app.py

import os
import requests
import json
import time
import subprocess
import gradio as gr
import uuid
from dotenv import load_dotenv

# Load environment variables
load_dotenv()

# API Keys
A_KEY = os.getenv("A_KEY")  # ElevenLabs API key
B_KEY = os.getenv("B_KEY")  # Lipsync API key
OPENAI_KEY = os.getenv("OPENAI_KEY")  # OpenAI API key

# URLs
API_URL = os.getenv("API_URL")
UPLOAD_URL = os.getenv("UPLOAD_URL")
OPENAI_API_URL = "https://api.openai.com/v1/audio/speech"

def get_elevenlabs_voices():
    url = "https://api.elevenlabs.io/v1/voices"
    headers = {
        "Accept": "application/json",
        "xi-api-key": A_KEY
    }
    
    response = requests.get(url, headers=headers)
    if response.status_code != 200:
        return []
    return [(voice['name'], voice['voice_id']) for voice in response.json().get('voices', [])]

def get_openai_voices():
    # OpenAI voices are predefined
    return [
        ("alloy", "alloy"),
        ("echo", "echo"),
        ("fable", "fable"),
        ("onyx", "onyx"),
        ("nova", "nova"),
        ("shimmer", "shimmer")
    ]

def text_to_speech_elevenlabs(voice_id, text, session_id):
    url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
    
    headers = {
        "Accept": "audio/mpeg",
        "Content-Type": "application/json",
        "xi-api-key": A_KEY
    }
    
    data = {
        "text": text,
        "model_id": "eleven_turbo_v2_5",
        "voice_settings": {
            "stability": 0.5,
            "similarity_boost": 0.5
        }
    }
    
    response = requests.post(url, json=data, headers=headers)
    if response.status_code != 200:
        return None
    
    audio_file_path = f'temp_voice_{session_id}.mp3'
    with open(audio_file_path, 'wb') as audio_file:
        audio_file.write(response.content)
    return audio_file_path

def text_to_speech_openai(voice, text, session_id):
    headers = {
        "Authorization": f"Bearer {OPENAI_KEY}",
        "Content-Type": "application/json"
    }
    
    data = {
        "model": "tts-1",
        "input": text,
        "voice": voice
    }
    
    response = requests.post(OPENAI_API_URL, headers=headers, json=data)
    if response.status_code != 200:
        return None
    
    audio_file_path = f'temp_voice_{session_id}.mp3'
    with open(audio_file_path, 'wb') as audio_file:
        audio_file.write(response.content)
    return audio_file_path

def upload_file(file_path):
    with open(file_path, 'rb') as file:
        files = {'fileToUpload': (os.path.basename(file_path), file)}
        data = {'reqtype': 'fileupload'}
        response = requests.post(UPLOAD_URL, files=files, data=data)
    
    if response.status_code == 200:
        return response.text.strip()
    return None

def lipsync_api_call(video_url, audio_url):
    headers = {
        "Content-Type": "application/json",
        "x-api-key": B_KEY
    }
    
    data = {
        "audioUrl": audio_url,
        "videoUrl": video_url,
        "maxCredits": 1000,
        "model": "sync-1.7.1-beta",
        "synergize": True,
        "pads": [0, 5, 0, 0],
        "synergizerStrength": 1
    }
    
    response = requests.post(API_URL, headers=headers, data=json.dumps(data))
    return response.json()

def check_job_status(job_id):
    headers = {"x-api-key": B_KEY}
    max_attempts = 30
    
    for _ in range(max_attempts):
        response = requests.get(f"{API_URL}/{job_id}", headers=headers)
        data = response.json()
        
        if data["status"] == "COMPLETED":
            return data["videoUrl"]
        elif data["status"] == "FAILED":
            return None
        
        time.sleep(10)
    return None

def get_media_duration(file_path):
    cmd = ['ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of', 'default=noprint_wrappers=1:nokey=1', file_path]
    result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    return float(result.stdout.strip())

def combine_audio_video(video_path, audio_path, output_path):
    video_duration = get_media_duration(video_path)
    audio_duration = get_media_duration(audio_path)

    if video_duration > audio_duration:
        cmd = [
            'ffmpeg', '-i', video_path, '-i', audio_path,
            '-t', str(audio_duration),
            '-map', '0:v', '-map', '1:a',
            '-c:v', 'copy', '-c:a', 'aac',
            '-y', output_path
        ]
    else:
        loop_count = int(audio_duration // video_duration) + 1
        cmd = [
            'ffmpeg', '-stream_loop', str(loop_count), '-i', video_path, '-i', audio_path,
            '-t', str(audio_duration),
            '-map', '0:v', '-map', '1:a',
            '-c:v', 'copy', '-c:a', 'aac',
            '-shortest', '-y', output_path
        ]

    subprocess.run(cmd, check=True)

def process_video(provider, voice, video_url, text, progress=gr.Progress()):
    session_id = str(uuid.uuid4())
    progress(0, desc="Generating speech...")
    
