Spaces:
Sleeping
Sleeping
File size: 13,710 Bytes
7323d5e |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 |
import io
import wave
from unittest.mock import MagicMock, patch
import pytest
import torch
import app as app_module
from app import preprocess_audio
def create_wav_buffer(
duration_seconds: float = 1.0, sample_rate: int = 16000
) -> io.BytesIO:
"""Create a valid WAV file buffer with silence."""
buffer = io.BytesIO()
n_frames = int(duration_seconds * sample_rate)
with wave.open(buffer, "wb") as wav_file:
wav_file.setnchannels(1)
wav_file.setsampwidth(2) # 16-bit audio
wav_file.setframerate(sample_rate)
wav_file.writeframes(b"\x00\x00" * n_frames)
buffer.seek(0)
return buffer
def test_health(client):
response = client.get("/health")
assert response.status_code == 200
assert response.json() == {"status": "ok"}
def test_list_speakers_empty(client):
"""Test listing speakers when none are enrolled."""
original_embeddings = app_module.speaker_embeddings.copy()
app_module.speaker_embeddings.clear()
try:
response = client.get("/speakers")
assert response.status_code == 200
assert response.json() == {"speakers": []}
finally:
app_module.speaker_embeddings.update(original_embeddings)
def test_list_speakers_with_data(client):
"""Test listing speakers with enrolled speakers."""
original_embeddings = app_module.speaker_embeddings.copy()
app_module.speaker_embeddings["alice"] = [0.1] * 192
app_module.speaker_embeddings["bob"] = [0.2] * 192
try:
response = client.get("/speakers")
assert response.status_code == 200
data = response.json()
assert set(data["speakers"]) == {"alice", "bob"}
finally:
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_enroll_speaker(client):
"""Test enrolling a new speaker."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
mock_model.encode_batch.return_value = torch.randn(1, 192)
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
try:
buffer = create_wav_buffer()
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
mock_storage = MagicMock()
original_storage = app_module.storage
app_module.storage = mock_storage
try:
response = client.post(
"/speakers/alice/enroll",
files={"file": ("test.wav", buffer, "audio/wav")},
)
assert response.status_code == 200
data = response.json()
assert "alice" in data["message"]
assert data["embedding_size"] == 192
assert "alice" in app_module.speaker_embeddings
finally:
app_module.storage = original_storage
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_enroll_speaker_model_not_loaded(client):
"""Test enrolling when model is not loaded."""
original_model = app_module.speaker_model
app_module.speaker_model = None
try:
buffer = create_wav_buffer()
with pytest.raises(RuntimeError, match="Model not loaded"):
client.post(
"/speakers/alice/enroll",
files={"file": ("test.wav", buffer, "audio/wav")},
)
finally:
app_module.speaker_model = original_model
def test_delete_speaker(client):
"""Test deleting a speaker."""
original_embeddings = app_module.speaker_embeddings.copy()
original_storage = app_module.storage
app_module.speaker_embeddings["alice"] = [0.1] * 192
mock_storage = MagicMock()
app_module.storage = mock_storage
try:
response = client.delete("/speakers/alice")
assert response.status_code == 200
assert "alice" not in app_module.speaker_embeddings
finally:
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
app_module.storage = original_storage
def test_delete_speaker_not_found(client):
"""Test deleting a non-existent speaker."""
