0plus1 / sottovoce / #4

Committed 10 Dec 2025 01:17PM UTC coverage: 72.956% (+0.4%) from 72.589%

Build # #4

Build Type

push

coveralls-python

Committed by

0plus1

Commit Message

Add speech synthesis

Run Details

89 of 133 new or added lines in 7 files covered. (66.92%)

6 existing lines in 2 files now uncovered.

232 of 318 relevant lines covered (72.96%)

0.73 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

28.57

/src/tts_engine.py

from __future__ import annotations

from pathlib import Path
from typing import Optional

import numpy
import sounddevice as sd  # type: ignore[import-not-found]
from piper import PiperVoice, SynthesisConfig  # type: ignore[import-not-found]

from src.config import Settings


class TtsEngine:
    """Thin wrapper around Piper TTS."""

    def __init__(self, settings: Settings):
        self.enabled = settings.tts_enabled
        self.voice_path = settings.tts_voice_path
        self.use_cuda = settings.tts_use_cuda
        self.length_scale = settings.tts_length_scale
        self.noise_scale = settings.tts_noise_scale
        self.noise_w_scale = settings.tts_noise_w_scale
        self.volume = settings.tts_volume
        self._voice: Optional[PiperVoice] = None

    def _ensure_voice(self) -> None:
        if self._voice is None:
            if not self.voice_path:
                raise RuntimeError("TTS voice path is not set. Set TTS_VOICE_PATH to a Piper voice .onnx.")
            self._voice = PiperVoice.load(self.voice_path, use_cuda=self.use_cuda)

    def _synth_config(self):
        assert self._voice is not None
        return SynthesisConfig(
            volume=self.volume,
            length_scale=self.length_scale,
            noise_scale=self.noise_scale,
            noise_w_scale=self.noise_w_scale,
        )

    def synthesize(self, text: str, output_path: Path) -> Path:
        if not self.enabled:
            raise RuntimeError("TTS is disabled")
        self._ensure_voice()
        assert self._voice is not None
        syn_config = self._synth_config()

        # Play back to the user (no disk write)
        chunks = list(self._voice.synthesize(text, syn_config=syn_config))
        if chunks:
            sample_rate = chunks[0].sample_rate
            channels = chunks[0].sample_channels
            audio_bytes = b"".join(chunk.audio_int16_bytes for chunk in chunks)
            audio = numpy.frombuffer(audio_bytes, dtype=numpy.int16)
            if channels > 1:
                audio = audio.reshape(-1, channels)
            sd.play(audio, sample_rate)
            sd.wait()

        return output_path

1	from __future__ import annotations	1✔
2
3	from pathlib import Path	1✔
4	from typing import Optional	1✔
5
6	import numpy	1✔
7	import sounddevice as sd # type: ignore[import-not-found]	1✔
8	from piper import PiperVoice, SynthesisConfig # type: ignore[import-not-found]	1✔
9
10	from src.config import Settings	1✔
11
12
13	class TtsEngine:	1✔
14	"""Thin wrapper around Piper TTS."""
15
16	def __init__(self, settings: Settings):	1✔
NEW 17	self.enabled = settings.tts_enabled	×
NEW 18	self.voice_path = settings.tts_voice_path	×
NEW 19	self.use_cuda = settings.tts_use_cuda	×
NEW 20	self.length_scale = settings.tts_length_scale	×
NEW 21	self.noise_scale = settings.tts_noise_scale	×
NEW 22	self.noise_w_scale = settings.tts_noise_w_scale	×
NEW 23	self.volume = settings.tts_volume	×
NEW 24	self._voice: Optional[PiperVoice] = None	×
25
26	def _ensure_voice(self) -> None:	1✔
NEW 27	if self._voice is None:	×
NEW 28	if not self.voice_path:	×
NEW 29	raise RuntimeError("TTS voice path is not set. Set TTS_VOICE_PATH to a Piper voice .onnx.")	×
NEW 30	self._voice = PiperVoice.load(self.voice_path, use_cuda=self.use_cuda)	×
31
32	def _synth_config(self):	1✔
NEW 33	assert self._voice is not None	×
NEW 34	return SynthesisConfig(	×
35	volume=self.volume,
36	length_scale=self.length_scale,
37	noise_scale=self.noise_scale,
38	noise_w_scale=self.noise_w_scale,
39	)
40
41	def synthesize(self, text: str, output_path: Path) -> Path:	1✔
NEW 42	if not self.enabled:	×
NEW 43	raise RuntimeError("TTS is disabled")	×
NEW 44	self._ensure_voice()	×
NEW 45	assert self._voice is not None	×
NEW 46	syn_config = self._synth_config()	×
47
48	# Play back to the user (no disk write)
NEW 49	chunks = list(self._voice.synthesize(text, syn_config=syn_config))	×
NEW 50	if chunks:	×
NEW 51	sample_rate = chunks[0].sample_rate	×
NEW 52	channels = chunks[0].sample_channels	×
NEW 53	audio_bytes = b"".join(chunk.audio_int16_bytes for chunk in chunks)	×
NEW 54	audio = numpy.frombuffer(audio_bytes, dtype=numpy.int16)	×
NEW 55	if channels > 1:	×
NEW 56	audio = audio.reshape(-1, channels)	×
NEW 57	sd.play(audio, sample_rate)	×
NEW 58	sd.wait()	×
59
NEW 60	return output_path	×

0plus1 / sottovoce / #4

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous