File size: 1,987 Bytes
de01f6a | 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 | """Tests for segmentation."""
from __future__ import annotations
from pathlib import Path
from video_vec2wav2_tokenizer.audio.extractor import load_wav
from video_vec2wav2_tokenizer.segmentation.segmenter import Segmenter, normalize_segments
from video_vec2wav2_tokenizer.utils.io import Segment
def test_normalize_drops_short_segments() -> None:
segs = [Segment(0.0, 0.4, "tiny"), Segment(1.0, 3.0, "ok")]
out = normalize_segments(segs, min_length=1.0, max_length=20.0)
assert len(out) == 1
assert out[0].text == "ok"
def test_normalize_splits_long_segments() -> None:
segs = [Segment(0.0, 50.0, "long")]
out = normalize_segments(segs, min_length=1.0, max_length=20.0)
assert len(out) == 3 # ceil(50/20)
assert out[0].text == "long"
assert out[1].text == "" # text not duplicated
assert all(s.duration <= 20.0 + 1e-6 for s in out)
def test_normalize_padding_clamped() -> None:
segs = [Segment(0.0, 2.0, "x")]
out = normalize_segments(segs, min_length=0.5, max_length=20.0, padding=1.0, total_duration=2.5)
assert out[0].start == 0.0 # clamped at 0
assert out[0].end == 2.5 # clamped at total_duration
def test_segment_file_writes_clips(sine_wav: Path, tmp_path: Path) -> None:
segs = [Segment(0.0, 1.0, "a"), Segment(1.0, 2.0, "b")]
segmenter = Segmenter(sample_rate=16000)
out_dir = tmp_path / "clips"
results = segmenter.segment_file(sine_wav, segs, out_dir, start_index=1)
assert [p.name for p, _ in results] == ["000001.wav", "000002.wav"]
samples, sr = load_wav(results[0][0])
assert sr == 16000
assert abs(samples.shape[0] - 16000) < 10 # ~1 second
def test_segment_index_continues(sine_wav: Path, tmp_path: Path) -> None:
segs = [Segment(0.0, 1.0, "a")]
segmenter = Segmenter(sample_rate=16000)
out_dir = tmp_path / "clips"
r1 = segmenter.segment_file(sine_wav, segs, out_dir, start_index=5)
assert r1[0][0].name == "000005.wav"
|