Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| import edge_tts | |
| import asyncio | |
| import tempfile | |
| import os | |
| import requests | |
| from bs4 import BeautifulSoup | |
| # Global voice cache | |
| _voice_cache = None | |
| async def get_cached_voices(): | |
| global _voice_cache | |
| if _voice_cache is None: | |
| _voice_cache = await edge_tts.list_voices() | |
| return _voice_cache | |
| async def get_voices(): | |
| voices = await get_cached_voices() | |
| return {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices} | |
| def fetch_html_text_from_url(url): | |
| try: | |
| response = requests.get(url, timeout=10) | |
| response.raise_for_status() | |
| soup = BeautifulSoup(response.text, "html.parser") | |
| # Extract meaningful readable text from HTML (e.g., from <p> tags) | |
| paragraphs = soup.find_all('p') | |
| text = "\n\n".join(p.get_text(strip=True) for p in paragraphs if p.get_text(strip=True)) | |
| if not text.strip(): | |
| return "No readable content found on the page." | |
| return text | |
| except Exception as e: | |
| return f"Error fetching or parsing URL: {str(e)}" | |
| async def text_to_speech(text, voice, rate, pitch): | |
| if not text.strip(): | |
| return None, None, "Please enter text to convert." | |
| if not voice: | |
| return None, None, "Please select a voice." | |
| voice_short_name = voice.split(" - ")[0] | |
| rate_str = f"{rate:+d}%" | |
| pitch_str = f"{pitch:+d}Hz" | |
| communicate = edge_tts.Communicate(text, voice_short_name, rate=rate_str, pitch=pitch_str) | |
| with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file: | |
| tmp_path = tmp_file.name | |
| try: | |
| await communicate.save(tmp_path) | |
| except Exception as e: | |
| return None, None, f"Error generating speech: {str(e)}" | |
| return tmp_path, tmp_path, None | |
| async def tts_interface(text, voice, rate, pitch): | |
| audio_path, file_path, warning = await text_to_speech(text, voice, rate, pitch) | |
| if warning: | |
| return None, None, gr.update(visible=True, value=warning) | |
| return audio_path, file_path, gr.update(visible=False) | |
| async def create_demo(): | |
| voices = await get_voices() | |
| with gr.Blocks(analytics_enabled=False) as demo: | |
| gr.Markdown("# 🎙️ Edge TTS Text-to-Speech") | |
| with gr.Row(): | |
| with gr.Column(): | |
| url_input = gr.Textbox(label="Optional: Enter URL to extract text (e.g., arXiv)", placeholder="https://arxiv.org/html/...") | |
| fetch_button = gr.Button("Fetch Text from URL") | |
| text_input = gr.Textbox(label="Input Text", lines=10) | |
| fetch_button.click(fn=fetch_html_text_from_url, inputs=url_input, outputs=text_input) | |
| example_btn = gr.Button("Use Example Text") | |
| example_btn.click(fn=lambda: "Hello, this is a sample sentence using Edge TTS.", outputs=text_input) | |
| voice_dropdown = gr.Dropdown( | |
| choices=[""] + list(voices.keys()), | |
| label="Select Voice", | |
| value="en-US-GuyNeural - en-US (Male)" | |
| ) | |
| rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate Adjustment (%)", step=1) | |
| pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1) | |
| generate_btn = gr.Button("Generate Speech", variant="primary") | |
| audio_output = gr.Audio(label="Generated Audio", type="filepath", autoplay=True) | |
| file_download = gr.File(label="Download MP3", visible=False) | |
| warning_md = gr.Markdown("Warning Message", visible=False) | |
| generate_btn.click( | |
| fn=tts_interface, | |
| inputs=[text_input, voice_dropdown, rate_slider, pitch_slider], | |
| outputs=[audio_output, file_download, warning_md] | |
| ) | |
| return demo | |
| async def main(): | |
| demo = await create_demo() | |
| demo.queue() | |
| demo.launch(show_api=False) | |
| if __name__ == "__main__": | |
| asyncio.run(main()) | |