app.py

import gradio as gr
import edge_tts
import asyncio
import tempfile
import os
from moviepy.editor import AudioFileClip

# Get all available voices
async def get_voices():
    voices = await edge_tts.list_voices()
    return {f"{v['ShortName']} - {v['Locale']} ({v['Gender']})": v['ShortName'] for v in voices}

# Text to speech functionality
async def text_to_speech(text, voice, rate, pitch):
    if not text.strip():
        return None, gr.Warning("Please enter the text to convert.")
    if not voice:
        return None, gr.Warning("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
        await communicate.save(tmp_path)
    return tmp_path, None

# Generate SRT file based on user preferences
def generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph):
    total_paragraphs = len(paragraphs)
    
    # Calculate how long each segment will be displayed
    segment_duration = audio_duration / total_paragraphs  # Total audio duration divided by total paragraphs
    
    current_time = 0
    with open(srt_path, 'w', encoding='utf-8') as srt_file:
        for i, paragraph in enumerate(paragraphs):
            words = paragraph.split()
            lines = [words[j:j + words_per_line] for j in range(0, len(words), words_per_line)]
            lines = [' '.join(line) for line in lines]

            start_time = current_time
            end_time = min(start_time + segment_duration, audio_duration)  # Ensure it doesn't exceed audio duration
            
            start_time_str = format_srt_time(start_time)
            end_time_str = format_srt_time(end_time)
            srt_file.write(f"{i + 1}\n{start_time_str} --> {end_time_str}\n")
            srt_file.write('\n'.join(lines) + '\n\n')

            current_time += segment_duration  # Update current time for the next segment

    return srt_path

def format_srt_time(seconds):
    millis = int((seconds - int(seconds)) * 1000)
    seconds = int(seconds)
    minutes = seconds // 60
    hours = seconds // 3600
    minutes %= 60
    seconds %= 60
    return f"{hours:02}:{minutes:02}:{seconds:02},{millis:03}"

# Text to audio and SRT functionality
async def text_to_audio_and_srt(text, voice, rate, pitch, words_per_line, lines_per_paragraph):
    audio_path, warning = await text_to_speech(text, voice, rate, pitch)
    if warning:
        return None, None, warning

    audio_clip = AudioFileClip(audio_path)
    audio_duration = audio_clip.duration
    
    # Generate SRT file based on the entire text
    base_name = os.path.splitext(audio_path)[0]
    srt_path = f"{base_name}_subtitle.srt"
    
    # Split input text into paragraphs based on larger gaps (two consecutive newlines)
    paragraphs = [p.strip() for p in text.split('\n\n') if p.strip()]
    
    generate_srt(paragraphs, audio_duration, srt_path, words_per_line, lines_per_paragraph)

    return audio_path, srt_path, None

# Gradio interface function
def tts_interface(text, voice, rate, pitch, words_per_line, lines_per_paragraph):
    audio_path, srt_path, warning = asyncio.run(text_to_audio_and_srt(text, voice, rate, pitch, words_per_line, lines_per_paragraph))
    return audio_path, srt_path, warning

# Create Gradio app
async def create_demo():
    voices = await get_voices()
    
    with gr.Blocks() as demo:
        gr.Markdown(
            """
            <h1 style="text-align: center; color: #333;">Text to Speech with Subtitles</h1>
            <p style="text-align: center; color: #555;">Convert your text to natural-sounding speech and generate subtitles (SRT) for your audio.</p>
            """, 
            elem_id="header"
        )

        with gr.Row():
            with gr.Column():
                text_input = gr.Textbox(label="Input Text", lines=5, placeholder="Enter text here...")
                voice_dropdown = gr.Dropdown(choices=[""] + list(voices.keys()), label="Select Voice", value="")
                rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Rate Adjustment (%)", step=1)
                pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1)

                words_per_line = gr.Slider(minimum=3, maximum=8, value=5, label="Words per Line", step=1)
                lines_per_paragraph = gr.Slider(minimum=1, maximum=5, value=2, label="Lines per Paragraph", step=1)
                
                generate_button = gr.Button("Generate Audio and Subtitles", variant="primary")

            with gr.Column():
                output_audio = gr.Audio(label="Generated Audio", type="filepath")
                output_srt = gr.File(label="Generated SRT", file_count="single")
                warning_msg = gr.Markdown(label="Warning", visible=False)

        generate_button.click(
            fn=tts_interface,
            inputs=[text_input, voice_dropdown, rate_slider, pitch_slider, words_per_line, lines_per_paragraph],
            outputs=[output_audio, output_srt, warning_msg]
        )

    return demo

# Run the app
if __name__ == "__main__":
    demo = asyncio.run(create_demo())
    demo.launch()