wfloat 0.0.1__tar.gz → 1.0.1__tar.gz

wfloat-1.0.1/MANIFEST.in

@@ -0,0 +1,2 @@
+include README.md
+recursive-include python *.py

wfloat-1.0.1/PKG-INFO

@@ -0,0 +1,172 @@
+Metadata-Version: 2.4
+Name: wfloat
+Version: 1.0.1
+Summary: High-level Python wrapper for Wfloat TTS
+Home-page: https://github.com/wfloat/wfloat-python
+Author: wfloat
+License: MIT
+Classifier: Programming Language :: Python :: 3
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Operating System :: MacOS :: MacOS X
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+Requires-Dist: wfloat-sherpa-onnx==1.12.24
+Dynamic: author
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# wfloat
+
+`wfloat` is the Python package for `wfloat-tts`, Wfloat's on-device English
+text-to-speech model.
+
+It runs speech locally in Python instead of calling a hosted inference API.
+The model supports 20 voices with emotion and intensity control.
+
+If you're building for the browser, use
+[`@wfloat/wfloat-web`](https://github.com/wfloat/wfloat-web). If you're
+building for React Native, use
+[`@wfloat/react-native-wfloat`](https://github.com/wfloat/react-native-wfloat).
+
+Try it in the browser: https://wfloat.com/demo
+
+<audio controls src="./sample.wav">
+  <a href="./sample.wav">Sample dialogue</a>
+</audio>
+
+[Sample dialogue](sample.wav)
+
+## Install
+
+```bash
+pip install wfloat
+```
+
+## Usage
+
+```python
+import wfloat
+
+model = wfloat.load("wfloat/wfloat-tts")
+
+result = model.generate(
+    text="No, no, that's not possible. The formula should have crystallized, but it adapted instead. Do you realize what that means for the rest of my work?",
+    voice_id="mad_scientist_woman",
+    emotion="surprise",
+    intensity=0.7,
+)
+
+result.audio.save("out.wav")
+```
+
+For multi-speaker dialogue:
+
+```python
+import wfloat
+
+model = wfloat.load("wfloat/wfloat-tts")
+
+result = model.generate_dialogue(
+    segments=[
+        {
+            "voice_id": "wise_elder_man",
+            "text": "Rain taps against the tavern shutters as you step inside.",
+            "emotion": "neutral",
+            "intensity": 0.5,
+        },
+        {
+            "voice_id": "strong_hero_man",
+            "text": "You're late. Two bandits stole the king's map over three hours ago.",
+            "emotion": "fear",
+            "intensity": 0.6,
+        },
+        {
+            "voice_id": "strong_hero_man",
+            "text": "They fled north, up into the woods.",
+            "emotion": "neutral",
+            "intensity": 0.5,
+        },
+    ],
+    silence_between_segments_sec=0.35,
+)
+
+result.audio.save("dialogue.wav")
+```
+
+You can also generate a WAV from the command line:
+
+```bash
+wfloat generate \
+  --text "Hello world!" \
+  --out out.wav \
+  --voice-id mad_scientist_woman \
+  --emotion surprise \
+  --intensity 0.7 \
+  --silence-padding-sec 0
+```
+
+For the full CLI help:
+
+```bash
+wfloat generate --help
+```
+
+The first load downloads the model assets. After that, the package uses the
+cached local copy.
+
+## Speaker IDs
+
+Use `voice_id` string names or numeric `sid` values:
+
+| Speaker | SID |
+| --- | ---: |
+| `skilled_hero_man` | 0 |
+| `skilled_hero_woman` | 1 |
+| `fun_hero_man` | 2 |
+| `fun_hero_woman` | 3 |
+| `strong_hero_man` | 4 |
+| `strong_hero_woman` | 5 |
+| `mad_scientist_man` | 6 |
+| `mad_scientist_woman` | 7 |
+| `clever_villain_man` | 8 |
+| `clever_villain_woman` | 9 |
+| `narrator_man` | 10 |
+| `narrator_woman` | 11 |
+| `wise_elder_man` | 12 |
+| `wise_elder_woman` | 13 |
+| `outgoing_anime_man` | 14 |
+| `outgoing_anime_woman` | 15 |
+| `scary_villain_man` | 16 |
+| `scary_villain_woman` | 17 |
+| `news_reporter_man` | 18 |
+| `news_reporter_woman` | 19 |
+
+## Emotions
+
+Supported emotion labels:
+
+- `neutral`
+- `joy`
+- `sadness`
+- `anger`
+- `fear`
+- `surprise`
+- `dismissive`
+- `confusion`
+
+`intensity` must be between `0.0` and `1.0`.
+
+## More
+
+- Docs: https://docs.wfloat.com
+- Model card, voices, emotions, and samples: https://huggingface.co/Wfloat/wfloat-tts
+- Web package: https://github.com/wfloat/wfloat-web
+- React Native package: https://github.com/wfloat/react-native-wfloat

