Skip to main content

A simple and easy-to-use library that connects to common artificial intelligence interfaces (powered by Gradio), allowing for quick development of local web applications. It includes mainstream AI capabilities such as LLM text interaction, image understanding, audio understanding, image generation, speech recognition, and speech synthesis.

Project description

aiZero-gr是一个简单易用的可以连接常用人工智能接口,快速搭建可视化本地web应用的Python第三方库,在aiZero库的基础上,调整为基于gradio框架的界面展示。

当前版本仅支持阿里云灵积模型服务的api key和有限的功能调用。

快速开始

from aiZero_gr import AIWebApp

# 设定web应用的功能
def my_ai_function():
    pass

app = AIWebApp(title='人工智能助手')    # 初始化web应用
app.set_apikey('YOUR_API_KEY')    # 设定AI接口的api key
app.add_input_text()    # 在页面中添加一个输入文本框
app.set_submit(my_ai_function)    # 设定提交按钮点击后执行的函数功能
app.run()    # 启动应用

启动后,程序将自动输出本地网址,你将可以在浏览器中访问并查看搭建好的web应用。

如果需要实现AI功能的可视化呈现,你只需要将'YOUR_API_KEY'替换为真实的api key,并完善my_ai_function函数。

AI功能实现例子

大模型文字交互

单轮对话
from aiZero_gr import AIWebApp, text_generation

def my_ai_function():
    text = app.input_text    # 获取输入文本框的文字内容
    reply = text_generation(text)    # 调用AI接口,获取回复反馈
    app.send(reply)    # 将结果推送至前端

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()
多轮对话
from aiZero_gr import AIWebApp, text_generation

def my_ai_function():
    text = app.input_text
    history = app.chat_history    # 获取对话历史记录
    reply = text_generation(text, history)    # 将历史记录同时提交
    app.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()
设置系统指令
from aiZero_gr import AIWebApp, text_generation

def my_ai_function():
    text = app.input_text
    history = app.chat_history
    # prompt参数可以设定系统指令,设定模型的行为要求
    reply = text_generation(text, history, prompt='你是一个10岁的小学生,名叫小幻')
    app.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()
流式输出
from aiZero_gr import AIWebApp, text_generation

def my_ai_function():
    text = app.input_text
    history = app.chat_history
    # stream=True可设置为流式输出,目前文字交互、图像理解、声音理解支持该参数
    reply = text_generation(text, history, stream=True)
    # 流式输出的返回结果是生成器,可以用for循环按顺序推送
    for r in reply:
	    app.results.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()

图像理解

from aiZero_gr import AIWebApp, image_understanding

def my_ai_function():
    text = app.input_text
    img = app.input_pic    # 获取输入的图像
    history = app.chat_history
    reply = image_understanding(img, text, history)
    app.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.add_input_pic()    # 添加图像输入组件
app.set_submit(my_ai_function)
app.run()

声音理解

from aiZero_gr import AIWebApp, audio_understanding

def my_ai_function():
    text = app.input_text
    audio = app.input_audio    # 获取输入的音频
    history = app.chat_history
    reply = audio_understanding(audio, text, history)
    app.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.add_input_audio()    # 添加音频输入组件
app.set_submit(my_ai_function)
app.run()

图像生成

from aiZero_gr import AIWebApp, image_generation

def my_ai_function():
    text = app.input_text
    reply = image_generation(text)
    app.send(reply, 'image')    # 推送图像时,需要明确类型为image

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()

人物图像风格重绘

from aiZero_gr import AIWebApp, human_repaint

def my_ai_function():
    img = app.input_pic
    reply = human_repaint(img)
    app.send(reply, 'image')

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_pic()
app.set_submit(my_ai_function)
app.run()

human_repaint函数可以接受style参数设定风格类型,可选值为0~9的数字(默认值为7)。

涂鸦作画

from aiZero_gr import AIWebApp, sketch_to_image

def my_ai_function():
    text = app.input_text    # 涂鸦作画的提示文字
    img = app.input_pic    # 涂鸦草图图像
    reply = sketch_to_image(img, text)
    app.send(reply, 'image')

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.add_input_pic()
app.set_submit(my_ai_function)
app.run()

sktech_to_image函数可以接受style参数设定风格类型,包括:

  • "<3d cartoon>":3D 卡通
  • "<anime>":二次元(默认值)
  • "<oil painting>":油画
  • "<watercolor>" :水彩
  • "<flat illustration>":扁平插画

语音识别

from aiZero_gr import AIWebApp, speech_recognition

def my_ai_function():
    audio = app.input_audio
    reply = speech_recognition(audio)
    app.send(reply)