    if provider == "ElevenLabs":
        audio_path = text_to_speech_elevenlabs(voice, text, session_id)
    else:  # OpenAI
        audio_path = text_to_speech_openai(voice, text, session_id)
    
    if not audio_path:
        return None, "Failed to generate speech audio."
    
    progress(0.2, desc="Processing video...")
    
    try:
        progress(0.3, desc="Uploading audio...")
        audio_url = upload_file(audio_path)
        
        if not audio_url:
            raise Exception("Failed to upload audio file")
        
        progress(0.4, desc="Initiating lipsync...")
        job_data = lipsync_api_call(video_url, audio_url)
        
        if "error" in job_data or "message" in job_data:
            raise Exception(job_data.get("error", job_data.get("message", "Unknown error")))
        
        job_id = job_data["id"]
        
        progress(0.5, desc="Processing lipsync...")
        result_url = check_job_status(job_id)
        
        if result_url:
            progress(0.9, desc="Downloading result...")
            response = requests.get(result_url)
            output_path = f"output_{session_id}.mp4"
            with open(output_path, "wb") as f:
                f.write(response.content)
            progress(1.0, desc="Complete!")
            return output_path, "Lipsync completed successfully!"
        else:
            raise Exception("Lipsync processing failed or timed out")
            
    except Exception as e:
        progress(0.8, desc="Falling back to simple combination...")
        try:
            video_response = requests.get(video_url)
            video_path = f"temp_video_{session_id}.mp4"
            with open(video_path, "wb") as f:
                f.write(video_response.content)
            
            output_path = f"output_{session_id}.mp4"
            combine_audio_video(video_path, audio_path, output_path)
            progress(1.0, desc="Complete!")
            return output_path, f"Used fallback method. Original error: {str(e)}"
        except Exception as fallback_error:
            return None, f"All methods failed. Error: {str(fallback_error)}"
    finally:
        if os.path.exists(audio_path):
            os.remove(audio_path)
        if os.path.exists(f"temp_video_{session_id}.mp4"):
            os.remove(f"temp_video_{session_id}.mp4")

def create_interface():
    elevenlabs_voices = get_elevenlabs_voices()
    openai_voices = get_openai_voices()
    
    with gr.Blocks() as app:
        gr.Markdown("# Voice Synthesis Application")
        with gr.Row():
            with gr.Column():
                provider_dropdown = gr.Dropdown(choices=["ElevenLabs", "OpenAI"], label="Select Provider", value="ElevenLabs")
                voice_dropdown = gr.Dropdown(choices=[v[0] for v in elevenlabs_voices], label="Select Voice", value=elevenlabs_voices[0][0] if elevenlabs_voices else None)
                video_url_input = gr.Textbox(label="Enter Video URL")
                text_input = gr.Textbox(label="Enter text", lines=3)
                generate_btn = gr.Button("Generate Video")
            with gr.Column():
                video_output = gr.Video(label="Generated Video")
                status_output = gr.Textbox(label="Status", interactive=False)
        
        def update_voices(provider):
            if provider == "ElevenLabs":
                return gr.Dropdown(choices=[v[0] for v in elevenlabs_voices], value=elevenlabs_voices[0][0] if elevenlabs_voices else None)
            else:  # OpenAI
                return gr.Dropdown(choices=[v[0] for v in openai_voices], value=openai_voices[0][0])
        
        provider_dropdown.change(fn=update_voices, inputs=[provider_dropdown], outputs=[voice_dropdown])
        
        def on_generate(provider, voice_name, video_url, text):
            if provider == "ElevenLabs":
                voice_id = next((v[1] for v in elevenlabs_voices if v[0] == voice_name), None)
            else:  # OpenAI
                voice_id = next((v[1] for v in openai_voices if v[0] == voice_name), None)
            
            if not voice_id:
                return None, "Invalid voice selected."
            return process_video(provider, voice_id, video_url, text)
        
        generate_btn.click(
            fn=on_generate,
            inputs=[provider_dropdown, voice_dropdown, video_url_input, text_input],
            outputs=[video_output, status_output]
        )
    
    return app

if __name__ == "__main__":
    app = create_interface()
    app.launch()