moss-tts-nano-cli 0.1.0__tar.gz

moss_tts_nano_cli-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,40 @@
+name: Publish to PyPI
+
+on:
+  push:
+    tags:
+      - "v*.*.*"
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      id-token: write
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v3
+        with:
+          version: "0.11.8"
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install build dependencies
+        run: uv sync --extra dev
+
+      - name: Run tests
+        run: uv run pytest
+
+      - name: Build package
+        run: uv build
+
+      - name: Publish to PyPI
+        env:
+          UV_PUBLISH_TOKEN: ${{ secrets.PYPI_API_TOKEN }}
+        run: uv publish

moss_tts_nano_cli-0.1.0/.gitignore

@@ -0,0 +1,10 @@
+# Python-generated files
+__pycache__/
+*.py[oc]
+build/
+dist/
+wheels/
+*.egg-info
+
+# Virtual environments
+.venv

moss_tts_nano_cli-0.1.0/.python-version

@@ -0,0 +1 @@
+3.12

moss_tts_nano_cli-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 moss-tts-nano-cli contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

moss_tts_nano_cli-0.1.0/PKG-INFO

@@ -0,0 +1,94 @@
+Metadata-Version: 2.4
+Name: moss-tts-nano-cli
+Version: 0.1.0
+Summary: One-command zero-shot voice cloning with MOSS-TTS-Nano via mlx-audio.
+Project-URL: Homepage, https://github.com/vra/moss-tts-nano-cli
+Project-URL: Repository, https://github.com/vra/moss-tts-nano-cli
+Project-URL: Issues, https://github.com/vra/moss-tts-nano-cli/issues
+Author-email: 蔚山 <weishan.wyf@alibaba-inc.com>
+License-File: LICENSE
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
+Requires-Python: >=3.12
+Requires-Dist: click>=8.0.0
+Requires-Dist: mlx-audio>=0.0.1
+Requires-Dist: numpy>=1.26.0
+Requires-Dist: soundfile>=0.12.0
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0.0; extra == 'dev'
+Requires-Dist: ruff>=0.6.0; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# moss-tts-nano-cli
+
+One-command zero-shot voice cloning with the [MOSS-TTS-Nano-100M](https://huggingface.co/mlx-community/MOSS-TTS-Nano-100M) model, powered by [mlx-audio](https://github.com/Blaizzy/mlx-audio).
+
+## Features
+
+- **No manual setup**: `uvx moss-tts-nano-cli` installs the environment automatically.
+- **Lazy model download**: the model is fetched from Hugging Face on first use.
+- **Voice cloning**: pass a reference WAV and text to synthesize speech in that voice.
+- **Apple Silicon optimized**: runs on MLX.
+
+## Requirements
+
+- Python 3.12+
+- macOS with Apple Silicon (MLX backend)
+- [uv](https://docs.astral.sh/uv/) (recommended)
+- FFmpeg (used indirectly by audio tooling)
+
+## Install
+
+```bash
+uv tool install moss-tts-nano-cli
+```
+
+Or run without installing:
+
+```bash
+uvx --from moss-tts-nano-cli moss-tts-nano --help
+```
+
+## Usage
+
+### Clone a voice
+
+```bash
+moss-tts-nano clone \
+  --ref ./my_voice.wav \
+  --text "Hello, this is my cloned voice speaking." \
+  --out output.wav
+```
+
+The first run downloads `mlx-community/MOSS-TTS-Nano-100M` and the audio tokenizer into the Hugging Face cache (`~/.cache/huggingface`).
+
+### Eagerly download the model
+
+```bash
+moss-tts-nano download
+```
+
+### Show info
+
+```bash
+moss-tts-nano info
+```
+
+## Development
+
+```bash
+git clone https://github.com/vra/moss-tts-nano-cli.git
+cd moss-tts-nano-cli
+uv sync --extra dev
+uv run pytest
+```
+
+## License
+
+MIT

