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
import mimetypes
import urllib.parse

# 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", 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 chunk_audio(audio_segment, chunk_size_ms=60000):
    chunks = []
    for i in range(0, len(audio_segment), chunk_size_ms):
        chunks.append(audio_segment[i:i+chunk_size_ms])
    return chunks

def transcribe_audio_chunks(chunks):
    transcriptions = []
    for chunk in chunks:
        with io.BytesIO() as audio_file:
            chunk.export(audio_file, format="wav")
            audio_file.seek(0)
            transcript = openai.Audio.transcribe("whisper-1", audio_file)
            transcriptions.append(transcript.get('text', ''))
    return ' '.join(transcriptions)

def process_media(file_path, is_url=False):
    global generated_file, transcription_text
    temp_file = None
    wav_path = None
    try:
        if is_url:
            logger.info(f"Processing URL: {file_path}")
            try:
                response = requests.get(file_path, stream=True)
                response.raise_for_status()
                content_type = response.headers.get('content-type', '')
                extension = mimetypes.guess_extension(content_type) or ''
                temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=extension)
                for chunk in response.iter_content(chunk_size=8192):
                    temp_file.write(chunk)
                temp_file.close()
                logger.info(f"URL content downloaded: {temp_file.name}")
            except Exception as e:
                logger.error(f"Error downloading URL content: {str(e)}")
                return f"Error downloading URL content: {str(e)}", False
        else:
            logger.info("Processing uploaded file")
            temp_file = tempfile.NamedTemporaryFile(delete=False)
            temp_file.write(file_path)
            temp_file.close()
            logger.info(f"Uploaded file saved: {temp_file.name}")

        # Try to extract audio using moviepy
        try:
            audio = AudioFileClip(temp_file.name)
            wav_path = temp_file.name + ".wav"
            audio.write_audiofile(wav_path)
            audio.close()
        except Exception as e:
            logger.warning(f"Could not process with moviepy: {str(e)}")
            # If moviepy fails, try with pydub
            try:
                audio = AudioSegment.from_file(temp_file.name)
                wav_path = temp_file.name + ".wav"
                audio.export(wav_path, format="wav")
            except Exception as e:
                logger.error(f"Could not process audio: {str(e)}")
                return f"Could not process audio: {str(e)}", False

        logger.info(f"Audio extracted to WAV: {wav_path}")

        # Chunk the audio file
        audio = AudioSegment.from_wav(wav_path)
        chunks = chunk_audio(audio)
        
        # Transcribe chunks
        transcription = transcribe_audio_chunks(chunks)
        
        # Diarization (simplified as OpenAI doesn't provide speaker diarization)
        formatted_transcript = f"Speaker 1: {transcription}"
        
        transcription_text = formatted_transcript
        generated_file = io.BytesIO(transcription_text.encode())
        logger.info("Transcription and diarization completed successfully")
        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

    # Clear the preview pane
    transcription_preview = ""

    if contents is not None:
        content_type, content_string = contents.split(',')
        decoded = base64.b64decode(content_string)
        status_message, success = process_media(decoded)
    elif 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, transcription_preview, 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.")