wfloat-1.0.1/README.md

@@ -0,0 +1,147 @@
+# wfloat
+
+`wfloat` is the Python package for `wfloat-tts`, Wfloat's on-device English
+text-to-speech model.
+
+It runs speech locally in Python instead of calling a hosted inference API.
+The model supports 20 voices with emotion and intensity control.
+
+If you're building for the browser, use
+[`@wfloat/wfloat-web`](https://github.com/wfloat/wfloat-web). If you're
+building for React Native, use
+[`@wfloat/react-native-wfloat`](https://github.com/wfloat/react-native-wfloat).
+
+Try it in the browser: https://wfloat.com/demo
+
+<audio controls src="./sample.wav">
+  <a href="./sample.wav">Sample dialogue</a>
+</audio>
+
+[Sample dialogue](sample.wav)
+
+## Install
+
+```bash
+pip install wfloat
+```
+
+## Usage
+
+```python
+import wfloat
+
+model = wfloat.load("wfloat/wfloat-tts")
+
+result = model.generate(
+    text="No, no, that's not possible. The formula should have crystallized, but it adapted instead. Do you realize what that means for the rest of my work?",
+    voice_id="mad_scientist_woman",
+    emotion="surprise",
+    intensity=0.7,
+)
+
+result.audio.save("out.wav")
+```
+
+For multi-speaker dialogue:
+
+```python
+import wfloat
+
+model = wfloat.load("wfloat/wfloat-tts")
+
+result = model.generate_dialogue(
+    segments=[
+        {
+            "voice_id": "wise_elder_man",
+            "text": "Rain taps against the tavern shutters as you step inside.",
+            "emotion": "neutral",
+            "intensity": 0.5,
+        },
+        {
+            "voice_id": "strong_hero_man",
+            "text": "You're late. Two bandits stole the king's map over three hours ago.",
+            "emotion": "fear",
+            "intensity": 0.6,
+        },
+        {
+            "voice_id": "strong_hero_man",
+            "text": "They fled north, up into the woods.",
+            "emotion": "neutral",
+            "intensity": 0.5,
+        },
+    ],
+    silence_between_segments_sec=0.35,
+)
+
+result.audio.save("dialogue.wav")
+```
+
+You can also generate a WAV from the command line:
+
+```bash
+wfloat generate \
+  --text "Hello world!" \
+  --out out.wav \
+  --voice-id mad_scientist_woman \
+  --emotion surprise \
+  --intensity 0.7 \
+  --silence-padding-sec 0
+```
+
+For the full CLI help:
+
+```bash
+wfloat generate --help
+```
+
+The first load downloads the model assets. After that, the package uses the
+cached local copy.
+
+## Speaker IDs
+
+Use `voice_id` string names or numeric `sid` values:
+
+| Speaker | SID |
+| --- | ---: |
+| `skilled_hero_man` | 0 |
+| `skilled_hero_woman` | 1 |
+| `fun_hero_man` | 2 |
+| `fun_hero_woman` | 3 |
+| `strong_hero_man` | 4 |
+| `strong_hero_woman` | 5 |
+| `mad_scientist_man` | 6 |
+| `mad_scientist_woman` | 7 |
+| `clever_villain_man` | 8 |
+| `clever_villain_woman` | 9 |
+| `narrator_man` | 10 |
+| `narrator_woman` | 11 |
+| `wise_elder_man` | 12 |
+| `wise_elder_woman` | 13 |
+| `outgoing_anime_man` | 14 |
+| `outgoing_anime_woman` | 15 |
+| `scary_villain_man` | 16 |
+| `scary_villain_woman` | 17 |
+| `news_reporter_man` | 18 |
+| `news_reporter_woman` | 19 |
+
+## Emotions
+
+Supported emotion labels:
+
+- `neutral`
+- `joy`
+- `sadness`
+- `anger`
+- `fear`
+- `surprise`
+- `dismissive`
+- `confusion`
+
+`intensity` must be between `0.0` and `1.0`.
+
+## More
+
+- Docs: https://docs.wfloat.com
+- Model card, voices, emotions, and samples: https://huggingface.co/Wfloat/wfloat-tts
+- Web package: https://github.com/wfloat/wfloat-web
+- React Native package: https://github.com/wfloat/react-native-wfloat