moss_tts_nano_cli-0.1.0/README.md

@@ -0,0 +1,67 @@
+# moss-tts-nano-cli
+
+One-command zero-shot voice cloning with the [MOSS-TTS-Nano-100M](https://huggingface.co/mlx-community/MOSS-TTS-Nano-100M) model, powered by [mlx-audio](https://github.com/Blaizzy/mlx-audio).
+
+## Features
+
+- **No manual setup**: `uvx moss-tts-nano-cli` installs the environment automatically.
+- **Lazy model download**: the model is fetched from Hugging Face on first use.
+- **Voice cloning**: pass a reference WAV and text to synthesize speech in that voice.
+- **Apple Silicon optimized**: runs on MLX.
+
+## Requirements
+
+- Python 3.12+
+- macOS with Apple Silicon (MLX backend)
+- [uv](https://docs.astral.sh/uv/) (recommended)
+- FFmpeg (used indirectly by audio tooling)
+
+## Install
+
+```bash
+uv tool install moss-tts-nano-cli
+```
+
+Or run without installing:
+
+```bash
+uvx --from moss-tts-nano-cli moss-tts-nano --help
+```
+
+## Usage
+
+### Clone a voice
+
+```bash
+moss-tts-nano clone \
+  --ref ./my_voice.wav \
+  --text "Hello, this is my cloned voice speaking." \
+  --out output.wav
+```
+
+The first run downloads `mlx-community/MOSS-TTS-Nano-100M` and the audio tokenizer into the Hugging Face cache (`~/.cache/huggingface`).
+
+### Eagerly download the model
+
+```bash
+moss-tts-nano download
+```
+
+### Show info
+
+```bash
+moss-tts-nano info
+```
+
+## Development
+
+```bash
+git clone https://github.com/vra/moss-tts-nano-cli.git
+cd moss-tts-nano-cli
+uv sync --extra dev
+uv run pytest
+```
+
+## License
+
+MIT

moss_tts_nano_cli-0.1.0/pyproject.toml

@@ -0,0 +1,57 @@
+[project]
+name = "moss-tts-nano-cli"
+version = "0.1.0"
+description = "One-command zero-shot voice cloning with MOSS-TTS-Nano via mlx-audio."
+readme = "README.md"
+authors = [
+    { name = "蔚山", email = "weishan.wyf@alibaba-inc.com" }
+]
+requires-python = ">=3.12"
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Multimedia :: Sound/Audio :: Speech",
+]
+dependencies = [
+    "click>=8.0.0",
+    "mlx-audio>=0.0.1",
+    "numpy>=1.26.0",
+    "soundfile>=0.12.0",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pytest>=8.0.0",
+    "ruff>=0.6.0",
+]
+
+[project.scripts]
+moss-tts-nano = "moss_tts_nano_cli:main"
+moss-tts-nano-cli = "moss_tts_nano_cli:main"
+
+[project.urls]
+Homepage = "https://github.com/vra/moss-tts-nano-cli"
+Repository = "https://github.com/vra/moss-tts-nano-cli"
+Issues = "https://github.com/vra/moss-tts-nano-cli/issues"
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/moss_tts_nano_cli"]
+
+[tool.ruff]
+line-length = 100
+target-version = "py312"
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "W"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]

moss_tts_nano_cli-0.1.0/src/moss_tts_nano_cli/__init__.py

@@ -0,0 +1,5 @@
+"""MOSS-TTS-Nano CLI: zero-shot voice cloning with one command."""
+
+from moss_tts_nano_cli.cli import main
+
+__all__ = ["main"]

moss_tts_nano_cli-0.1.0/src/moss_tts_nano_cli/cache.py

@@ -0,0 +1,10 @@
+"""Cache and path helpers for the MOSS-TTS-Nano CLI."""
+
+from pathlib import Path
+
+DEFAULT_MODEL_ID = "mlx-community/MOSS-TTS-Nano-100M"
+
+
+def default_cache_dir() -> Path:
+    """Return the default Hugging Face cache directory used by mlx-audio."""
+    return Path.home() / ".cache" / "huggingface"

