vision-agents-plugins-pocket 0.2.10__tar.gz

vision_agents_plugins_pocket-0.2.10/.gitignore

@@ -0,0 +1,90 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.cursor/*
+# Distribution / packaging
+.Python
+build/
+dist/
+downloads/
+develop-eggs/
+eggs/
+.eggs/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+pip-wheel-metadata/
+MANIFEST
+*.egg-info/
+*.egg
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+coverage.xml
+nosetests.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Type checker / lint caches
+.mypy_cache/
+.dmypy.json
+dmypy.json
+.pytype/
+.pyre/
+.ruff_cache/
+
+# Environments
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+.env
+.env.local
+.env.*.local
+.env.bak
+pyvenv.cfg
+.python-version
+
+# Editors / IDEs
+.vscode/
+.idea/
+
+# Jupyter Notebook
+.ipynb_checkpoints/
+
+# OS / Misc
+.DS_Store
+*.log
+
+# Tooling & repo-specific
+pyrightconfig.json
+shell.nix
+bin/*
+lib/*
+stream-py/
+
+# Artifacts / assets
+*.pt
+*.kef
+*.onnx
+profile.html
+
+/opencode.json

vision_agents_plugins_pocket-0.2.10/PKG-INFO

@@ -0,0 +1,68 @@
+Metadata-Version: 2.4
+Name: vision-agents-plugins-pocket
+Version: 0.2.10
+Summary: Pocket TTS integration for Vision Agents - lightweight CPU-based text-to-speech
+Project-URL: Documentation, https://visionagents.ai/
+Project-URL: Website, https://visionagents.ai/
+Project-URL: Source, https://github.com/GetStream/Vision-Agents
+License-Expression: MIT
+Keywords: AI,TTS,agents,kyutai,pocket-tts,text-to-speech,voice agents
+Requires-Python: >=3.10
+Requires-Dist: pocket-tts>=0.1.0
+Requires-Dist: vision-agents
+Description-Content-Type: text/markdown
+
+# Pocket TTS Plugin
+
+A lightweight Text-to-Speech (TTS) plugin for [Vision Agents](https://github.com/GetStream/Vision-Agents) powered by Kyutai's [Pocket TTS](https://huggingface.co/kyutai/pocket-tts) model. Runs efficiently on CPU with low latency (~200ms) and supports voice cloning.
+
+## Features
+
+- Runs on CPU - no GPU required
+- Small model size (100M parameters)
+- Low latency (~200ms to first audio)
+- Voice cloning support
+- Built-in voice selection
+
+## Installation
+
+```bash
+uv add vision-agents[pocket]
+```
+
+## Usage
+
+```python
+from vision_agents.plugins import pocket
+
+# Create TTS with default voice
+tts = pocket.TTS()
+
+# Or specify a built-in voice
+tts = pocket.TTS(voice="marius")
+
+# Or use a custom voice for cloning
+tts = pocket.TTS(voice="path/to/your/voice.wav")
+```
+
+## Configuration
+
+| Parameter | Description | Values |
+|-----------|-------------|--------|
+| `voice` | Built-in voice name or path to custom wav file | `"alba"` (default), `"marius"`, `"javert"`, `"jean"`, `"fantine"`, `"cosette"`, `"eponine"`, `"azelma"`, or custom path |
+
+## Built-in Voices
+
+- `alba` - Default voice
+- `marius`
+- `javert`
+- `jean`
+- `fantine`
+- `cosette`
+- `eponine`
+- `azelma`
+
+## Dependencies
+
+- pocket-tts>=0.1.0
+- PyTorch 2.5+

