app.py

import ggwave
import gradio as gr
import numpy as np
from pydub import AudioSegment


def encode_text_to_audio(text, protocol_id=1, volume=20):
    """
    Encode text to audio waveform using ggwave

    Args:
        text: Text to encode
        protocol_id: ggwave protocol (0-8, default 1)
        volume: Audio volume (0-100, default 20)

    Returns:
        Tuple of (sample_rate, audio_data) for Gradio audio output
    """
    if not text:
        return None

    try:
        waveform = ggwave.encode(text, protocolId=protocol_id, volume=volume)
        audio_data = np.frombuffer(waveform, dtype=np.float32)
        return 48000, audio_data
    except Exception as e:
        raise gr.Error(f"Encoding failed: {str(e)}")


def decode_audio_to_text(filepath):
    """
    Decode audio waveform to text using ggwave

    Args:
        filepath: Path to the audio file from Gradio

    Returns:
        Decoded text string
    """
    if filepath is None:
        return "No audio provided"

    try:
        segment = AudioSegment.from_file(filepath).set_channels(1)
        sample_rate = segment.frame_rate
        audio = np.array(segment.get_array_of_samples(), dtype=np.float32)
        audio /= 2 ** (segment.sample_width * 8 - 1)
        if sample_rate != 48000:
            duration = len(audio) / sample_rate
            new_length = int(duration * 48000)
            audio = np.interp(
                np.linspace(0, len(audio), new_length), np.arange(len(audio)), audio
            )

        instance = ggwave.init()
        chunk_size = 1024
        decoded_text = ""

        for i in range(0, len(audio), chunk_size):
            chunk = audio[i : i + chunk_size]
            chunk_bytes = chunk.tobytes()

            res = ggwave.decode(instance, chunk_bytes)
            if res is not None:
                try:
                    decoded_text = res.decode("utf-8")
                    break
                except:
                    pass

        ggwave.free(instance)

        if decoded_text:
            return decoded_text
        else:
            return "❌ No ggwave signal detected in audio"

    except Exception as e:
        return f"❌ Decoding failed: {str(e)}"


encode_interface = gr.Interface(
    fn=encode_text_to_audio,
    inputs=[
        gr.Textbox(
            label="Text to Encode",
            placeholder="Enter text to convert to audio...",
            lines=3,
        ),
        gr.Slider(
            minimum=0,
            maximum=8,
            value=1,
            step=1,
            label="Protocol ID",
            info="ggwave protocol (0-8, affects speed/reliability)",
        ),
        gr.Slider(
            minimum=0,
            maximum=100,
            value=20,
            step=1,
            label="Volume",
            info="Audio volume (0-100)",
        ),
    ],
    outputs=gr.Audio(label="Generated Audio", type="numpy"),
    title="📤 Encode Text to Audio",
    description="Convert text into an audio signal using ggwave protocol",
    examples=[
        ["Hello World", 1, 20],
        ["GGWave is cool!", 1, 30],
        ["Testing 123", 2, 25],
    ],
    theme="default",
)

decode_interface = gr.Interface(
    fn=decode_audio_to_text,
    inputs=gr.Audio(
        label="Upload Audio File", type="filepath", sources=["upload", "microphone"]
    ),
    outputs=gr.Textbox(label="Decoded Text", lines=5),
    title="📥 Decode Audio to Text",
    description="Extract text from an audio signal using ggwave protocol",
    examples = [
        ["./audio.wav"],
    ],
    theme="default",
)

demo = gr.TabbedInterface(
    [encode_interface, decode_interface],
    tab_names=["📤 Encode", "📥 Decode"],
    title="🎵 GGWave Audio Encoder/Decoder",
)

demo.launch(debug=True, mcp_server=True)