y=10.0,
metadata={"match_score": score, "matcher_source": source},
)
class Task3ReferenceCacheTests(unittest.TestCase):
def test_v3_spec_auto_routing_matches_phase_b_reference_table(self) -> None:
cache = ReferenceCache()
loaded = cache.preload_from_directory(Path("data/references/2026_baseline"))
self.assertEqual(loaded, 12)
self.assertEqual(cache.get_candidate_suppression_mode(), "per_reference_top_1")
expected = {
"ref_01": ("rgb", "orb", ["rgb"]),
"ref_02": ("rgb", "orb", ["rgb"]),
"ref_03": ("rgb", "orb", ["rgb"]),
"ref_04": ("thermal", "yoloe", ["thermal"]),
"ref_05": ("rgb", "yoloe", ["rgb"]),
"ref_06": ("rgb", "yoloe", ["rgb"]),
"ref_07": ("rgb", "orb", ["rgb"]),
"ref_08": ("rgb", "yoloe", ["rgb"]),
"ref_09": ("rgb", "yoloe", ["rgb"]),
"ref_10": ("rgb", "yoloe", ["rgb"]),
"ref_11": ("thermal", "both", ["rgb", "thermal"]),
"ref_12": ("thermal", "yoloe", ["thermal"]),
}
for reference_id, (modality, detector, modalities) in expected.items():
item = cache.get(reference_id)
self.assertIsNotNone(item, reference_id)
self.assertEqual(item["reference_modality"], modality, reference_id)
self.assertEqual(item["detector"], detector, reference_id)
self.assertEqual(item["detector_modalities"], modalities, reference_id)
self.assertEqual(len(cache.get_overrides_applied()), 1)
self.assertEqual(cache.get_overrides_applied()[0]["reference_id"], "ref_07")
diagnostics = cache.get_routing_diagnostics()
self.assertEqual(diagnostics["ref_07"]["detector"], "orb")
self.assertEqual(diagnostics["ref_07"]["detector_modalities"], ["rgb"])
self.assertEqual(diagnostics["ref_07"]["override"]["modality"], "rgb")
def test_manual_override_replaces_auto_assignment_and_logs_warning(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
shutil.copyfile(f"data/references/2026_baseline/{OFFICIAL_REFERENCE_FILES['ref_01']}", temp_path / "ref_01.jpg")
(temp_path / "manifest.json").write_text(
json.dumps({"spec_path": "spec.json"}),
encoding="utf-8",
)
(temp_path / "spec.json").write_text(
json.dumps(
{
"references": {"ref_01": {"file": "ref_01.jpg"}},
"overrides": {
"ref_01": {
"detector": "yoloe",
"detector_modalities": ["rgb"],
"rationale": "manual override for test",
}
},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
with self.assertLogs("src.task3.reference_cache", level="WARNING") as captured:
cache.preload_from_directory(temp_path)
self.assertEqual(cache.get_detector("ref_01"), "yoloe")
self.assertIn("manual override", "\n".join(captured.output))
self.assertEqual(len(cache.get_overrides_applied()), 1)
def test_malformed_override_missing_detector_raises(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
shutil.copyfile(f"data/references/2026_baseline/{OFFICIAL_REFERENCE_FILES['ref_01']}", temp_path / "ref_01.jpg")
(temp_path / "manifest.json").write_text(json.dumps({"spec_path": "spec.json"}), encoding="utf-8")
(temp_path / "spec.json").write_text(
json.dumps(
{
"references": {"ref_01": {"file": "ref_01.jpg"}},
"overrides": {"ref_01": {"rationale": "missing detector"}},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
with self.assertRaises(ValueError):
cache.preload_from_directory(temp_path)
def test_dual_path_detector_both_routes_reference_through_both_paths(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
image = np.zeros((64, 64, 3), dtype=np.uint8)
gradient = np.tile(np.linspace(100, 150, 64, dtype=np.uint8), (64, 1))
image[..., 0] = gradient
image[..., 1] = np.clip(gradient + 8, 0, 255)
image[..., 2] = np.clip(gradient + 12, 0, 255)
Image.fromarray(image, mode="RGB").save(temp_path / "ref_ambiguous.png")
(temp_path / "manifest.json").write_text(json.dumps({"spec_path": "spec.json"}), encoding="utf-8")
(temp_path / "spec.json").write_text(
json.dumps(
{
"references": {"ref_ambiguous": {"file": "ref_ambiguous.png"}},
"overrides": {
"ref_ambiguous": {