vision_agents_plugins_pocket-0.2.10/README.md

@@ -0,0 +1,54 @@
+# Pocket TTS Plugin
+
+A lightweight Text-to-Speech (TTS) plugin for [Vision Agents](https://github.com/GetStream/Vision-Agents) powered by Kyutai's [Pocket TTS](https://huggingface.co/kyutai/pocket-tts) model. Runs efficiently on CPU with low latency (~200ms) and supports voice cloning.
+
+## Features
+
+- Runs on CPU - no GPU required
+- Small model size (100M parameters)
+- Low latency (~200ms to first audio)
+- Voice cloning support
+- Built-in voice selection
+
+## Installation
+
+```bash
+uv add vision-agents[pocket]
+```
+
+## Usage
+
+```python
+from vision_agents.plugins import pocket
+
+# Create TTS with default voice
+tts = pocket.TTS()
+
+# Or specify a built-in voice
+tts = pocket.TTS(voice="marius")
+
+# Or use a custom voice for cloning
+tts = pocket.TTS(voice="path/to/your/voice.wav")
+```
+
+## Configuration
+
+| Parameter | Description | Values |
+|-----------|-------------|--------|
+| `voice` | Built-in voice name or path to custom wav file | `"alba"` (default), `"marius"`, `"javert"`, `"jean"`, `"fantine"`, `"cosette"`, `"eponine"`, `"azelma"`, or custom path |
+
+## Built-in Voices
+
+- `alba` - Default voice
+- `marius`
+- `javert`
+- `jean`
+- `fantine`
+- `cosette`
+- `eponine`
+- `azelma`
+
+## Dependencies
+
+- pocket-tts>=0.1.0
+- PyTorch 2.5+

vision_agents_plugins_pocket-0.2.10/pyproject.toml

@@ -0,0 +1,40 @@
+[build-system]
+requires = ["hatchling", "hatch-vcs"]
+build-backend = "hatchling.build"
+
+[project]
+name = "vision-agents-plugins-pocket"
+dynamic = ["version"]
+description = "Pocket TTS integration for Vision Agents - lightweight CPU-based text-to-speech"
+readme = "README.md"
+keywords = ["pocket-tts", "TTS", "text-to-speech", "AI", "voice agents", "agents", "kyutai"]
+requires-python = ">=3.10"
+license = "MIT"
+dependencies = [
+    "vision-agents",
+    "pocket-tts>=0.1.0",
+]
+
+[project.urls]
+Documentation = "https://visionagents.ai/"
+Website = "https://visionagents.ai/"
+Source = "https://github.com/GetStream/Vision-Agents"
+
+[tool.hatch.version]
+source = "vcs"
+raw-options = { root = "..", search_parent_directories = true, fallback_version = "0.0.0" }
+
+[tool.hatch.build.targets.wheel]
+packages = [".", "vision_agents"]
+
+[tool.hatch.build.targets.sdist]
+include = ["/vision_agents"]
+
+[tool.uv.sources]
+vision-agents = { workspace = true }
+
+[dependency-groups]
+dev = [
+    "pytest>=8.4.1",
+    "pytest-asyncio>=1.0.0",
+]

vision_agents_plugins_pocket-0.2.10/vision_agents/plugins/pocket/__init__.py

@@ -0,0 +1,3 @@
+from .tts import TTS
+
+__all__ = ["TTS"]