response = client.delete("/speakers/unknown")
assert response.status_code == 404
def test_identify_speaker(client):
"""Test identifying a speaker."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
mock_model.encode_batch.return_value = torch.randn(1, 192)
mock_model.similarity.return_value = torch.tensor([[0.85]])
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
response = client.post(
"/identify", files={"file": ("test.wav", buffer, "audio/wav")}
)
assert response.status_code == 200
data = response.json()
assert data["identified"] is True
assert data["speaker"] == "alice"
assert "confidence" in data
assert "threshold" in data
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_identify_speaker_no_match(client):
"""Test identification when no speaker matches."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
mock_model.encode_batch.return_value = torch.randn(1, 192)
mock_model.similarity.return_value = torch.tensor([[0.1]]) # Below threshold
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
response = client.post(
"/identify", files={"file": ("test.wav", buffer, "audio/wav")}
)
assert response.status_code == 200
data = response.json()
assert data["identified"] is False
assert data["speaker"] is None
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_identify_no_speakers_enrolled(client):
"""Test identification when no speakers are enrolled."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
try:
buffer = create_wav_buffer()
response = client.post(
"/identify", files={"file": ("test.wav", buffer, "audio/wav")}
)
assert response.status_code == 400
assert "No speakers enrolled" in response.json()["detail"]
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_verify_speaker(client):
"""Test verifying a specific speaker."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
mock_model.encode_batch.return_value = torch.randn(1, 192)
mock_model.similarity.return_value = torch.tensor([[0.85]])
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
response = client.post(
"/verify",
params={"name": "alice"},
files={"file": ("test.wav", buffer, "audio/wav")},
)
assert response.status_code == 200
data = response.json()
assert data["verified"] is True
assert data["speaker"] == "alice"
assert "confidence" in data
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_verify_speaker_not_found(client):
"""Test verifying a non-existent speaker."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
try:
buffer = create_wav_buffer()
response = client.post(
"/verify",
params={"name": "unknown"},
files={"file": ("test.wav", buffer, "audio/wav")},
)
assert response.status_code == 404
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_verify_speaker_failed(client):
"""Test verification when speaker doesn't match."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
mock_model = MagicMock()
mock_model.encode_batch.return_value = torch.randn(1, 192)
mock_model.similarity.return_value = torch.tensor([[0.1]]) # Below threshold
app_module.speaker_model = mock_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
response = client.post(
"/verify",
params={"name": "alice"},
files={"file": ("test.wav", buffer, "audio/wav")},
)
assert response.status_code == 200
data = response.json()
assert data["verified"] is False
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_identify_model_not_loaded(client):
"""Test identify when model is not loaded."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
app_module.speaker_model = None
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with pytest.raises(RuntimeError, match="Model not loaded"):
client.post("/identify", files={"file": ("test.wav", buffer, "audio/wav")})
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_verify_model_not_loaded(client):
"""Test verify when model is not loaded."""
original_model = app_module.speaker_model
original_embeddings = app_module.speaker_embeddings.copy()
app_module.speaker_model = None
app_module.speaker_embeddings["alice"] = [0.1] * 192
try:
buffer = create_wav_buffer()
with pytest.raises(RuntimeError, match="Model not loaded"):
client.post(
"/verify",
params={"name": "alice"},
files={"file": ("test.wav", buffer, "audio/wav")},
)
finally:
app_module.speaker_model = original_model
app_module.speaker_embeddings.clear()
app_module.speaker_embeddings.update(original_embeddings)
def test_preprocess_audio_with_resampling():
"""Test audio preprocessing with resampling from 44.1kHz to 16kHz."""
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 44100), 44100)
with patch("app.torchaudio.transforms.Resample") as mock_resample:
mock_resampler = MagicMock()
mock_resampler.return_value = torch.zeros(1, 16000)
mock_resample.return_value = mock_resampler
result = preprocess_audio(b"fake_audio_data")
mock_resample.assert_called_once_with(orig_freq=44100, new_freq=16000)
assert result.shape == (1, 16000)
def test_preprocess_audio_stereo_to_mono():
"""Test audio preprocessing converting stereo to mono."""
with patch("app.torchaudio.load") as mock_load:
# Stereo audio (2 channels)
mock_load.return_value = (torch.zeros(2, 16000), 16000)
result = preprocess_audio(b"fake_audio_data")
# Should be converted to mono (1 channel)
assert result.shape[0] == 1
def test_preprocess_audio_no_changes_needed():
"""Test audio preprocessing when no resampling or mono conversion needed."""
with patch("app.torchaudio.load") as mock_load:
mock_load.return_value = (torch.zeros(1, 16000), 16000)
result = preprocess_audio(b"fake_audio_data")
assert result.shape == (1, 16000)
|