app.py

import time
import gradio as gr
import requests
import tempfile
import os

token = os.environ['apikey']
API_HOST = "https://api.voicemod.net"




def getPayload(index, audio_file):
    new_name = audio_file.replace("-0-100.wav", ".m4a")

    command = f"ffmpeg -y -f wav -i {audio_file} -f ipod {new_name}"
    print("Executing command: " + command)
    os.system(command)

    voiceIds = [        
        "banana",
        "bob",
        "pilot",
        "drone", 
        "evil", 
        "ogre",
        "cave"
    ]


    payload={'voiceId': voiceIds[index]}
    files=[
        ('audioFile',('audio.m4a',open(new_name,'rb'),'application/octet-stream'))
    ]


    return payload, files

 

def cleanUpLines(lines):
    return list(filter(None, lines))


def greet(index, audio_file):
    url = API_HOST + "/v1/cloud/audio?convertMp3=true"


    print("URL: " + url)
    print("Audio file path: ", audio_file)

    payload, files = getPayload(index, audio_file)

    headers = {
        'x-api-key': token,
    }


    print(payload)
    print("Before the call")
    response = requests.request("POST", url, headers=headers, data=payload, files=files)
    print("After the call...")
    print(response.status_code)
    print(response.text)

    jsonResp = response.json()
    print(jsonResp)

    return gr.make_waveform(download_file(jsonResp['convertedFileUrl']))


def download_file(url):
    response = requests.get(url)
    if response.status_code == 200:
        with tempfile.NamedTemporaryFile(delete=False) as tmp_file:
            tmp_file.write(response.content)
            tmp_file.flush()
            return tmp_file.name
    else:
        print("Error: Unable to download file")


with gr.Blocks() as demo:

    gr.Markdown("""
        ## Voicemod's Speech-to-Speech API

        **Try some of our voice filters on your own voice with this demo!**

        To use the demo, follow the instructions:

        1. First, select the voice you want from the dropdown
        2. Then record yourself saying someting funny (or not)
        3. ???
        4. PROFIT! (or rather, play the resulting audio)

        ### Recommendation
        For better results, make sure you leave a small amount of silence at the end of the recording (1 second or less is fine).
        Also, a little bit of acting never hurts. Especially for the ones like **Evil**, results are a lot more fun when you do some acting while speaking :)

        ## Join our Community
        If you'd like to know more and meet other developers working with this technology, join our [Discord Server](https://discord.gg/vm-dev-community)!
        """)

    voices = [
        "Banana",
        "Bob",
        "Pilot",
        "Drone",
        "Evil",
        "Ogre",
        "Cave"
    ]

    with gr.Row():                
        with gr.Column():
            with gr.Row():
                dd = gr.Dropdown(choices=voices, type="index", label="1. Select the voice...")
                audioInput = gr.Audio(type="filepath", 
                                        label="2. Record yourself saying something...", 
                                        source="microphone",
                                       )

            with gr.Row():
                btn = gr.Button("Run")
        with gr.Column():
            video = gr.Video(label="3. Play the generated audio")
            video.style(height=300)
        
    gr.Markdown("""
        ## Want to use this API for your project?

        If you'd like to use this API for your own project, request access through our [form here](https://voicemod.typeform.com/to/KqeNN6bO) and join
        our [Dev Community on Discord](https://discord.gg/vm-dev-community)!
    """)

    btn.click(fn=greet, 
             inputs=[
                    dd,
                    audioInput
              ],
                outputs=video)

demo.launch()