vision_agents_plugins_pocket-0.2.10/vision_agents/plugins/pocket/tts.py

@@ -0,0 +1,129 @@
+import asyncio
+import logging
+from concurrent.futures import ThreadPoolExecutor
+from typing import Any, AsyncIterator, Iterator, Literal
+
+import numpy as np
+
+from getstream.video.rtc.track_util import AudioFormat, PcmData
+from vision_agents.core import tts
+from vision_agents.core.warmup import Warmable
+
+from pocket_tts import TTSModel
+
+logger = logging.getLogger(__name__)
+
+Voice = Literal[
+    "alba", "marius", "javert", "jean", "fantine", "cosette", "eponine", "azelma"
+]
+
+VOICE_PATHS = {
+    "alba": "hf://kyutai/tts-voices/alba-mackenna/casual.wav",
+    "marius": "hf://kyutai/tts-voices/marius/casual.wav",
+    "javert": "hf://kyutai/tts-voices/javert/casual.wav",
+    "jean": "hf://kyutai/tts-voices/jean/casual.wav",
+    "fantine": "hf://kyutai/tts-voices/fantine/casual.wav",
+    "cosette": "hf://kyutai/tts-voices/cosette/casual.wav",
+    "eponine": "hf://kyutai/tts-voices/eponine/casual.wav",
+    "azelma": "hf://kyutai/tts-voices/azelma/casual.wav",
+}
+
+
+class TTS(tts.TTS, Warmable[tuple[TTSModel, Any]]):
+    """
+    Pocket TTS Text-to-Speech implementation.
+
+    A lightweight CPU-based TTS model from Kyutai with ~200ms latency
+    and voice cloning support.
+    """
+
+    def __init__(
+        self,
+        voice: Voice | str = "alba",
+        client: TTSModel | None = None,
+    ) -> None:
+        """
+        Initialize Pocket TTS.
+
+        Args:
+            voice: Built-in voice name or path to custom wav file for voice cloning.
+            client: Optional pre-initialized TTSModel instance.
+        """
+        super().__init__(provider_name="pocket")
+
+        self.voice = voice
+        self._model: TTSModel | None = client
+        self._voice_state = None
+        self._executor = ThreadPoolExecutor(max_workers=4)
+
+    async def on_warmup(self) -> tuple[TTSModel, Any]:
+        if self._model is not None and self._voice_state is not None:
+            return (self._model, self._voice_state)
+
+        loop = asyncio.get_running_loop()
+
+        if self._model is not None:
+            model = self._model
+        else:
+            logger.info("Loading Pocket TTS model...")
+            model = await loop.run_in_executor(self._executor, TTSModel.load_model)
+            logger.info("Pocket TTS model loaded")
+
+        voice_path = VOICE_PATHS.get(self.voice, self.voice)
+        logger.info(f"Loading voice state for: {self.voice}")
+        voice_state = await loop.run_in_executor(
+            self._executor,
+            lambda: model.get_state_for_audio_prompt(voice_path),
+        )
+        logger.info("Voice state loaded")
+        return (model, voice_state)
+
+    def on_warmed_up(self, resource: tuple[TTSModel, Any]) -> None:
+        self._model, self._voice_state = resource
+
+    async def _ensure_loaded(self) -> None:
+        """Ensure model and voice state are loaded."""
+        if self._model is None or self._voice_state is None:
+            resource = await self.on_warmup()
+            self.on_warmed_up(resource)
+
+    async def stream_audio(
+        self, text: str, *_, **__
+    ) -> PcmData | Iterator[PcmData] | AsyncIterator[PcmData]:
+        """
+        Convert text to speech using Pocket TTS.
+
+        Args:
+            text: The text to convert to speech.
+
+        Returns:
+            PcmData containing the synthesized audio.
+        """
+        await self._ensure_loaded()
+        assert self._model is not None
+        assert self._voice_state is not None
+
+        model = self._model
+        voice_state = self._voice_state
+
+        def _generate():
+            audio_tensor = model.generate_audio(voice_state, text)
+            audio_np = audio_tensor.numpy()
+            pcm16 = (np.clip(audio_np, -1.0, 1.0) * 32767.0).astype(np.int16)
+            return pcm16, model.sample_rate
+
+        loop = asyncio.get_running_loop()
+        samples, sample_rate = await loop.run_in_executor(self._executor, _generate)
+
+        return PcmData.from_numpy(
+            samples, sample_rate=sample_rate, channels=1, format=AudioFormat.S16
+        )
+
+    async def stop_audio(self) -> None:
+        """Stop audio playback (no-op for Pocket TTS)."""
+        logger.info("Pocket TTS stop requested (no-op)")
+
+    async def close(self) -> None:
+        """Close the TTS and cleanup resources."""
+        await super().close()
+        self._executor.shutdown(wait=False)