yu auto reply

yu-auto-reply Automatic reply

Author: Programmer Fish Skin

Programming Learning Circle: Programming Navigation Knowledge Planet

[toc]

The platform monitoring and automatic reply tool based on Java Spring Boot supports flexible configuration of multiple monitoring tasks and supports one-click deployment!

Demo video: https://www.bilibili.com/video/BV1WX4y1o7aL

This project adopts multiple design modes to decouple monitors and respondents, and can flexibly configure monitoring on multiple different platforms and bind different types of automatic replies.

? If you like this project, thank you for starting Diandian Star. The author may provide more platform monitoring support in the future.

Functional Features

Monitoring capability

• Knowledge Planet Question Monitoring
• Default monitoring (simulated data)

Response ability

• OpenAI answer (supports self-selected models, such as gpt-4)
• Default monitoring (simulated data)

Configuration capability

• Supports configuring multiple tasks
• Each task can be flexibly specified for monitoring and answering methods

Deployment capability

• Supports Docker containerized deployment
• Supports Railway one-click deployment
• Support dynamically specifying environment variables to change configuration

Quick Start

1) Modify the application.yml configuration, which mainly contains 3 parts:

• openAI configuration (need to have an API key)
• Knowledge Planet Configuration (need to obtain cookies yourself)
• Task configuration

The detailed configuration is as follows:

 # openAI 配置
# https://platform.openai.com/docs/api-reference
openai :
  model : ${OPENAI_MODEL:text-davinci-003}
  apiKey : ${OPENAI_API_KEY:你的apiKey}
# 知识星球配置
# https://zsxq.com/
zsxq :
  cookie : ${ZSXQ_COOKIE:你的星球cookie}
  groupId : ${ZSXQ_GROUP_ID:你的星球id}
  # 是否提醒提问者
  silenced : ${ZSXQ_SILENCED:true}
# 任务配置
task :
  # 任务列表，支持配置多个
  list :
    - name : task1 #任务名
      monitor : zsxq #监控者
      answerer : openai #回答者
      cron : ' 0/30 * * * * ? ' #执行周期

2) Just run the main class MainApplication directly

One-click deployment

After clicking the above deployment button, the environment variable will be automatically recognized and changed to your own:

Architectural design

One picture is better than a thousand words:

The most critical design of this project is to decouple the monitor and the respondent . You can monitor any platform and bind different automatic answers to each platform (such as OpenAI).

Implementation key:

1. Define Answerer interface, method of unified answering
2. Define the Monitor monitor interface, unify the monitoring method, and automatically reply to the monitored messages through Answerer callback parameters.
3. Use the mediator mode, use the JobMediator class to combine Monitor and Answerer instead of forcibly binding the responder and monitor
4. Use factory mode to generate monitors and responders based on configuration
5. Use Spring Scheduler to read the yml configuration to automatically create multitasking

Development

Custom monitoring

1) Write a class to implement monitor/Monitor abstract class

2) Modify the createMonitor method of factory/MonitorFactory to supplement the creation of your own monitor

Custom answers

1) Write a class to implement answerer/Answerer interface

2) Modify the createAnswerer method of factory/AnswererFactory to supplement the creation of your own answerer

Free ChatGPT Exchange Group

Welcome to contribute

The author is usually very busy , and this project only took a few hours to do it. It is open source for everyone to refer to, but the PR and Issues response will not be very timely. Thank you for your understanding!

If you have any questions about the project itself, please contact issues and PR;

If you have programming problems or need project teaching, please see Programming Navigation Knowledge Planet

Q&A

1) Question: Why support the Knowledge Planet first?

Answer: Because OpenAI's API is not free, the planet can limit the number of questions to prevent the interface from being brushed.