wfloat-1.0.1/pyproject.toml

@@ -0,0 +1,3 @@
+[build-system]
+requires = ["setuptools>=68", "wheel"]
+build-backend = "setuptools.build_meta"

wfloat-1.0.1/python/wfloat/__init__.py

@@ -0,0 +1,47 @@
+from ._constants import SPEAKER_IDS, VALID_EMOTIONS, VALID_SIDS
+from ._model import Model, load
+from ._results import Audio, GenerationResult, Timeline, TimelineChunk
+from ._version import __version__
+
+_LOW_LEVEL_EXPORTS = {
+    "GenerationConfig",
+    "OfflineTts",
+    "OfflineTtsConfig",
+    "OfflineTtsModelConfig",
+    "OfflineTtsWfloatModelConfig",
+    "WfloatPreparedText",
+    "git_date",
+    "git_sha1",
+    "prepare_wfloat_text",
+    "version",
+    "write_wave",
+}
+
+
+__all__ = [
+    "Audio",
+    "GenerationResult",
+    "Model",
+    "SPEAKER_IDS",
+    "Timeline",
+    "TimelineChunk",
+    "VALID_EMOTIONS",
+    "VALID_SIDS",
+    "load",
+]
+__all__.extend(sorted(_LOW_LEVEL_EXPORTS))
+
+
+def __getattr__(name):
+    if name not in _LOW_LEVEL_EXPORTS:
+        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
+
+    from . import _bindings
+
+    value = getattr(_bindings, name)
+    globals()[name] = value
+    return value
+
+
+def __dir__():
+    return sorted(set(globals()) | _LOW_LEVEL_EXPORTS)

wfloat-1.0.1/python/wfloat/__main__.py

@@ -0,0 +1,5 @@
+from ._cli import main
+
+
+if __name__ == "__main__":  # pragma: no cover
+    raise SystemExit(main())

