c86e323d7a
Quality-first multi-pass audiobook generation from English manuscripts. Includes: - umbraaudio/ core package (pipeline, config, voices, web review UI) - scripts/ CLI utilities + model download script - tts_server/ remote TTS server - specs/ design docs, tests/, pyproject.toml + uv.lock - Qwen3-TTS git submodule (pinned 022e286) with local demo.py UI patch saved at scripts/qwen3_tts_demo_umbra_patch.patch - .env with runtime config/secrets Excluded (see .gitignore): models/ (downloaded via scripts/download_models.sh), output/, logs/, .venv, .umbra runtime data, and book projects (DealHunterShip_*), which are committed separately.
94 lines
2.8 KiB
Python
94 lines
2.8 KiB
Python
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
from umbraaudio.assembly import AudioStitcher
|
|
from umbraaudio.assembly.metadata import (
|
|
BookMetadata,
|
|
ChapterMeta,
|
|
build_chapter_metadata,
|
|
read_metadata_json,
|
|
write_ffmpeg_metadata,
|
|
write_metadata_json,
|
|
)
|
|
|
|
|
|
def test_stitch_utterances_output_exists(
|
|
temp_dir: Path, sample_wavs: list[Path], stitcher: AudioStitcher
|
|
):
|
|
output = temp_dir / "stitched.wav"
|
|
result = stitcher.stitch_utterances(sample_wavs, output)
|
|
assert result == output
|
|
assert output.exists()
|
|
assert output.stat().st_size > 0
|
|
|
|
|
|
def test_stitch_utterances_empty_list(stitcher: AudioStitcher, temp_dir: Path):
|
|
output = temp_dir / "empty.wav"
|
|
result = stitcher.stitch_utterances([], output)
|
|
assert result == output
|
|
assert output.exists()
|
|
|
|
|
|
def test_stitch_chapters_output_exists(
|
|
temp_dir: Path, sample_wavs: list[Path], stitcher: AudioStitcher
|
|
):
|
|
output = temp_dir / "chapters.wav"
|
|
result = stitcher.stitch_chapters(sample_wavs, output)
|
|
assert result == output
|
|
assert output.exists()
|
|
|
|
|
|
def test_make_silence(stitcher: AudioStitcher):
|
|
silence = stitcher._make_silence(100)
|
|
assert len(silence) == int(24000 * 0.1)
|
|
assert (silence == 0).all()
|
|
|
|
|
|
def test_build_chapter_metadata():
|
|
titles = ["Ch 1", "Ch 2", "Ch 3"]
|
|
durations = [10.0, 15.0, 20.0]
|
|
meta = build_chapter_metadata(titles, durations, book_title="Test Book", author="Author")
|
|
assert meta.title == "Test Book"
|
|
assert meta.author == "Author"
|
|
assert len(meta.chapters) == 3
|
|
assert meta.chapters[0].title == "Ch 1"
|
|
assert meta.chapters[0].start_time_sec == 0.0
|
|
assert meta.chapters[0].end_time_sec == 10.0
|
|
assert meta.chapters[1].start_time_sec == 10.0
|
|
assert meta.chapters[1].end_time_sec == 25.0
|
|
|
|
|
|
def test_metadata_json_roundtrip(temp_dir: Path):
|
|
meta = BookMetadata(
|
|
title="Book",
|
|
author="Author",
|
|
narrator="Narrator",
|
|
chapters=[ChapterMeta(title="Ch 1", start_time_sec=0.0, end_time_sec=10.0, index=0)],
|
|
)
|
|
path = temp_dir / "meta.json"
|
|
write_metadata_json(meta, path)
|
|
assert path.exists()
|
|
loaded = read_metadata_json(path)
|
|
assert loaded.title == "Book"
|
|
assert loaded.author == "Author"
|
|
assert len(loaded.chapters) == 1
|
|
assert loaded.chapters[0].title == "Ch 1"
|
|
|
|
|
|
def test_write_ffmpeg_metadata(temp_dir: Path):
|
|
meta = BookMetadata(
|
|
title="Test Book",
|
|
author="Author",
|
|
chapters=[ChapterMeta(title="Ch 1", start_time_sec=0.0, end_time_sec=10.0, index=0)],
|
|
)
|
|
path = temp_dir / "ffmetadata.txt"
|
|
write_ffmpeg_metadata(meta, path)
|
|
assert path.exists()
|
|
content = path.read_text()
|
|
assert "title=Test Book" in content
|
|
assert "artist=Author" in content
|
|
assert "[CHAPTER]" in content
|
|
assert "START=0" in content
|
|
assert "END=10000" in content
|