tomriddle/cuda

1import pathlib2import uuid3import os4import gradio as gr5from tqdm import tqdm6import requests7import urllib.request8import json9import time10output_mp3="output.mp3"11 12def upload_image(img: str, d_id_key: str):13    url = "https://api.d-id.com/images"14 15    files = {"image": ("hero.jpg", open(img, "rb"), "image/jpg")}16    headers = {17        "accept": "application/json",18        "authorization": "Basic "+d_id_key19    }20 21    response = requests.post(url, files=files, headers=headers)22 23    response_dict = response.json()24    img_url = response_dict["url"]25    # return audio_url26    print(img_url)27    return img_url28 29def upload_audio(audio: str, d_id_key: str):30    url = "https://api.d-id.com/audios"31 32    files = {"audio": (audio, open(audio, "rb"), "audio/mpeg")}33    headers = {34        "accept": "application/json",35        "authorization": "Basic "+d_id_key36    }37    response = requests.post(url, files=files, headers=headers)38    response_dict = response.json()39    audio_url = response_dict["url"]40    # return audio_url41    print(audio_url)42    return audio_url43 44def get_did_video(process_video_url,d_id_key):45    url = "https://api.d-id.com/talks/"+process_video_url46 47    headers = {48        "accept": "application/json",49        "authorization": "Basic "+d_id_key50    }51    response_dict = {}52 53    while "result_url" not in response_dict:54      # make API call and get response dictionary55      response = requests.get(url, headers=headers)56      response_dict = response.json()57      58      print(response.text)59      60      # wait for 1 second before checking again61      time.sleep(1)62      63    # "result_url" key is now present in the dictionary64    result_url = response_dict["result_url"]65 66    print("From did_video \n\n\n")67    print("/n/n/n")68 69    # response_dict = response.json()70    result_url = response_dict["result_url"]71 72    print(result_url)73    return result_url74 75def text_to_speach_api(text: str, elv_key,voice_id: str):76    url = "https://api.elevenlabs.io/v1/text-to-speech/"+voice_id+"/stream"77    headers = {78        "accept": "*/*",79        "xi-api-key": elv_key,80        "Content-Type": "application/json",81    }82    data = {83        "text": text,84        "voice_settings": {85            "stability": 0,86            "similarity_boost": 087        }88    }89 90    response = requests.post(url, headers=headers, json=data)91    # print(response.text)92 93    if response.ok:94        with open("output.mp3", "wb") as f:95            f.write(response.content)96    else:97        print("Error: ", response.text)98 99def get_voice_names():100    with open("data.json") as f:101        data = json.load(f)102    return [voice["name"] for voice in data["voices"]]103 104 105 106# define a function to get voice id by name107def get_voice_id(name):108    # load the JSON data109    with open("data.json") as f:110        data = json.load(f)111    for voice in data['voices']:112        if voice['name'] == name:113            return voice['voice_id']114    return None115 116#D-id API117def d_id_api(image_url, d_id_key,audio_url):118    print("D-id API")119    url = "https://api.d-id.com/talks"120    payload = {121        122    "source_url": image_url,123    "script": {124        "type": "audio",125        "audio_url": audio_url,126    }127 128    }129    headers = {130      "accept": "application/json",131      "content-type": "application/json",132    "authorization": "Basic "+d_id_key133  }134 135    response = requests.post(url, json=payload, headers=headers)136    print("From D-id API \n\n\n")137    print(response.text)138    response_dict = response.json()139    process_video = response_dict["id"]140    # return audio_url141    print(process_video)142    return process_video143    144 145 146def transcribe_video(d_id_key: str, elv_key: str, full_text: str,voice_name: str,img):147    print(voice_name)148    voice_id=get_voice_id(voice_name)149    text_to_speach_api(full_text, elv_key,voice_id)150    audio_url=upload_audio(output_mp3,d_id_key)151    image_url=upload_image(img,d_id_key)152    process_video_url=d_id_api(image_url, d_id_key,audio_url)153    video_url=get_did_video(process_video_url,d_id_key)154    file_name = 'hero.mp4'155    urllib.request.urlretrieve(video_url, file_name)156    return file_name157    158 159examples = [["", "","Good morning, it's great to see you! I hope you're having a wonderful day. I just wanted to say thank you for taking the time to speak with me. Is there anything new or exciting happening in your life? I'd love to hear about it. Let's catch up soon!",160             "Arnold","./images/hero.jpg"],["","","Hello there, I'm a talking photo! I can speak any text you type here. Try it out!", "Domi","./images/3.jpg"],["","","Hello there, I'm a talking photo! I can speak any text you type here. Try it out!", "Domi","./images/2.jpg"]]161 162demo = gr.Interface(fn=transcribe_video, inputs=[163    gr.Textbox(label="D-Id API Key",placeholder="Paste your D-Id",type='password'),164    gr.Textbox(label="Elevenlabs API Keys",placeholder="Paste Elevenlabs",type='password'),165    gr.Textbox(lines=4, label=" Please input the text you wish to generate in order to make the photo speak.", placeholder="English Text here"),166    gr.Dropdown(choices=get_voice_names(), label="Select a voice"),167    gr.Image(label="photo of a Person", type="filepath")168  ], outputs="video",title="Bring your images to life with the talking animation feature now!",examples=examples,cache_examples=False)169 170demo.launch()