Spaces:
Running
Running
Refactor .gitignore and enhance Pocket TTS and Moonshine STT: Simplify .gitignore by removing unnecessary entries and adding environment-specific files. Update Pocket TTS to use a configurable sample rate and replace UUID with shortuuid for request IDs. Modify Moonshine STT to use NotGivenOr for language parameter and update stream class inheritance for better clarity.
Browse files- .gitignore +23 -182
- src/agent/agent.py +1 -0
- src/plugins/moonshine_stt/stt.py +6 -7
- src/plugins/pocket_tts/tts.py +8 -7
.gitignore
CHANGED
|
@@ -1,191 +1,32 @@
|
|
| 1 |
-
#
|
| 2 |
-
|
|
|
|
|
|
|
| 3 |
|
| 4 |
-
#
|
| 5 |
-
# Byte-compiled / optimized / DLL files
|
| 6 |
__pycache__/
|
| 7 |
-
*.
|
| 8 |
-
*
|
| 9 |
-
|
| 10 |
-
# C extensions
|
| 11 |
-
*.so
|
| 12 |
-
|
| 13 |
-
# Distribution / packaging
|
| 14 |
-
.Python
|
| 15 |
-
build/
|
| 16 |
-
develop-eggs/
|
| 17 |
-
dist/
|
| 18 |
-
downloads/
|
| 19 |
-
eggs/
|
| 20 |
-
.eggs/
|
| 21 |
-
lib/
|
| 22 |
-
lib64/
|
| 23 |
-
parts/
|
| 24 |
-
sdist/
|
| 25 |
-
var/
|
| 26 |
-
wheels/
|
| 27 |
-
share/python-wheels/
|
| 28 |
*.egg-info/
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
MANIFEST
|
| 32 |
-
|
| 33 |
-
# PyInstaller
|
| 34 |
-
# Usually these files are written by a python script from a template
|
| 35 |
-
# before PyInstaller builds the exe, so as to inject date/other infos into it.
|
| 36 |
-
*.manifest
|
| 37 |
-
*.spec
|
| 38 |
-
|
| 39 |
-
# Installer logs
|
| 40 |
-
pip-log.txt
|
| 41 |
-
pip-delete-this-directory.txt
|
| 42 |
-
|
| 43 |
-
# Translations
|
| 44 |
-
*.mo
|
| 45 |
-
*.pot
|
| 46 |
-
|
| 47 |
-
# Django stuff:
|
| 48 |
-
*.log
|
| 49 |
-
local_settings.py
|
| 50 |
-
db.sqlite3
|
| 51 |
-
db.sqlite3-journal
|
| 52 |
-
|
| 53 |
-
# Flask stuff:
|
| 54 |
-
instance/
|
| 55 |
-
.webassets-cache
|
| 56 |
-
|
| 57 |
-
# Scrapy stuff:
|
| 58 |
-
.scrapy
|
| 59 |
-
|
| 60 |
-
# Sphinx documentation
|
| 61 |
-
docs/_build/
|
| 62 |
-
|
| 63 |
-
# PyBuilder
|
| 64 |
-
.pybuilder/
|
| 65 |
-
target/
|
| 66 |
-
|
| 67 |
-
# Jupyter Notebook
|
| 68 |
-
.ipynb_checkpoints
|
| 69 |
-
|
| 70 |
-
# IPython
|
| 71 |
-
profile_default/
|
| 72 |
-
ipython_config.py
|
| 73 |
-
|
| 74 |
-
# pyenv
|
| 75 |
-
# For a library or package, you might want to ignore these files since the code is
|
| 76 |
-
# intended to run in multiple environments; otherwise, check them in:
|
| 77 |
-
# .python-version
|
| 78 |
-
|
| 79 |
-
# pipenv
|
| 80 |
-
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
| 81 |
-
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
| 82 |
-
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
| 83 |
-
# install all needed dependencies.
|
| 84 |
-
#Pipfile.lock
|
| 85 |
-
|
| 86 |
-
# poetry
|
| 87 |
-
# Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
|
| 88 |
-
# This is especially recommended for binary packages to ensure reproducibility, and is more
|