wfloat-1.0.1/python/wfloat/_assets.py

@@ -0,0 +1,124 @@
+import json
+import os
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Dict, Optional
+from urllib.parse import urlencode, urlparse
+from urllib.request import Request, urlopen
+
+from ._version import __version__ as PACKAGE_VERSION
+
+
+DEFAULT_MODEL_ASSET_HOST = "https://wfloat.com"
+DEFAULT_MODEL_ASSET_PATH = "/api/model-assets"
+
+
+@dataclass(frozen=True)
+class ModelAssets:
+    model_onnx: str
+    model_onnx_checksum: str
+    model_tokens: str
+    model_tokens_checksum: str
+    espeak_data: str
+    espeak_checksum: str
+    persistent_id: Optional[str] = None
+
+    @classmethod
+    def from_dict(cls, data: Dict[str, object]) -> "ModelAssets":
+        required_fields = (
+            "model_onnx",
+            "model_onnx_checksum",
+            "model_tokens",
+            "model_tokens_checksum",
+            "espeak_data",
+            "espeak_checksum",
+        )
+
+        missing = [
+            field_name
+            for field_name in required_fields
+            if not isinstance(data.get(field_name), str) or not str(data.get(field_name)).strip()
+        ]
+        if missing:
+            raise ValueError(
+                "Model asset response is missing required fields: %s"
+                % ", ".join(missing)
+            )
+
+        return cls(
+            model_onnx=str(data["model_onnx"]),
+            model_onnx_checksum=str(data["model_onnx_checksum"]),
+            model_tokens=str(data["model_tokens"]),
+            model_tokens_checksum=str(data["model_tokens_checksum"]),
+            espeak_data=str(data["espeak_data"]),
+            espeak_checksum=str(data["espeak_checksum"]),
+            persistent_id=str(data["persistent_id"]).strip()
+            if isinstance(data.get("persistent_id"), str) and str(data.get("persistent_id")).strip()
+            else None,
+        )
+
+    def to_dict(self) -> Dict[str, str]:
+        return {
+            "model_onnx": self.model_onnx,
+            "model_onnx_checksum": self.model_onnx_checksum,
+            "model_tokens": self.model_tokens,
+            "model_tokens_checksum": self.model_tokens_checksum,
+            "espeak_data": self.espeak_data,
+            "espeak_checksum": self.espeak_checksum,
+            **({"persistent_id": self.persistent_id} if self.persistent_id else {}),
+        }
+
+
+def get_package_version(default: str = "0.0.0") -> str:
+    return PACKAGE_VERSION or default
+
+
+def get_model_asset_host() -> str:
+    return os.environ.get("WFLOAT_MODEL_ASSET_HOST", DEFAULT_MODEL_ASSET_HOST).rstrip("/")
+
+
+def filename_from_url(url: str, fallback: str) -> str:
+    parsed = urlparse(url)
+    filename = Path(parsed.path).name
+    return filename or fallback
+
+
+def fetch_model_assets(
+    model_name: str,
+    *,
+    persistent_id: Optional[str] = None,
+    package_version_override: Optional[str] = None,
+    timeout: float = 60.0,
+) -> ModelAssets:
+    version = package_version_override or get_package_version()
+    query = {
+        "platform": "python",
+        "version": version,
+        "model_name": model_name,
+    }
+    if persistent_id:
+        query["persistent_id"] = persistent_id
+
+    params = urlencode(query)
+    url = "%s%s?%s" % (get_model_asset_host(), DEFAULT_MODEL_ASSET_PATH, params)
+    request = Request(
+        url,
+        headers={
+            "Accept": "application/json",
+            "User-Agent": "wfloat-python/%s" % version,
+        },
+        method="GET",
+    )
+
+    with urlopen(request, timeout=timeout) as response:
+        payload = response.read().decode("utf-8")
+
+    try:
+        data = json.loads(payload)
+    except json.JSONDecodeError as exc:
+        raise RuntimeError("Failed to decode model asset response JSON.") from exc
+
+    if not isinstance(data, dict):
+        raise RuntimeError("Model asset response must be a JSON object.")
+
+    return ModelAssets.from_dict(data)

wfloat-1.0.1/python/wfloat/_bindings.py

@@ -0,0 +1,58 @@
+try:
+    import sherpa_onnx
+except ImportError as exc:
+    raise ImportError(
+        "Failed to import sherpa_onnx. "
+        "Reinstall wfloat so pip can install the matching wfloat-sherpa-onnx dependency."
+    ) from exc
+
+
+_REQUIRED_EXPORTS = (
+    "GenerationConfig",
+    "OfflineTts",
+    "OfflineTtsConfig",
+    "OfflineTtsModelConfig",
+    "OfflineTtsWfloatModelConfig",
+    "WfloatPreparedText",
+    "git_date",
+    "git_sha1",
+    "prepare_wfloat_text",
+    "version",
+    "write_wave",
+)
+
+missing_exports = [name for name in _REQUIRED_EXPORTS if not hasattr(sherpa_onnx, name)]
+if missing_exports:
+    raise ImportError(
+        "Installed sherpa_onnx is missing required exports: "
+        f"{', '.join(missing_exports)}. "
+        "Reinstall wfloat so pip can install a compatible wfloat-sherpa-onnx build."
+    )
+
+
+GenerationConfig = sherpa_onnx.GenerationConfig
+OfflineTts = sherpa_onnx.OfflineTts
+OfflineTtsConfig = sherpa_onnx.OfflineTtsConfig
+OfflineTtsModelConfig = sherpa_onnx.OfflineTtsModelConfig
+OfflineTtsWfloatModelConfig = sherpa_onnx.OfflineTtsWfloatModelConfig
+WfloatPreparedText = sherpa_onnx.WfloatPreparedText
+git_date = sherpa_onnx.git_date
+git_sha1 = sherpa_onnx.git_sha1
+prepare_wfloat_text = sherpa_onnx.prepare_wfloat_text
+version = sherpa_onnx.version
+write_wave = sherpa_onnx.write_wave
+
+
+__all__ = [
+    "GenerationConfig",
+    "OfflineTts",
+    "OfflineTtsConfig",
+    "OfflineTtsModelConfig",
+    "OfflineTtsWfloatModelConfig",
+    "WfloatPreparedText",
+    "git_date",
+    "git_sha1",
+    "prepare_wfloat_text",
+    "version",
+    "write_wave",
+]