app.py

import os
import gradio as gr
import requests
import json
from moviepy import VideoFileClip
import uuid
import time

ELEVENLABS_API_KEY = os.environ.get("ELEVENLABS_API_KEY", None)

def extract_audio(video_path, output_format="mp3"):
    if not video_path:
        return None, "No video provided"
    
    output_path = f"extracted_audio_{uuid.uuid4().hex[:8]}.{output_format}"
    
    try:
        video = VideoFileClip(video_path)
        video.audio.write_audiofile(output_path, logger=None)
        video.close()
        return output_path, f"Audio extracted successfully"
    except Exception as e:
        return None, f"Error extracting audio: {str(e)}"

def save_transcription(transcription):
    if "error" in transcription:
        return None, transcription["error"]
    transcript_filename = f"transcription_{uuid.uuid4().hex[:8]}.txt"
    
    try:
        with open(transcript_filename, "w", encoding="utf-8") as f:
            f.write(transcription.get('text', 'No text found'))
        
        return transcript_filename, "Transcription saved as text file"
    except Exception as e:
        return None, f"Error saving transcription: {str(e)}"

def process_video_file(video_file, output_format, api_key, model_id):
    if video_file is None:
        return None, "Please upload a video file", None, "No video provided"
    
    audio_path, message = extract_audio(video_file, output_format)
    
    if audio_path and os.path.exists(audio_path):
        transcription = transcribe_audio(audio_path, api_key, model_id)
        transcript_file, transcript_message = save_transcription(transcription)
        return audio_path, message, transcript_file, transcript_message
    else:
        return None, message, None, "Audio extraction failed, cannot transcribe"

def process_video_url(video_url, output_format, api_key, model_id):
    if not video_url.strip():
        return None, "Please enter a video URL", None, "No URL provided"
    
    video_path, error = download_video_from_url(video_url)
    if error:
        return None, error, None, "Video download failed, cannot transcribe"
    
    audio_path, message = extract_audio(video_path, output_format) 
    if video_path and os.path.exists(video_path):
        try:
            os.remove(video_path)
        except:
            pass
    
    if audio_path and os.path.exists(audio_path):
        transcription = transcribe_audio(audio_path, api_key, model_id)
        transcript_file, transcript_message = save_transcription(transcription)
        return audio_path, message, transcript_file, transcript_message
    else:
        return None, message, None, "Audio extraction failed, cannot transcribe"


def transcribe_audio(audio_path, api_key, model_id="elevenlabs_1"):

    start_time = time.time()
    
    if not api_key:
        return {"error": "Please provide an API key"}
    
    url = "https://api.elevenlabs.io/v1/speech-to-text"
    headers = {
        "xi-api-key": api_key,
        "Accept": "application/json"
    }
    
    try:
        with open(audio_path, "rb") as f:
            files = {
                "file": (os.path.basename(audio_path), f, "audio/mpeg"),
                "model_id": (None, model_id)
            }
            
            # Use requests.post with explicit content type handling
            response = requests.post(
                url, 
                headers=headers, 
                files=files
            )
            
            # Check for specific HTTP errors
            if response.status_code == 401:
                return {"error": "Unauthorized. Please check your API key."}
            if response.status_code == 422:
                return {"error": "Unprocessable Entity. Check file format or API usage."}
            
            response.raise_for_status()
            result = response.json()
    except requests.exceptions.RequestException as e:
        return {"error": f"API request failed: {str(e)}"}
    except json.JSONDecodeError:
        return {"error": "Failed to parse API response"}
    
    end_time = time.time()
    processing_time = end_time - start_time
    
    file_size = os.path.getsize(audio_path) / (1024 * 1024)
    
    try:
        audio_data, sample_rate = sf.read(audio_path)
        audio_duration = len(audio_data) / sample_rate
    except:
        try:
            import librosa
            audio_duration = librosa.get_duration(filename=audio_path)
        except:
            audio_duration = 0
    
    # Ensure text is extracted correctly from the new response format
    text = result.get('text', '')
    
    return {
        "service": "ElevenLabs",
        "text": text,
        "processing_time": processing_time,
        "file_size_mb": file_size,
        "audio_duration": audio_duration,
        "real_time_factor": processing_time / audio_duration if audio_duration > 0 else None,
        "processing_speed": audio_duration / processing_time if audio_duration > 0 else None,
        "raw_response": result,
        "language_code": result.get('language_code'),
        "language_probability": result.get('language_probability')
    }

with gr.Blocks(title="Video to Audio to Transcription") as app:
    gr.Markdown("# Video => Audio => Transcription")
    
    api_key = gr.Textbox(
        placeholder="Enter your ElevenLabs API key",
        label="ElevenLabs API Key",
        type="password",
        value=ELEVENLABS_API_KEY
    )
    
    model_id = gr.Dropdown(
        choices=["scribe_v1"],
        value="scribe_v1",
        label="Transcription Model"
    )
    
    with gr.Tabs():
        with gr.TabItem("Upload Video"):
            with gr.Row():
                with gr.Column():
                    video_input = gr.Video(label="Upload Video")
                    format_choice_file = gr.Radio(["mp3", "wav"], value="mp3", label="Output Format")
                    extract_button_file = gr.Button("Extract Audio & Transcribe")
                
                with gr.Column():
                    audio_output_file = gr.Audio(label="Extracted Audio", type="filepath")
                    status_output_file = gr.Textbox(label="Audio Extraction Status")
                    transcript_file_output = gr.File(label="Transcription Text File")
                    transcript_status_output = gr.Textbox(label="Transcription Status")
            
            extract_button_file.click(
                fn=process_video_file,
                inputs=[video_input, format_choice_file, api_key, model_id],
                outputs=[audio_output_file, status_output_file, transcript_file_output, transcript_status_output]
            )

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