aboutsummaryrefslogtreecommitdiff
path: root/app/converter/clients/__init__.py
diff options
context:
space:
mode:
authorhistoria <historiavg@proton.me>2026-09-02 01:26:09 -0400
committerhistoria <historiavg@proton.me>2026-09-02 01:26:09 -0400
commit8579517a35ef1865fc9b428899d73d52dcb27a14 (patch)
treedba52f8d99cfe4014e0b787367de99f238e5a0db /app/converter/clients/__init__.py
parent391f50da7a085bec75155c0eb9b47910266058cc (diff)
downloadtts-audiobook-generator-8579517a35ef1865fc9b428899d73d52dcb27a14.tar.gz
feat: sglang backend support
Diffstat (limited to 'app/converter/clients/__init__.py')
-rw-r--r--app/converter/clients/__init__.py16
1 files changed, 10 insertions, 6 deletions
diff --git a/app/converter/clients/__init__.py b/app/converter/clients/__init__.py
index cdb7912..d2a7f8d 100644
--- a/app/converter/clients/__init__.py
+++ b/app/converter/clients/__init__.py
@@ -1,9 +1,9 @@
"""TTS client implementations — one module per backend server.
-Public API: the three client classes (QwenTTSClient, FasterTTSClient,
-AudioCppTTSClient), the backend/voice-mode vocabulary, and the shared
-helpers (normalize_language, speaker tables, whisper transcription) that
-the UIs and setup wizards build on.
+Public API: the client classes (QwenTTSClient, FasterTTSClient,
+AudioCppTTSClient, SgOmniTTSClient), the backend/voice-mode vocabulary,
+and the shared helpers (normalize_language, speaker tables, whisper
+transcription) that the UIs and setup wizards build on.
"""
# The TTS backends a conversion can use, in Convert-form order. Each has a
@@ -12,7 +12,8 @@ the UIs and setup wizards build on.
BACKEND_QWEN = "qwen"
BACKEND_FASTER = "faster"
BACKEND_AUDIOCPP = "audiocpp"
-BACKENDS = (BACKEND_AUDIOCPP, BACKEND_QWEN, BACKEND_FASTER)
+BACKEND_SGLOMNI = "sglomni"
+BACKENDS = (BACKEND_AUDIOCPP, BACKEND_QWEN, BACKEND_FASTER, BACKEND_SGLOMNI)
from .base import BaseTTSClient, ConversionCancelled, VOICE_MODE_CLONE, \
VOICE_MODE_CUSTOM, VOICE_MODE_DESIGN, VOICE_MODES, resolve_request_seed
@@ -25,6 +26,7 @@ from .transcribe import (transcribe_reference_audio,
whisper_backend_available, whisper_backend_problem)
from .qwen import CUSTOM_VOICE_MODEL_ID, MODEL_SIZE, QwenTTSClient
from .faster import SAMPLE_RATE, FasterTTSClient
+from .sglomni import SgOmniTTSClient
from .audiocpp import (
AUDIOCPP_CLONE_ONLY_FAMILIES,
AUDIOCPP_DEFAULT_FAMILY_PROFILE,
@@ -58,11 +60,13 @@ from .audiocpp import (
__all__ = [
# vocabulary
- "BACKEND_QWEN", "BACKEND_FASTER", "BACKEND_AUDIOCPP", "BACKENDS",
+ "BACKEND_QWEN", "BACKEND_FASTER", "BACKEND_AUDIOCPP", "BACKEND_SGLOMNI",
+ "BACKENDS",
"VOICE_MODE_CUSTOM", "VOICE_MODE_CLONE", "VOICE_MODE_DESIGN", "VOICE_MODES",
# clients
"BaseTTSClient", "ConversionCancelled", "resolve_request_seed",
"QwenTTSClient", "FasterTTSClient", "AudioCppTTSClient",
+ "SgOmniTTSClient",
# model facts
"MODEL_SIZE", "CUSTOM_VOICE_MODEL_ID", "SAMPLE_RATE",
# languages