| 89 |
-
# commonly ignored for libraries.
|
| 90 |
-
# https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
|
| 91 |
-
#poetry.lock
|
| 92 |
-
|
| 93 |
-
# pdm
|
| 94 |
-
# Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
|
| 95 |
-
#pdm.lock
|
| 96 |
-
# pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
|
| 97 |
-
# in version control.
|
| 98 |
-
# https://pdm.fming.dev/#use-with-ide
|
| 99 |
-
.pdm.toml
|
| 100 |
-
|
| 101 |
-
# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
|
| 102 |
-
__pypackages__/
|
| 103 |
-
|
| 104 |
-
# Celery stuff
|
| 105 |
-
celerybeat-schedule
|
| 106 |
-
celerybeat.pid
|
| 107 |
|
| 108 |
-
#
|
| 109 |
-
|
|
|
|
|
|
|
|
|
|
| 110 |
|
| 111 |
-
#
|
| 112 |
-
.env
|
| 113 |
-
.venv
|
| 114 |
-
env/
|
| 115 |
-
venv/
|
| 116 |
-
ENV/
|
| 117 |
-
env.bak/
|
| 118 |
-
venv.bak/
|
| 119 |
dev/
|
| 120 |
-
nvidia_services/cache/asr/
|
| 121 |
-
nvidia_services/cache/tts/
|
| 122 |
-
.claude/
|
| 123 |
-
.cursor/
|
| 124 |
.pytest_cache/
|
| 125 |
-
# Spyder project settings
|
| 126 |
-
.spyderproject
|
| 127 |
-
.spyproject
|
| 128 |
-
|
| 129 |
-
# Rope project settings
|
| 130 |
-
.ropeproject
|
| 131 |
-
|
| 132 |
-
# mkdocs documentation
|
| 133 |
-
/site
|
| 134 |
-
|
| 135 |
-
# mypy
|
| 136 |
-
.mypy_cache/
|
| 137 |
-
.dmypy.json
|
| 138 |
-
dmypy.json
|
| 139 |
-
|
| 140 |
-
# Pyre type checker
|
| 141 |
-
.pyre/
|
| 142 |
-
|
| 143 |
-
# pytype static type analyzer
|
| 144 |
-
.pytype/
|
| 145 |
-
|
| 146 |
-
# Cython debug symbols
|
| 147 |
-
cython_debug/
|
| 148 |
-
|
| 149 |
-
# PyCharm
|
| 150 |
-
# JetBrains specific template is maintained in a separate JetBrains.gitignore that can
|
| 151 |
-
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
|
| 152 |
-
# and can be added to the global gitignore or merged into this file. For a more nuclear
|
| 153 |
-
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
|
| 154 |
-
#.idea/
|
| 155 |
-
|
| 156 |
-
### Python Patch ###
|
| 157 |
-
# Poetry local configuration file - https://python-poetry.org/docs/configuration/#local-configuration
|
| 158 |
-
poetry.toml
|
| 159 |
-
|
| 160 |
-
# ruff
|
| 161 |
-
.ruff_cache/
|
| 162 |
-
|
| 163 |
-
# LSP config files
|
| 164 |
-
pyrightconfig.json
|
| 165 |
-
|
| 166 |
-
### PythonVanilla ###
|
| 167 |
-
# Byte-compiled / optimized / DLL files
|
| 168 |
-
|
| 169 |
-
# C extensions
|
| 170 |
-
|
| 171 |
-
# Distribution / packaging
|
| 172 |
-
|
| 173 |
-
# Installer logs
|
| 174 |
-
|
| 175 |
-
# Translations
|
| 176 |
-
|
| 177 |
-
# pyenv
|
| 178 |
-
# For a library or package, you might want to ignore these files since the code is
|
| 179 |
-
# intended to run in multiple environments; otherwise, check them in:
|
| 180 |
-
# .python-version
|
| 181 |
-
|
| 182 |
-
# pipenv
|
| 183 |
-
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
| 184 |
-
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