"detector": "both",
"detector_modalities": ["rgb", "thermal"],
"rationale": "dual-path test override",
}
},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
cache.preload_from_directory(temp_path)
matcher = Task3Matcher(reference_cache=cache, runtime_settings=MvpRuntimeSettings(task3_mode="yoloe_vp_lightglue"))
decoded = DecodedFrame(
bgr=object(),
gray=object(),
width=16,
height=16,
channel_count=3,
modality="rgb",
frame_index=0,
)
frame = FrameEnvelope(
frame_url="http://mock/frames/1/",
image_url="/frame.jpg",
video_name="rgb_reference_session",
translation_x=0.0,
translation_y=0.0,
translation_z=0.0,
health_status="1",
)
with patch.object(
Task3Matcher,
"_match_with_yoloe_vp_lightglue",
return_value=[_candidate("ref_ambiguous", 0.51, "task3_yoloe_vp_lightglue")],
) as yoloe_mock, patch.object(
Task3Matcher,
"_match_with_real_orb_only",
return_value=[_candidate("ref_ambiguous", 0.83, "task3_orb_bf_homography")],
) as orb_mock:
matches = matcher.match(frame, b"", cache.list_ids(), decoded_frame=decoded, mode="yoloe_vp_lightglue")
self.assertEqual([item.metadata["matcher_source"] for item in matches], ["task3_orb_bf_homography", "task3_yoloe_vp_lightglue"])
self.assertEqual(yoloe_mock.call_args.args[-1], ["ref_ambiguous"])
self.assertEqual(orb_mock.call_args.args[-1], ["ref_ambiguous"])
self.assertEqual(matcher.last_run_info["yoloe_routed_refs"], ["ref_ambiguous"])
self.assertEqual(matcher.last_run_info["orb_routed_refs"], ["ref_ambiguous"])
self.assertIn("ref_ambiguous", matcher.last_run_info["auto_routing"])
self.assertEqual(len(matcher.last_run_info["overrides_applied"]), 1)
def test_v2_style_spec_still_loads_while_legacy_manual_fields_are_ignored(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
shutil.copyfile(f"data/references/2026_baseline/{OFFICIAL_REFERENCE_FILES['ref_05']}", temp_path / "ref_05.jpg")
(temp_path / "manifest.json").write_text(json.dumps({"spec_path": "spec.json"}), encoding="utf-8")
(temp_path / "spec.json").write_text(
json.dumps(
{
"version": "v2",
"references": {
"ref_05": {
"file": "ref_05.jpg",
"modality": "thermal",
"detector": "orb",
"detector_modalities": ["thermal"],
}
},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
loaded = cache.preload_from_directory(temp_path)
self.assertEqual(loaded, 1)
self.assertEqual(cache.get_detector("ref_05"), "yoloe")
self.assertEqual(cache.get("ref_05")["reference_modality"], "rgb")
def test_spec_can_opt_in_to_per_reference_suppression(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
shutil.copyfile(f"data/references/2026_baseline/{OFFICIAL_REFERENCE_FILES['ref_05']}", temp_path / "ref_05.jpg")
(temp_path / "manifest.json").write_text(json.dumps({"spec_path": "spec.json"}), encoding="utf-8")
(temp_path / "spec.json").write_text(
json.dumps(
{
"per_reference_suppression": True,
"references": {
"ref_05": {
"file": "ref_05.jpg",
}
},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
loaded = cache.preload_from_directory(temp_path)
self.assertEqual(loaded, 1)
self.assertEqual(cache.get_candidate_suppression_mode(), "per_reference_top_1")
def test_spec_can_explicitly_opt_out_of_per_reference_suppression(self) -> None:
with tempfile.TemporaryDirectory() as tmp_dir:
temp_path = Path(tmp_dir)
shutil.copyfile(f"data/references/2026_baseline/{OFFICIAL_REFERENCE_FILES['ref_05']}", temp_path / "ref_05.jpg")
(temp_path / "manifest.json").write_text(json.dumps({"spec_path": "spec.json"}), encoding="utf-8")
(temp_path / "spec.json").write_text(
json.dumps(
{
"per_reference_suppression": False,
"references": {
"ref_05": {
"file": "ref_05.jpg",
}
},
}
),
encoding="utf-8",
)
cache = ReferenceCache()
loaded = cache.preload_from_directory(temp_path)
self.assertEqual(loaded, 1)
self.assertEqual(cache.get_candidate_suppression_mode(), "global_top_1")
if __name__ == "__main__":
unittest.main()
|