app.py

import gradio as gr
# from TTS.api import TTS
import librosa
import http.client
# Python 2.x引入urllib模块。
from restful_aliyun_tts import *
import urllib.parse
from base_chat import *
import json
from training import *
from process import *

title = "文本转语音"
# embed_record() # 把历史聊天记录向量化

def generateAudio(text):
    token = "e329d9b11a1d4fa68e2a212c5a7892d7"
    appKey = "YdRfHrZVeusHKfv6"
    # text = '我是A i Vtuber,很高兴认识你,你有什么开心的事情嘛。'
    # 采用RFC 3986规范进行urlencode编码。
    result = run_question(text)
    # textUrlencode = text
    textUrlencode = result
    # Python 2.x请使用urllib.quote。
    # textUrlencode = urllib.quote(textUrlencode, '')
    # Python 3.x请使用urllib.parse.quote_plus。
    textUrlencode = urllib.parse.quote_plus(textUrlencode)
    textUrlencode = textUrlencode.replace("+", "%20")
    textUrlencode = textUrlencode.replace("*", "%2A")
    textUrlencode = textUrlencode.replace("%7E", "~")
    # print('text: ' + textUrlencode)
    audioSaveFile = 'syAudios.wav'
    format = 'wav'
    sampleRate = 16000

    processGETRequest(appKey, token, textUrlencode, audioSaveFile, format, sampleRate)
    
    # audio, sr = librosa.load(path="syAudios.wav")
    
    # return  sr,audio
    return audioSaveFile

def generateAudio_from_file(file):
    # 验证获取文件的内容 ,需要把加载文件的内容 进行学习训练，然后再进行生成
    # print(file.name)
    files_path = "/".join((file.name).split("/")[0:-1]) + "/"
    print(files_path)
    res = train(files_path)
    print(res)
    # with open(file.name,"r",encoding="utf-8") as f:
    #     content = f.read()
    #     print(content)

def generateAudio_wav_file(text):
    token = "e329d9b11a1d4fa68e2a212c5a7892d7"
    appKey = "YdRfHrZVeusHKfv6"
    # print(text.name)
    print(text)
    res = runPrompt(text)
    print(res)
    textUrlencode = res
    textUrlencode = urllib.parse.quote_plus(textUrlencode)
    textUrlencode = textUrlencode.replace("+", "%20")
    textUrlencode = textUrlencode.replace("*", "%2A")
    textUrlencode = textUrlencode.replace("%7E", "~")
    # print('text: ' + textUrlencode)
    audioSaveFile = 'syAudios.wav'
    format = 'wav'
    sampleRate = 16000

    processGETRequest(appKey, token, textUrlencode, audioSaveFile, format, sampleRate)

    # audio, sr = librosa.load(path="syAudios.wav")

    # return  sr,audio
    return audioSaveFile


example = [["Ai Vtuber是什么"],["where is nanjing"],["who are you"],["今天吃什么好呢"],["武汉今天天气咋样"],["what are the Annie skills"]]

app = gr.Blocks()
with app :
    gr.Markdown("# 文本转语音以及根据加载的文本来回答\n\n"
                "Text：根据提问的问题回答之后转语音\n\n"
                "File：根据加载的文本通过训练后，按照文本来进行回答，只回答相关的问题\n\n")

    with gr.Tabs():
        with gr.TabItem("TTS"):
            with gr.Row():
                with gr.Column():
                    textbox = gr.Textbox(label="Text",
                                         placeholder="Type your sentence here",
                                         value="what is Ai Vtuber",elem_id="tts-input")
                    filebox = gr.File(file_count="single",
                                      type="file",
                                      label=None
                                    )
                with gr.Column():
                    # text_output = gr.Textbox(label="Message")
                    # 通过对话
                    audio_output = gr.Audio(label="Output Audio",elem_id="tts-audio")
                    btn = gr.Button("Generate")
                    btn.click(fn=generateAudio,
                              inputs=[textbox],
                              outputs=[audio_output])
                    # 通过文本进行学习之后的 只回答和文本相关的内容 并且输出成音频还是
                    audio_output2 = gr.Audio(label="Output Audio", elem_id="tts-audio-from-file")
                    btn2 = gr.Button("Training data(click onece can train data)")
                    btn2.click(fn=generateAudio_from_file,
                              inputs=[filebox],
                              outputs=[audio_output2])
                    # 经过训练后，点击进行生成语音
                    btn3 = gr.Button("Generate from file")
                    btn3.click(fn=generateAudio_wav_file,
                               inputs=[textbox],   # 传入的是文本 搜索的是经过向量化后的文本
                               outputs=[audio_output2])


        gr.Examples(examples=example,
                    inputs=[textbox],
                    outputs=[audio_output],
                    fn=generateAudio)

#app.launch(share=True)
app.launch()

# gr.Markdown("通过提问，然后输出结果是语音\n 加载文本，然后根据文本的内容回答问题")

# app = gr.Interface(
#     fn=generateAudio,
#     inputs="text",
#     outputs="audio",
#     title=title,
#     # streaming=True,
#     # live=True,
#     examples=[("Ai Vtuber是什么"),("where is nanjing"),("who are you"),("今天吃什么好呢"),("武汉今天天气咋样")]
#     )

# app.launch(server_name="0.0.0.0", server_port=7865, share=True)
# app.launch(share=True)
# app.launch()