Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,49 +1,70 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
import gradio as gr
|
| 4 |
import tempfile
|
| 5 |
-
import
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
language_dict = tts_order_voice
|
| 8 |
|
|
|
|
|
|
|
|
|
|
| 9 |
async def text_to_speech_edge(text, language_code):
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
tmp_path = tmp_file.name
|
| 14 |
|
| 15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
-
|
| 18 |
-
return "Speech synthesis complete: {}".format(text), tmp_path
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
| 21 |
-
|
| 22 |
-
#
|
| 23 |
input_text = gr.Textbox(lines=5, label="Input Text")
|
| 24 |
-
# Output Component: Textbox label translated
|
| 25 |
output_text = gr.Textbox(label="Output Text")
|
| 26 |
-
# Output Component: Audio label translated
|
| 27 |
output_audio = gr.Audio(type="filepath", label="Generated Audio File")
|
| 28 |
-
|
| 29 |
default_language = list(language_dict.keys())[0]
|
| 30 |
-
|
| 31 |
-
# Input Component: Dropdown label translated
|
| 32 |
language = gr.Dropdown(
|
| 33 |
choices=list(language_dict.keys()),
|
| 34 |
value=default_language,
|
| 35 |
label="Language / Voice"
|
| 36 |
)
|
| 37 |
|
| 38 |
-
|
|
|
|
| 39 |
interface = gr.Interface(
|
| 40 |
fn=text_to_speech_edge,
|
| 41 |
inputs=[input_text, language],
|
| 42 |
outputs=[output_text, output_audio],
|
| 43 |
-
title="Edge TTS Text-to-Speech",
|
| 44 |
-
description="Convert text into high-quality audio using various Edge TTS voices."
|
| 45 |
)
|
| 46 |
|
| 47 |
|
|
|
|
| 48 |
if __name__ == "__main__":
|
| 49 |
interface.launch()
|
|
|
|
| 1 |
+
# WARNING: This file assumes you have a tts_voice.py file with English-translated keys.
|
| 2 |
+
|
| 3 |
import gradio as gr
|
| 4 |
import tempfile
|
| 5 |
+
import edge_tts
|
| 6 |
+
import anyio
|
| 7 |
+
# Import the translated voice mapping (assuming you kept it)
|
| 8 |
+
from tts_voice import tts_order_voice
|
| 9 |
|
| 10 |
language_dict = tts_order_voice
|
| 11 |
|
| 12 |
+
# HARDCODED ENDPOINT: This is the fix to bypass connection instability
|
| 13 |
+
VOICE_SERVICE_ENDPOINT = "wss://speech.platform.bing.com/consumer/speech/synthesize/readaloud/v1"
|
| 14 |
+
|
| 15 |
async def text_to_speech_edge(text, language_code):
|
| 16 |
+
# Input Validation (kept from previous successful step)
|
| 17 |
+
if not text or not text.strip():
|
| 18 |
+
return "ERROR: Input text cannot be empty. Please enter text.", None
|
|
|
|
| 19 |
|
| 20 |
+
try:
|
| 21 |
+
voice = language_dict[language_code]
|
| 22 |
+
|
| 23 |
+
# Initialize Communicate with the specific endpoint
|
| 24 |
+
communicate = edge_tts.Communicate(
|
| 25 |
+
text,
|
| 26 |
+
voice,
|
| 27 |
+
# FORCED FIX: Use a specific, known WebSocket endpoint
|
| 28 |
+
websocket_url=VOICE_SERVICE_ENDPOINT
|
| 29 |
+
)
|
| 30 |
+
|
| 31 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
|
| 32 |
+
tmp_path = tmp_file.name
|
| 33 |
|
| 34 |
+
await communicate.save(tmp_path)
|
|
|
|
| 35 |
|
| 36 |
+
return "Speech synthesis complete: {}".format(text), tmp_path
|
| 37 |
+
|
| 38 |
+
except edge_tts.exceptions.NoAudioReceived:
|
| 39 |
+
# Catch the specific error with a clear message
|
| 40 |
+
return "ERROR: Failed to generate audio. The TTS service is rejecting the request. Check your text length or try a different voice.", None
|
| 41 |
+
except Exception as e:
|
| 42 |
+
# Catch any other unexpected errors
|
| 43 |
+
return f"An unexpected error occurred: {str(e)}", None
|
| 44 |
|
| 45 |
+
|
| 46 |
+
# --- Modern Gradio Component Syntax ---
|
| 47 |
input_text = gr.Textbox(lines=5, label="Input Text")
|
|
|
|
| 48 |
output_text = gr.Textbox(label="Output Text")
|
|
|
|
| 49 |
output_audio = gr.Audio(type="filepath", label="Generated Audio File")
|
|
|
|
| 50 |
default_language = list(language_dict.keys())[0]
|
|
|
|
|
|
|
| 51 |
language = gr.Dropdown(
|
| 52 |
choices=list(language_dict.keys()),
|
| 53 |
value=default_language,
|
| 54 |
label="Language / Voice"
|
| 55 |
)
|
| 56 |
|
| 57 |
+
|
| 58 |
+
# --- Gradio Interface Definition ---
|
| 59 |
interface = gr.Interface(
|
| 60 |
fn=text_to_speech_edge,
|
| 61 |
inputs=[input_text, language],
|
| 62 |
outputs=[output_text, output_audio],
|
| 63 |
+
title="Edge TTS Text-to-Speech",
|
| 64 |
+
description="Convert text into high-quality audio using various Edge TTS voices."
|
| 65 |
)
|
| 66 |
|
| 67 |
|
| 68 |
+
# --- Standard Launch Command ---
|
| 69 |
if __name__ == "__main__":
|
| 70 |
interface.launch()
|