moss_tts_nano_cli-0.1.0/src/moss_tts_nano_cli/cli.py

@@ -0,0 +1,88 @@
+"""Command-line interface for MOSS-TTS-Nano."""
+
+from pathlib import Path
+
+import click
+
+from moss_tts_nano_cli.cache import DEFAULT_MODEL_ID, default_cache_dir
+from moss_tts_nano_cli.tts import clone_voice, save_audio
+
+
+@click.group()
+@click.version_option(version="0.1.0", prog_name="moss-tts-nano")
+def cli():
+    """MOSS-TTS-Nano CLI: zero-shot multilingual voice cloning."""
+
+
+@cli.command()
+@click.option(
+    "--ref",
+    "ref_path",
+    required=True,
+    type=click.Path(exists=True, dir_okay=False, path_type=Path),
+    help="Path to the reference WAV audio file for voice cloning.",
+)
+@click.option(
+    "--text",
+    required=True,
+    help="Text to synthesize in the cloned voice.",
+)
+@click.option(
+    "--out",
+    "out_path",
+    default="output.wav",
+    show_default=True,
+    type=click.Path(path_type=Path),
+    help="Output WAV file path.",
+)
+@click.option(
+    "--model",
+    "model_id",
+    default=DEFAULT_MODEL_ID,
+    show_default=True,
+    help="Hugging Face model id for MOSS-TTS-Nano.",
+)
+@click.option(
+    "--quiet",
+    is_flag=True,
+    help="Suppress progress messages.",
+)
+def clone(ref_path: Path, text: str, out_path: Path, model_id: str, quiet: bool) -> None:
+    """Clone a voice from REF audio and speak TEXT."""
+    audio, sample_rate = clone_voice(
+        text=text,
+        ref_audio_path=ref_path,
+        model_id=model_id,
+        verbose=not quiet,
+    )
+    save_audio(out_path, audio, sample_rate)
+    click.echo(f"Saved output to {out_path}")
+
+
+@cli.command()
+@click.option(
+    "--model",
+    "model_id",
+    default=DEFAULT_MODEL_ID,
+    show_default=True,
+    help="Hugging Face model id for MOSS-TTS-Nano.",
+)
+def download(model_id: str) -> None:
+    """Eagerly download the model and audio tokenizer."""
+    click.echo(f"Downloading model {model_id!r}...")
+    from mlx_audio.tts.utils import load_model
+
+    load_model(model_id)
+    click.echo("Model download complete.")
+
+
+@cli.command()
+def info() -> None:
+    """Show cache location, default model, and package metadata."""
+    click.echo(f"Default model: {DEFAULT_MODEL_ID}")
+    click.echo(f"Cache directory: {default_cache_dir()}")
+    click.echo("Commands: clone, download, info")
+
+
+def main() -> None:
+    cli()

moss_tts_nano_cli-0.1.0/src/moss_tts_nano_cli/tts.py

@@ -0,0 +1,62 @@
+"""Text-to-speech inference helpers for MOSS-TTS-Nano."""
+
+from pathlib import Path
+
+import numpy as np
+import soundfile as sf
+from mlx_audio.tts.utils import load_model
+
+from moss_tts_nano_cli.cache import DEFAULT_MODEL_ID
+
+
+def load_reference_audio(path: Path) -> tuple[np.ndarray, int]:
+    """Load a reference WAV file and return mono audio + sample rate."""
+    audio, sample_rate = sf.read(str(path))
+    if audio.ndim > 1:
+        audio = audio.mean(axis=1)
+    return audio.astype(np.float32), int(sample_rate)
+
+
+def clone_voice(
+    text: str,
+    ref_audio_path: Path,
+    model_id: str = DEFAULT_MODEL_ID,
+    verbose: bool = True,
+):
+    """Generate speech cloned from *ref_audio_path* speaking *text*.
+
+    Returns the generated waveform as a NumPy array and the model sample rate.
+    """
+    ref_audio, ref_sr = load_reference_audio(ref_audio_path)
+    if verbose:
+        print(f"Loaded reference audio: {ref_audio.shape[0]} samples @ {ref_sr} Hz")
+
+    if verbose:
+        print(f"Loading model {model_id!r}... (first run will download files)")
+    model = load_model(model_id)
+    if verbose:
+        print(f"Model loaded; sample rate = {model.sample_rate} Hz")
+
+    if verbose:
+        print(f"Generating speech for: {text!r}")
+    results = list(
+        model.generate(
+            text=text,
+            ref_audio=ref_audio,
+            ref_audio_sample_rate=ref_sr,
+        )
+    )
+    result = results[0]
+    audio = np.array(result.audio)
+    if verbose:
+        print(
+            f"Generated {audio.shape[0]} samples "
+            f"({result.audio_duration}) @ {result.sample_rate} Hz "
+            f"RTF={result.real_time_factor:.3f}"
+        )
+    return audio, int(result.sample_rate)
+
+
+def save_audio(path: Path, audio: np.ndarray, sample_rate: int) -> None:
+    """Save a waveform to a WAV file."""
+    sf.write(str(path), audio, sample_rate)

moss_tts_nano_cli-0.1.0/tests/test_cli.py

@@ -0,0 +1,19 @@
+"""Smoke tests for the CLI entry points."""
+
+from click.testing import CliRunner
+
+from moss_tts_nano_cli.cli import cli
+
+
+def test_info_command():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["info"])
+    assert result.exit_code == 0
+    assert "Default model" in result.output
+
+
+def test_clone_requires_ref_and_text():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["clone"])
+    assert result.exit_code != 0
+    assert "Missing option" in result.output