Some toy projects I made:
-
A platform-independent bot framework that is interacted purely based on string. Developed new supports on (infinite) subcommands and keyword args like
/cmd subcmd args kwarg=val. Developed and pruned based on telegrambot. -
Example usage of
dumb_botthat queries and subscribe tv programs. The bot interactively ask for tv information like channel name and program title then subscribe that program. Also supports scheduled queries. -
A collection of small automated WeCom bots, including automated website check-in bot with clicking captcha solvers, scheduled RSS crawl and notification bot.
-
A Python wrapper for wecom APIs based on pydantic. Used by all my wecombot-related projects.
-
-
A hub providing RSS feeds for all kinds of sources.
I added support for crawling and generating RSS feeds for 8 hosts, such as this and this and so on.
-
A movie metadata scraper that works with Emby, a self-hosted multimedia platform.
I contributed 8 major bug fixes and 2 features through PR #475 and PR #476.
-
A tool that restores LAN multiplayer support for a game after the original servers shut down.
I contributed through PR #12, mainly include
- Redesign the packet hook logic for more stable discovery and connection.
- Implement new features to improve connections with multiple interfaces, like automatic troubleshooting, logging, hosts file effective checking.
The work mainly requires analyzing packets, configuring route table, and use pydivert to capture and rewrite packets automatically.
-
Tampermonkey scripts (Tampermonkey is a Web browser extension for running custom JavaScripts)
-
Assist class enrollments.
-
Resolve hidden download links and add download buttons.
-
An web crawling automation script that reads your course schedule and convert them to .ics files to be imported to mobile phone.
-
Scripts for automatically making check-in reports.
-
A Flask-based personal web service. It is mainly used as an endpoint for responding my WeCom bot messages, with some other features like a dashboard for redirecting to all web services I have deployed, and exposing webhook endpoints for FreshrssCrawler. Also include two WeCom bots:
TVSubscriberBotfor subscribing TV programs andDrinkBotfor recording daily drink intake. Both bots support interactive dialogues and scheduled tasks.
-
A MineSweeper game based on PyQt5 that simulates classic minesweeper game interface, with a customized solver algorithm and a DQN model solver.
-
An CLI automation tool that clean closed captions of tv programs.
-
An GUI automation tool based on PyQt5 that monitors and creates torrents for local files, and automatically publish them to BT sites.
-
Used as a crontab task on servers. Monitors connections made to specified ports, block and log their info to a table if connections were made too often.
-
Data reconstruction attack in Federated Learning based on Inference-time Conditional Diffusion Models. Proficient in PyTorch.
-
MIPS five-level pipeline CPU based on Verilog.
-
A collection of notes gradually accumulated from everyday learning over my entire life. Temporarily public for showcase purposes.
-
wol:
An asus router plugin of koolshare firmware that fixes the original feature of sending WOL (Wake-on-LAN) packets.
我做过的一些玩具项目:
-
一个不依赖通信应用平台的机器人框架,纯基于字符串进行交互。实现(无限)链式子命令、键值参数功能,例如
/cmd subcmd args kwarg=val,以及交互式对话。基于 telegrambot 改进和精简而来。 -
dumb_bot的一个实际用例,用于查询和订阅电视节目。机器人会通过交互方式询问频道名、节目标题等电视节目信息,然后订阅该节目;同时也支持定时自动查询。 -
一组小型自动化企业微信机器人,包括支持顺序点击验证码的网站自动签到机器人、RSS定时爬虫推送机器人。
-
一个基于pydantic的企业微信 API 的 Python 封装,我所有企业微信机器人相关项目都在使用。
-
-
一个RSS源中心,提供各种网站的RSS源。
-
一个电影元数据刮削器,配合自建媒体平台Emby使用。
-
一个恢复游戏LAN联机功能的工具。
我通过PR#12贡献了改进,涉及数据包分析、路由表配置、pydivert 自动抓取并重写数据包等工作。改进主要包括:
- 重新设计数据包hook逻辑,提升局域网房间搜索和连接稳定性。
- 增加自动排障、运行日志、hosts生效检查等功能,方便定位多网卡/VPN环境下的连接问题。
-
Tampermonkey 脚本(Tampermonkey 是用于运行自定义 JavaScript 的浏览器扩展)
-
辅助抢课。
-
解析隐藏下载链接并添加下载按钮。
-
一个网页爬取自动化脚本,用于读取课程表并转换为可导入手机日历的 .ics 文件。
-
自动打卡脚本。
-
一个基于 Flask 的个人网络服务。主要作为企业微信机器人的消息回复端,同时也包含一些其他功能如 自建服务重定向 Dashboard,以及为 FreshrssCrawler 暴露 webhook 接口。项目还包含两个企业微信机器人:
TVSubscriberBot用于订阅电视节目,DrinkBot用于记录每日饮水量。两个机器人都支持交互式对话和定时任务。
-
一个基于 pyqt5 的扫雷游戏,模拟经典扫雷界面,并包含自定义求解算法和 DQN 模型求解器。
-
一个用于清理电视节目台词字幕的命令行自动化工具。
-
一个基于 PyQt5 的 GUI 自动化工具,用于监控本地文件、创建种子,并自动发布到 BT 站点。
-
作为服务器上的 crontab 任务使用。监控指定端口的连接,如果连接过于频繁,则阻断并将连接信息记录到表格中。
-
基于推理时条件扩散模型的联邦学习数据重建攻击。熟练掌握PyTorch使用。
-
基于 Verilog 的 MIPS 五级流水线 CPU。


