Spaces:
Running
Running
zach
commited on
Commit
·
7f69677
1
Parent(s):
557e7ca
Update design to accomodate easily adding other TTS providers in the future
Browse files- src/app.py +16 -11
- src/constants.py +4 -0
src/app.py
CHANGED
|
@@ -28,6 +28,7 @@ from src.constants import (
|
|
| 28 |
PROMPT_MIN_LENGTH,
|
| 29 |
SAMPLE_PROMPTS,
|
| 30 |
TROPHY_EMOJI,
|
|
|
|
| 31 |
VOTE_FOR_OPTION_A,
|
| 32 |
VOTE_FOR_OPTION_B,
|
| 33 |
)
|
|
@@ -108,24 +109,28 @@ def text_to_speech(
|
|
| 108 |
logger.warning("Skipping text-to-speech due to empty text.")
|
| 109 |
raise gr.Error("Please generate or enter text to synthesize.")
|
| 110 |
|
|
|
|
|
|
|
| 111 |
# If not using generated text, then only compare Hume to Hume
|
| 112 |
-
|
| 113 |
-
random.
|
| 114 |
)
|
| 115 |
|
| 116 |
try:
|
| 117 |
with ThreadPoolExecutor(max_workers=2) as executor:
|
| 118 |
-
provider_a = HUME_AI
|
| 119 |
future_audio_a = executor.submit(text_to_speech_with_hume, prompt, text)
|
| 120 |
|
| 121 |
-
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
|
| 125 |
-
|
| 126 |
-
|
| 127 |
-
|
| 128 |
-
|
|
|
|
|
|
|
|
|
|
| 129 |
|
| 130 |
audio_a = future_audio_a.result()
|
| 131 |
audio_b = future_audio_b.result()
|
|
|
|
| 28 |
PROMPT_MIN_LENGTH,
|
| 29 |
SAMPLE_PROMPTS,
|
| 30 |
TROPHY_EMOJI,
|
| 31 |
+
TTS_PROVIDERS,
|
| 32 |
VOTE_FOR_OPTION_A,
|
| 33 |
VOTE_FOR_OPTION_B,
|
| 34 |
)
|
|
|
|
| 109 |
logger.warning("Skipping text-to-speech due to empty text.")
|
| 110 |
raise gr.Error("Please generate or enter text to synthesize.")
|
| 111 |
|
| 112 |
+
# Hume AI always included in comparison
|
| 113 |
+
provider_a = HUME_AI
|
| 114 |
# If not using generated text, then only compare Hume to Hume
|
| 115 |
+
provider_b = (
|
| 116 |
+
HUME_AI if text != generated_text_state else random.choice(TTS_PROVIDERS)
|
| 117 |
)
|
| 118 |
|
| 119 |
try:
|
| 120 |
with ThreadPoolExecutor(max_workers=2) as executor:
|
|
|
|
| 121 |
future_audio_a = executor.submit(text_to_speech_with_hume, prompt, text)
|
| 122 |
|
| 123 |
+
match provider_b:
|
| 124 |
+
case ELEVENLABS:
|
| 125 |
+
future_audio_b = executor.submit(
|
| 126 |
+
text_to_speech_with_elevenlabs, prompt, text
|
| 127 |
+
)
|
| 128 |
+
case HUME_AI:
|
| 129 |
+
future_audio_b = executor.submit(
|
| 130 |
+
text_to_speech_with_hume, prompt, text
|
| 131 |
+
)
|
| 132 |
+
case _:
|
| 133 |
+
raise ValueError(f"Unsupported provider: {provider_b}")
|
| 134 |
|
| 135 |
audio_a = future_audio_a.result()
|
| 136 |
audio_b = future_audio_b.result()
|
src/constants.py
CHANGED
|
@@ -4,6 +4,9 @@ constants.py
|
|
| 4 |
This module defines global constants used throughout the project.
|
| 5 |
"""
|
| 6 |
|
|
|
|
|
|
|
|
|
|
| 7 |
# Third-Party Library Imports
|
| 8 |
from src.types import OptionKey, TTSProviderName
|
| 9 |
|
|
@@ -11,6 +14,7 @@ from src.types import OptionKey, TTSProviderName
|
|
| 11 |
# UI constants
|
| 12 |
HUME_AI: TTSProviderName = "Hume AI"
|
| 13 |
ELEVENLABS: TTSProviderName = "ElevenLabs"
|
|
|
|
| 14 |
|
| 15 |
PROMPT_MIN_LENGTH: int = 20
|
| 16 |
PROMPT_MAX_LENGTH: int = 800
|
|
|
|
| 4 |
This module defines global constants used throughout the project.
|
| 5 |
"""
|
| 6 |
|
| 7 |
+
# Standard Library Imports
|
| 8 |
+
from typing import List
|
| 9 |
+
|
| 10 |
# Third-Party Library Imports
|
| 11 |
from src.types import OptionKey, TTSProviderName
|
| 12 |
|
|
|
|
| 14 |
# UI constants
|
| 15 |
HUME_AI: TTSProviderName = "Hume AI"
|
| 16 |
ELEVENLABS: TTSProviderName = "ElevenLabs"
|
| 17 |
+
TTS_PROVIDERS: List[TTSProviderName]
|
| 18 |
|
| 19 |
PROMPT_MIN_LENGTH: int = 20
|
| 20 |
PROMPT_MAX_LENGTH: int = 800
|