app.py

import gradio as gr
import requests
import time
import os
from dotenv import load_dotenv

# Load AssemblyAI API key from .env
load_dotenv()
ASSEMBLYAI_API_KEY = os.getenv("ASSEMBLYAI_API_KEY")
if not ASSEMBLYAI_API_KEY:
    raise ValueError("Missing ASSEMBLYAI_API_KEY in environment variables.")


def transcribe_with_assemblyai_file(file_path, api_key):
    headers = {"authorization": api_key}
    upload_url = "https://api.assemblyai.com/v2/upload"

    def read_file(filename, chunk_size=5242880):
        with open(filename, 'rb') as f:
            while True:
                data = f.read(chunk_size)
                if not data:
                    break
                yield data

    print("Uploading file to AssemblyAI...")
    upload_response = requests.post(upload_url, headers=headers, data=read_file(file_path))
    audio_url = upload_response.json().get("upload_url")
    if not audio_url:
        return "Error during file upload."

    transcript_endpoint = "https://api.assemblyai.com/v2/transcript"
    transcript_request = {"audio_url": audio_url}
    transcript_response = requests.post(transcript_endpoint, json=transcript_request, headers=headers)
    transcript_id = transcript_response.json().get("id")
    if not transcript_id:
        return "Error initiating transcription."

    polling_url = f"{transcript_endpoint}/{transcript_id}"
    print("Transcription in progress...")
    while True:
        polling_response = requests.get(polling_url, headers=headers).json()
        status = polling_response.get("status")
        if status == "completed":
            return polling_response.get("text", "No transcription text available.")
        elif status == "error":
            return f"Error transcribing audio: {polling_response.get('error')}"
        time.sleep(5)


def transcribe(audio_file_path):
    # If user provides no audio, fallback to CommodityReport.mp3
    file_path = audio_file_path if audio_file_path else "CommodityReport.mp3"
    return transcribe_with_assemblyai_file(file_path, ASSEMBLYAI_API_KEY)


# Create a simplified UI without the upload button
with gr.Blocks() as demo:
    gr.Markdown("# Transcription")

    # Instructions with HTML for better styling
    gr.HTML("""
    <div class="instructions-container">

        <h4 class="section-header">USING AN AUDIO FILE:</h4>
        <ul>
            <li>Upload an audio file by dropping it or uploading it in the panel below. (If you don't have an audio file, just click 'Transcribe' and it will transcribe a default file.)</li>
            <li>Click Transcribe</li>
        </ul>

        <h4 class="section-header">RECORD YOUR OWN AUDIO AND TRANSCRIBE IT:</h4>
        <ul>
            <li>Click the grey microphone icon in the middle above the 'Transcribe' button.</li>
            <li>Then click the 'Record' button and click the 'Stop' button when finished.</li>
            <li>Click Transcribe</li>
        </ul>
    </div>
    """)

    # Use the original Audio component but make it smaller
    audio_input = gr.Audio(
        label="",  # Remove the label to reduce spacing
        type="filepath",
        interactive=True
    )

    # Add styling to make the audio box smaller and improve visuals
    gr.HTML("""
    <style>
    /* Style the instructions container */
    .instructions-container {
        background-color: #f8f9fa;
        padding: 5px;
        border-radius: 5px;
        margin-bottom: 5px;
        border-left: 4px solid #FF6B00;
    }

    /* Make the section headers larger */
    .section-header {
        font-size: 14px;
        font-weight: bold;
        margin-top: 5px;
        margin-bottom: 5px;
        color: #333;
    }

    /* Style list items */
    .instructions-container ul {
        margin-bottom: 5px;
    }

    .instructions-container li {
        margin-bottom: 5px;
    }

    /* Highlight the microphone icon */
    .audio-player-footer button:last-child svg {
        color: #888 !important;
        transform: scale(1.2);
    }

    /* Make the audio box smaller */
    .gradio-audio {
        max-height: 50px;
        overflow: hidden;
    }

    /* Ensure the audio controls remain visible */
    .audio-player-footer {
        position: relative;
        z-index: 10;
    }

    /* Make the Transcribe button more prominent */
    .gradio-button.primary {
        font-size: 16px;
        padding: 10px 20px;
        margin: 5px 0;
        background-color: #FF6B00;
    }
    </style>
    """)

    # Transcription controls - make button more prominent
    transcribe_button = gr.Button("Transcribe", variant="primary", size="lg")
    transcript_output = gr.Textbox(label="Transcript", lines=10)

    # Connect the transcribe button
    transcribe_button.click(fn=transcribe, inputs=audio_input, outputs=transcript_output)

demo.launch()