| 185 |
-
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
| 186 |
-
# install all needed dependencies.
|
| 187 |
-
|
| 188 |
-
# PEP 582; used by e.g. github.com/David-OConnor/pyflow
|
| 189 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 190 |
|
| 191 |
-
#
|
|
|
|
|
|
| 1 |
+
# Environment
|
| 2 |
+
.env
|
| 3 |
+
.venv/
|
| 4 |
+
.streamlit/
|
| 5 |
|
| 6 |
+
# Python
|
|
|
|
| 7 |
__pycache__/
|
| 8 |
+
*.pyc
|
| 9 |
+
*.pyo
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
*.egg-info/
|
| 11 |
+
dist/
|
| 12 |
+
build/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
+
# IDE
|
| 15 |
+
.cursor/
|
| 16 |
+
.cursorignore
|
| 17 |
+
.claude/
|
| 18 |
+
CLAUDE.md
|
| 19 |
|
| 20 |
+
# Dev files
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
dev/
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
.pytest_cache/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
|
| 24 |
+
# Model weights (never commit)
|
| 25 |
+
*.onnx
|
| 26 |
+
*.pt
|
| 27 |
+
*.bin
|
| 28 |
+
*.safetensors
|
| 29 |
+
model_cache/
|
| 30 |
|
| 31 |
+
# OS
|
| 32 |
+
.DS_Store
|
src/agent/agent.py
CHANGED
|
@@ -91,6 +91,7 @@ async def session_handler(ctx: agents.JobContext) -> None:
|
|
| 91 |
voice=settings.voice.POCKET_TTS_VOICE,
|
| 92 |
temperature=settings.voice.POCKET_TTS_TEMPERATURE,
|
| 93 |
lsd_decode_steps=settings.voice.POCKET_TTS_LSD_DECODE_STEPS,
|
|
|
|
| 94 |
metrics_callback=tts_metrics_callback,
|
| 95 |
)
|
| 96 |
|
|
|
|
| 91 |
voice=settings.voice.POCKET_TTS_VOICE,
|
| 92 |
temperature=settings.voice.POCKET_TTS_TEMPERATURE,
|
| 93 |
lsd_decode_steps=settings.voice.POCKET_TTS_LSD_DECODE_STEPS,
|
| 94 |
+
sample_rate=settings.voice.SAMPLE_RATE_OUTPUT,
|
| 95 |
metrics_callback=tts_metrics_callback,
|
| 96 |
)
|
| 97 |
|
src/plugins/moonshine_stt/stt.py
CHANGED
|
@@ -10,8 +10,8 @@ import torch
|
|
| 10 |
from transformers import AutoProcessor, MoonshineStreamingForConditionalGeneration
|
| 11 |
from livekit import rtc
|
| 12 |
from livekit.agents import stt
|
| 13 |
-
from livekit.agents.types import APIConnectOptions, DEFAULT_API_CONNECT_OPTIONS
|
| 14 |
-
from livekit.agents.utils import AudioBuffer
|
| 15 |
|
| 16 |
|
| 17 |
@dataclass
|
|
@@ -49,7 +49,7 @@ class MoonshineSTT(stt.STT):
|
|
| 49 |
self,
|
| 50 |
buffer: AudioBuffer,
|
| 51 |
*,
|
| 52 |
-
language: str
|
| 53 |
conn_options: APIConnectOptions,
|
| 54 |
) -> stt.SpeechEvent:
|
| 55 |
config = self._sanitize_options(language=language)
|
|
@@ -85,7 +85,7 @@ class MoonshineSTT(stt.STT):
|
|
| 85 |
*,
|
| 86 |
language: str | None = None,
|
| 87 |
conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
|
| 88 |
-
) -> stt.
|
| 89 |
config = self._sanitize_options(language=language)
|
| 90 |
return MoonshineSTTStream(
|
| 91 |
stt=self,
|
|
@@ -98,7 +98,7 @@ class MoonshineSTT(stt.STT):
|
|
| 98 |
)
|
| 99 |
|
| 100 |
|
| 101 |
-
class MoonshineSTTStream(stt.
