app.py

import base64
import io
import os
import threading
import tempfile
import logging
import openai
from dash import Dash, dcc, html, Input, Output, State, callback, callback_context
import dash_bootstrap_components as dbc
from pydub import AudioSegment
import requests
from pytube import YouTube

# Configure logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)

# Try to import moviepy with the simpler import statement
try:
    from moviepy import VideoFileClip
    logger.info("MoviePy (VideoFileClip) successfully imported")
except ImportError as e:
    logger.error(f"Error importing MoviePy (VideoFileClip): {str(e)}")
    logger.error("Please ensure moviepy is installed correctly")
    raise

# Supported file formats
AUDIO_FORMATS = ['.wav', '.mp3', '.ogg', '.flac', '.aac', '.m4a', '.wma']
VIDEO_FORMATS = ['.mp4', '.avi', '.mov', '.flv', '.wmv', '.mkv', '.webm']
SUPPORTED_FORMATS = AUDIO_FORMATS + VIDEO_FORMATS

# Initialize the Dash app
app = Dash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])

# Global variables
generated_file = None
transcription_text = ""

# Set up OpenAI API key
openai.api_key = os.getenv("OPENAI_API_KEY")

# Layout
app.layout = dbc.Container([
    html.H1("Audio/Video Transcription and Diarization App", className="text-center my-4"),
    dbc.Card([
        dbc.CardBody([
            dcc.Upload(
                id='upload-media',
                children=html.Div([
                    'Drag and Drop or ',
                    html.A('Select Audio/Video File')
                ]),
                style={
                    'width': '100%',
                    'height': '60px',
                    'lineHeight': '60px',
                    'borderWidth': '1px',
                    'borderStyle': 'dashed',
                    'borderRadius': '5px',
                    'textAlign': 'center',
                    'margin': '10px'
                },
                multiple=False
            ),
            html.Div(id='output-media-upload'),
            dbc.Input(id="url-input", type="text", placeholder="Enter audio/video URL (including YouTube)", className="mb-3"),
            dbc.Button("Process Media", id="process-url-button", color="primary", className="mb-3"),
            dbc.Spinner(html.Div(id='transcription-status'), color="primary", type="grow"),
            html.H4("Diarized Transcription Preview", className="mt-4"),
            html.Div(id='transcription-preview', style={'whiteSpace': 'pre-wrap'}),
            html.Br(),
            dbc.Button("Download Transcription", id="btn-download", color="primary", className="mt-3", disabled=True),
            dcc.Download(id="download-transcription")
        ])
    ])
], fluid=True)

def process_media(file_path, is_url=False):
    global generated_file, transcription_text
    temp_file = None
    wav_path = None
    try:
        if is_url:
            if 'youtube.com' in file_path or 'youtu.be' in file_path:
                yt = YouTube(file_path)
                stream = yt.streams.filter(progressive=True, file_extension='mp4').order_by('resolution').desc().first()
                temp_file = tempfile.NamedTemporaryFile(delete=False, suffix='.mp4')
                stream.download(output_path=os.path.dirname(temp_file.name), filename=os.path.basename(temp_file.name))
            else:
                response = requests.get(file_path)
                content_type = response.headers.get('content-type', '')
                if 'audio' in content_type:
                    suffix = '.mp3'
                elif 'video' in content_type:
                    suffix = '.mp4'
                else:
                    suffix = ''
                temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=suffix)
                temp_file.write(response.content)
                temp_file.close()
        else:
            temp_file = tempfile.NamedTemporaryFile(delete=False)
            temp_file.write(file_path)
            temp_file.close()

        file_extension = os.path.splitext(temp_file.name)[1].lower()
        
        if file_extension in VIDEO_FORMATS:
            video = VideoFileClip(temp_file.name)
            audio = video.audio
            wav_path = temp_file.name + ".wav"
            audio.write_audiofile(wav_path)
            video.close()
        elif file_extension in AUDIO_FORMATS:
            audio = AudioSegment.from_file(temp_file.name)
            wav_path = temp_file.name + ".wav"
            audio.export(wav_path, format="wav")
        else:
            return f"Unsupported file format: {file_extension}. Please upload a supported audio or video file.", False

        with open(wav_path, "rb") as audio_file:
            transcript = openai.Audio.transcribe("whisper-1", audio_file)
            audio_file.seek(0)
            diarized_transcript = openai.Audio.transcribe("whisper-1", audio_file, response_format="verbose_json")
        
        formatted_transcript = ""
        if 'segments' in diarized_transcript:
            for segment in diarized_transcript["segments"]:
                speaker = segment.get('speaker', 'Unknown')
                text = segment.get('text', '')
                formatted_transcript += f"Speaker {speaker}: {text}\n\n"
        else:
            formatted_transcript = transcript.get('text', 'No transcription available.')
        
        transcription_text = formatted_transcript
        generated_file = io.BytesIO(transcription_text.encode())
        return "Transcription and diarization completed successfully!", True
    except Exception as e:
        logger.error(f"Error during processing: {str(e)}")
        return f"An error occurred: {str(e)}", False
    finally:
        if temp_file and os.path.exists(temp_file.name):
            os.unlink(temp_file.name)
        if wav_path and os.path.exists(wav_path):
            os.unlink(wav_path)
            
@app.callback(
    [Output('output-media-upload', 'children'),
     Output('transcription-status', 'children'),
     Output('transcription-preview', 'children'),
     Output('btn-download', 'disabled')],
    [Input('upload-media', 'contents'),
     Input('process-url-button', 'n_clicks')],
    [State('upload-media', 'filename'),
     State('url-input', 'value')]
)
def update_output(contents, n_clicks, filename, url):
    ctx = callback_context
    if not ctx.triggered:
        return "No file uploaded or URL processed.", "", "", True

    trigger_id = ctx.triggered[0]['prop_id'].split('.')[0]

    if contents is not None:
        # Process file upload
        content_type, content_string = contents.split(',')
        decoded = base64.b64decode(content_string)
        status_message, success = process_media(decoded)
    elif url:
        # Process URL
        status_message, success = process_media(url, is_url=True)
    else:
        return "No file uploaded or URL processed.", "", "", True

    if success:
        preview = transcription_text[:1000] + "..." if len(transcription_text) > 1000 else transcription_text
        return f"Media processed successfully.", status_message, preview, False
    else:
        return "Processing failed.", status_message, "", True

@app.callback(
    Output("download-transcription", "data"),
    Input("btn-download", "n_clicks"),
    prevent_initial_call=True,
)
def download_transcription(n_clicks):
    if n_clicks is None:
        return None
    return dcc.send_bytes(generated_file.getvalue(), "diarized_transcription.txt")

if __name__ == '__main__':
    print("Starting the Dash application...")
    app.run(debug=True, host='0.0.0.0', port=7860)
    print("Dash application has finished running.")