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_audio()
app.set_submit(my_ai_function)
app.run()

所用接口支持中英文双语的语音识别。

语音合成

from aiZero_gr import AIWebApp, speech_synthesis

def my_ai_function():
    text = app.input_text
    reply = speech_synthesis(text)
    app.send(reply, 'audio')    # 将音频结果推送到前端

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()

speech_synthesis函数可以接受以下参数:

  • rate:设定语速快慢,取值范围0.5~2,默认值为1。
  • pitch:设定语调高低,取值范围0.5~2,默认值为1。
  • voice:设定使用的音色,详细列表参见链接

进阶使用

多种AI功能的混合

aiZero-gr支持方便地将不同的AI功能进行链式整合,例如,将文字交互、语音识别、语音合成结合起来,可以实现语音交互。

运用时,注意系统默认的chat_history为前端可见的对话历史记录,如果需要将前端不可见的数据作为传递给AI的对话历史,则需要手动维护对话历史记录

from aiZero_gr import AIWebApp, text_generation, speech_recognition, speech_synthesis

def my_ai_function():
    audio = app.input_audio
    text = speech_recognition(audio)
    reply = text_generation(text, history)
    # 文字内容发生未推送到前端,需要手动添加
    history.add_user_content(text)
    history.add_ai_content(reply)
    reply_audio = speech_synthesis(reply)
    app.send(reply_audio, 'audio')

app = AIWebApp(title='人工智能助手')
history = app.create_history()    # 手动创建对话历史
app.set_apikey('YOUR_API_KEY')
app.add_input_audio()
app.set_submit(my_ai_function)
app.run()

同时输出多种不同类型的信息

aiZero支持同时输出文字、音频和图像等多种不同类型的信息,但每种类型的信息只能为一个。例如,我们可以输入文字,同时以文本和音频形式输出AI的响应。

from aiZero_gr import AIWebApp, text_generation, speech_synthesis

def my_ai_function():
    text = app.input_text
    history = app.chat_history
    reply = text_generation(text, history)
    reply_audio = speech_synthesis(reply)
    app.send(reply)
    app.send(reply_audio, 'audio')

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()

同次分时输出多个信息

aiZero也支持在一次大模型交互的输出时,分不同时间输出不同信息。例如,我们制作一个绘图助手,它可以将用户输入的简易文字信息扩充为详细的画面信息后,再交给图像生成功能完成图像的创作。在这个功能设计中,AI的文字回复响应很快而图像生成响应较慢,因此应当在不同时间推送至前端。

from aiZero_gr import AIWebApp, text_generation, image_generation

def my_ai_function():
    text = app.input_text
    reply = text_generation(text, prompt='你是一个人工智能绘图助手,你负责将用户输入信息中描述的画面扩充为详细的绘图AI的提示词')
    app.send(reply)
    img = image_generation(reply)
    app.send(img, 'image')

app = AIWebApp(title='人工智能助手')
app.set_apikey('YOUR_API_KEY')
app.add_input_text()
app.set_submit(my_ai_function)
app.run()

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

aiZero-gr-0.0.2a1.tar.gz (11.6 kB view details)

Uploaded Source

Built Distribution

aiZero_gr-0.0.2a1-py3-none-any.whl (9.5 kB view details)

Uploaded Python 3

File details

Details for the file aiZero-gr-0.0.2a1.tar.gz.

File metadata

  • Download URL: aiZero-gr-0.0.2a1.tar.gz
  • Upload date:
  • Size: 11.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.10.3

File hashes

Hashes for aiZero-gr-0.0.2a1.tar.gz
Algorithm Hash digest
SHA256 a13311e994f18903f4567788046443a5dd849cdc6757a10636d8e702a7c243c0
MD5 555c100f574eddaca670b3f386a71a8b
BLAKE2b-256 7688c50be0e6e2e2c60e9e2a4356a0886f6dd7fd4b2312049e0ceafdadf26ac2

See more details on using hashes here.

File details

Details for the file aiZero_gr-0.0.2a1-py3-none-any.whl.

File metadata

  • Download URL: aiZero_gr-0.0.2a1-py3-none-any.whl
  • Upload date:
  • Size: 9.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.10.3

File hashes

Hashes for aiZero_gr-0.0.2a1-py3-none-any.whl
Algorithm Hash digest
SHA256 4c072e4e15b50a72c94e285ae01cdbc43055abd061c301e6db6e9ace07dc8680
MD5 8cfbb4e5fa94bfa0302af2aaba28fd0c
BLAKE2b-256 386cb7ba7025392566062d922da3e98d374a205b58e95227c5f43ae673517fe4

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page