|
| 102 |
def __init__(
|
| 103 |
self,
|
| 104 |
*,
|
|
@@ -148,8 +148,7 @@ class MoonshineSTTStream(stt.SpeechStream):
|
|
| 148 |
|
| 149 |
async def _finalize_segment(self) -> None:
|
| 150 |
# Generate a unique request ID for this segment
|
| 151 |
-
|
| 152 |
-
request_id = str(uuid.uuid4())
|
| 153 |
|
| 154 |
if len(self._buffer) == 0:
|
| 155 |
# Don't emit metrics for empty segments - just return
|
|
|
|
| 10 |
from transformers import AutoProcessor, MoonshineStreamingForConditionalGeneration
|
| 11 |
from livekit import rtc
|
| 12 |
from livekit.agents import stt
|
| 13 |
+
from livekit.agents.types import APIConnectOptions, DEFAULT_API_CONNECT_OPTIONS, NOT_GIVEN, NotGivenOr
|
| 14 |
+
from livekit.agents.utils import AudioBuffer, shortuuid
|
| 15 |
|
| 16 |
|
| 17 |
@dataclass
|
|
|
|
| 49 |
self,
|
| 50 |
buffer: AudioBuffer,
|
| 51 |
*,
|
| 52 |
+
language: NotGivenOr[str] = NOT_GIVEN,
|
| 53 |
conn_options: APIConnectOptions,
|
| 54 |
) -> stt.SpeechEvent:
|
| 55 |
config = self._sanitize_options(language=language)
|
|
|
|
| 85 |
*,
|
| 86 |
language: str | None = None,
|
| 87 |
conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
|
| 88 |
+
) -> stt.RecognizeStream:
|
| 89 |
config = self._sanitize_options(language=language)
|
| 90 |
return MoonshineSTTStream(
|
| 91 |
stt=self,
|
|
|
|
| 98 |
)
|
| 99 |
|
| 100 |
|
| 101 |
+
class MoonshineSTTStream(stt.RecognizeStream):
|
| 102 |
def __init__(
|
| 103 |
self,
|
| 104 |
*,
|
|
|
|
| 148 |
|
| 149 |
async def _finalize_segment(self) -> None:
|
| 150 |
# Generate a unique request ID for this segment
|
| 151 |
+
request_id = shortuuid("STT_")
|
|
|
|
| 152 |
|
| 153 |
if len(self._buffer) == 0:
|
| 154 |
# Don't emit metrics for empty segments - just return
|
src/plugins/pocket_tts/tts.py
CHANGED
|
@@ -4,7 +4,6 @@ from __future__ import annotations
|
|
| 4 |
import asyncio
|
| 5 |
import logging
|
| 6 |
import time
|
| 7 |
-
import uuid
|
| 8 |
from typing import Callable
|
| 9 |
|
| 10 |
import numpy as np
|
|
@@ -14,9 +13,9 @@ from scipy import signal
|
|
| 14 |
|
| 15 |
from livekit.agents import tts
|
| 16 |
from livekit.agents.types import APIConnectOptions, DEFAULT_API_CONNECT_OPTIONS
|
|
|
|
| 17 |
|
| 18 |
-
|
| 19 |
-
from src.core.settings import settings
|
| 20 |
|
| 21 |
# Reduce verbosity of pocket_tts library to avoid console spam
|
| 22 |
logging.getLogger("pocket_tts").setLevel(logging.WARNING)
|
|
@@ -34,6 +33,7 @@ class PocketTTS(tts.TTS):
|
|
| 34 |
voice: str = "alba",
|
| 35 |
temperature: float = 0.7,
|
| 36 |
lsd_decode_steps: int = 1,
|
|
|
|
| 37 |
metrics_callback: OptionalTTSMetricsCallback = None,
|
| 38 |
) -> None:
|
| 39 |
"""Initialize Pocket TTS plugin.
