Hugging Face's logo

Spaces:
zxsipola123456
/
tts
Paused

App Files Community
tts
File size: 5,575 Bytes 
9653cf4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
 
import gradio as gr
import os
import tempfile
from openai import OpenAI
from tts_voice import tts_order_voice
import edge_tts
import tempfile
import anyio


import torch
import torchaudio
import gradio as gr
from scipy.io import wavfile
from scipy.io.wavfile import write


knn_vc = torch.hub.load('bshall/knn-vc', 'knn_vc', prematched=True, trust_repo=True, pretrained=True, device='cpu')

language_dict = tts_order_voice

async def text_to_speech_edge(text, language_code):
    voice = language_dict[language_code]
    communicate = edge_tts.Communicate(text, voice)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
        tmp_path = tmp_file.name

    await communicate.save(tmp_path)

    return "语音合成完成：{}".format(text), tmp_path


def voice_change(audio_in, audio_ref):
    samplerate1, data1 = wavfile.read(audio_in)
    samplerate2, data2 = wavfile.read(audio_ref)
    write("./audio_in.wav", samplerate1, data1)
    write("./audio_ref.wav", samplerate2, data2)

    query_seq = knn_vc.get_features("./audio_in.wav")
    matching_set = knn_vc.get_matching_set(["./audio_ref.wav"])
    out_wav = knn_vc.match(query_seq, matching_set, topk=4)
    torchaudio.save('output.wav', out_wav[None], 16000)
    return 'output.wav'


def tts(text, model, voice, api_key,base_url):
    if len(text)>300:
        raise gr.Error('您输入的文本字符多于300个，请缩短您的文本')
    if api_key == '':
        raise gr.Error('Please enter your OpenAI API Key')
    else:
        try:
            client = OpenAI(api_key=api_key,base_url='https://lmzh.top/v1')

            response = client.audio.speech.create(
                model=model, # "tts-1","tts-1-hd"
                voice=voice, # 'alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'
                input=text,
            )

        except Exception as error:
            # Handle any exception that occurs
            raise gr.Error("An error occurred while generating speech. Please check your API key and try again.")
            print(str(error))

    # Create a temp file to save the audio
    with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as temp_file:
        temp_file.write(response.content)

    # Get the file path of the temp file
    temp_file_path = temp_file.name
        
    return temp_file_path


app = gr.Blocks()

with app:
    gr.Markdown("# <center>🌟 - OpenAI TTS + AI变声</center>")
    gr.Markdown("### <center>🎶 地表最强文本转语音模型 + 3秒实时AI变声，支持中文！Powered by [OpenAI TTS](https://platform.openai.com/docs/guides/text-to-speech) and [KNN-VC](https://github.com/bshall/knn-vc) </center>")
    with gr.Tab("🤗 OpenAI TTS"):
        with gr.Row(variant='panel'):
          api_key = gr.Textbox(type='password', label='OpenAI API Key', placeholder='请在此填写您的OpenAI API Key')
          model = gr.Dropdown(choices=['tts-1','tts-1-hd'], label='请选择模型（tts-1推理更快，tts-1-hd音质更好）', value='tts-1')
          voice = gr.Dropdown(choices=['alloy', 'echo', 'fable', 'onyx', 'nova', 'shimmer'], label='请选择一个说话人', value='alloy')
        with gr.Row():
          with gr.Column():
            inp_text = gr.Textbox(label="请填写您想生成的文本（中英文皆可）", placeholder="想说却还没说的 还很多 攒着是因为想写成歌", lines=5)
            btn_text = gr.Button("一键开启真实拟声吧", variant="primary")
    
          with gr.Column():
            inp1 = gr.Audio(type="filepath", label="OpenAI TTS真实拟声", interactive=False)
            inp2 = gr.Audio(type="filepath", label="请上传AI变声的参照音频（决定变声后的语音音色）")
            btn1 = gr.Button("一键开启AI变声吧", variant="primary")
          with gr.Column():
            out1 = gr.Audio(type="filepath", label="AI变声后的专属音频")
          btn_text.click(tts, [inp_text, model, voice, api_key], inp1)
          btn1.click(voice_change, [inp1, inp2], out1)
    with gr.Tab("⚡ Edge TTS"):
        with gr.Row():
            input_text = gr.Textbox(lines=5, placeholder="想说却还没说的 还很多 攒着是因为想写成歌", label="请填写您想生成的文本（中英文皆可）")
            default_language = list(language_dict.keys())[15]
            language = gr.Dropdown(choices=list(language_dict.keys()), value=default_language, label="请选择文本对应的语言")
            btn_edge = gr.Button("一键开启真实拟声吧", variant="primary")
            output_text = gr.Textbox(label="输出文本", visible=False)
            output_audio = gr.Audio(type="filepath", label="Edge TTS真实拟声")

        with gr.Row():
            inp_vc = gr.Audio(type="filepath", label="请上传AI变声的参照音频（决定变声后的语音音色）")
            btn_vc = gr.Button("一键开启AI变声吧", variant="primary")
            out_vc = gr.Audio(type="filepath", label="AI变声后的专属音频")

        btn_edge.click(text_to_speech_edge, [input_text, language], [output_text, output_audio])
        btn_vc.click(voice_change, [output_audio, inp_vc], out_vc)

        
    gr.Markdown("### <center>注意❗：请不要生成会对个人以及组织造成侵害的内容，此程序仅供科研、学习及个人娱乐使用。Get your OpenAI API Key [here](https://platform.openai.com/api-keys).</center>")
    gr.HTML('''
        <div class="footer">
         <p>Power by sipola </p>
        </div>
    ''')

app.launch(show_error=True)