|
|
@@ -43,9 +43,10 @@ class PocketTTS(tts.TTS):
|
|
| 43 |
or path to audio file for custom voice
|
| 44 |
temperature: Sampling temperature (0.0-2.0)
|
| 45 |
lsd_decode_steps: LSD decoding steps (higher = better quality, slower)
|
|
|
|
| 46 |
"""
|
| 47 |
# Use the configured output sample rate (default 48000 Hz)
|
| 48 |
-
self._output_sample_rate =
|
| 49 |
self._native_sample_rate = 24000 # Pocket TTS native rate
|
| 50 |
|
| 51 |
super().__init__(
|
|
@@ -149,7 +150,7 @@ class PocketSynthesizeStream(tts.SynthesizeStream):
|
|
| 149 |
Args:
|
| 150 |
output_emitter: Audio emitter for pushing generated audio
|
| 151 |
"""
|
| 152 |
-
request_id =
|
| 153 |
|
| 154 |
output_emitter.initialize(
|
| 155 |
request_id=request_id,
|
|
@@ -165,7 +166,7 @@ class PocketSynthesizeStream(tts.SynthesizeStream):
|
|
| 165 |
if isinstance(data, self._FlushSentinel):
|
| 166 |
if text_buffer.strip():
|
| 167 |
# Create a new segment for each text chunk
|
| 168 |
-
segment_id =
|
| 169 |
output_emitter.start_segment(segment_id=segment_id)
|
| 170 |
await self._synthesize_segment(text_buffer, output_emitter, segment_id)
|
| 171 |
output_emitter.end_segment()
|
|
@@ -177,7 +178,7 @@ class PocketSynthesizeStream(tts.SynthesizeStream):
|
|
| 177 |
|
| 178 |
# Process any remaining text
|
| 179 |
if text_buffer.strip():
|
| 180 |
-
segment_id =
|
| 181 |
output_emitter.start_segment(segment_id=segment_id)
|
| 182 |
await self._synthesize_segment(text_buffer, output_emitter, segment_id)
|
| 183 |
output_emitter.end_segment()
|
|
|
|
| 4 |
import asyncio
|
| 5 |
import logging
|
| 6 |
import time
|
|
|
|
| 7 |
from typing import Callable
|
| 8 |
|
| 9 |
import numpy as np
|
|
|
|
| 13 |
|
| 14 |
from livekit.agents import tts
|
| 15 |
from livekit.agents.types import APIConnectOptions, DEFAULT_API_CONNECT_OPTIONS
|
| 16 |
+
from livekit.agents.utils import shortuuid
|
| 17 |
|
| 18 |
+
logger = logging.getLogger(__name__)
|
|
|
|
| 19 |
|
| 20 |
# Reduce verbosity of pocket_tts library to avoid console spam
|
| 21 |
logging.getLogger("pocket_tts").setLevel(logging.WARNING)
|
|
|
|
| 33 |
voice: str = "alba",
|
| 34 |
temperature: float = 0.7,
|
| 35 |
lsd_decode_steps: int = 1,
|
| 36 |
+
sample_rate: int = 48000,
|
| 37 |
metrics_callback: OptionalTTSMetricsCallback = None,
|
| 38 |
) -> None:
|
| 39 |
"""Initialize Pocket TTS plugin.
|
|
|
|
| 43 |
or path to audio file for custom voice
|
| 44 |
temperature: Sampling temperature (0.0-2.0)
|
| 45 |
lsd_decode_steps: LSD decoding steps (higher = better quality, slower)
|
| 46 |
+
sample_rate: Output sample rate in Hz (default 48000)
|
| 47 |
"""
|
| 48 |
# Use the configured output sample rate (default 48000 Hz)
|
| 49 |
+
self._output_sample_rate = sample_rate
|
| 50 |
self._native_sample_rate = 24000 # Pocket TTS native rate
|
| 51 |
|
| 52 |
super().__init__(
|
|
|
|
| 150 |
Args:
|
| 151 |
output_emitter: Audio emitter for pushing generated audio
|
| 152 |
"""
|
| 153 |
+
request_id = shortuuid("TTS_")
|
| 154 |
|
| 155 |
output_emitter.initialize(
|
| 156 |
request_id=request_id,
|
|
|
|
| 166 |
if isinstance(data, self._FlushSentinel):
|
| 167 |
if text_buffer.strip():
|
| 168 |
# Create a new segment for each text chunk
|
| 169 |
+
segment_id = shortuuid("SEG_")
|
| 170 |
output_emitter.start_segment(segment_id=segment_id)
|
| 171 |
await self._synthesize_segment(text_buffer, output_emitter, segment_id)
|
| 172 |
output_emitter.end_segment()
|
|
|
|
| 178 |
|
| 179 |
# Process any remaining text
|
| 180 |
if text_buffer.strip():
|
| 181 |
+
segment_id = shortuuid("SEG_")
|
| 182 |
output_emitter.start_segment(segment_id=segment_id)
|
| 183 |
await self._synthesize_segment(text_buffer, output_emitter, segment_id)
|
| 184 |
output_emitter.end_segment()
|