Druid is a cross-platform, data-driven GUI toolkit for Rust, providing a reactive widget system with immutable state management via the `Data` trait and composable `Lens` for data access. It supports multiple backends (macOS, Windows, GTK, X11, Wayland, WebAssembly) through a platform abstraction layer (`druid-shell`). The framework includes a rich set of built-in widgets (Flex, Scroll, Split, TextBox, etc.), a layout system inspired by Flutter's `BoxConstraints`, an environment-based theming system (`Env`), and a command/notification system for inter-widget communication. The project is explicitly discontinued and unmaintained, with its successor being Xilem. Despite this, the codebase exhibits high craftsmanship with clean abstractions, well-documented patterns (Controller, Scope, Lens combinators), and a modular architecture that influenced later Rust UI frameworks. However, technical drift is high due to outdated dependencies, incomplete backend implementations (especially Wayland and Web), and deprecated API usage.
This repository contains educational labs for MIT's Introduction to Deep Learning course (6.S191). It consists of Jupyter notebooks covering deep learning fundamentals (Lab1: music generation with RNNs), computer vision (Lab2: MNIST classification and facial detection with VAEs), and LLM fine-tuning (Lab3: style adaptation using LoRA). The repository also includes extra labs for reinforcement learning and uncertainty quantification. While structured as educational material, the labs involve substantive data processing, model training, and evaluation pipelines. The mitdeeplearning package provides reusable utilities for training visualization and data handling. The code assumes execution in Google Colab with GPU access and relies on external APIs (Comet ML, OpenRouter) and datasets hosted on third-party services (Dropbox).
This repository provides a collection of scripts and lightweight HTTP servers for automated deployment and management of sing-box multi-protocol proxy nodes. It supports VLESS-REALITY, VMess-WebSocket (with optional Cloudflare Argo tunnel), Hysteria2, and TUIC v5 protocols, along with Nezha monitoring agent integration. The primary use case is setting up proxy services on VPS or free hosting platforms like Serv00/CT8, with additional 'toy' implementations for game hosting environments using Go, Node.js, PHP, and Python. The architecture is script-driven: large Bash scripts (sing-box.sh, sb_serv00.sh) handle installation, configuration, and service management via interactive menus or environment variables. They download precompiled binaries, generate sing-box JSON configurations, and manage systemd or OpenRC services. The supplementary HTTP servers (Node.js, Python, PHP) provide subscription endpoints and process lifecycle management, downloading and running a core binary (sbsh) that performs the actual proxy setup. The PHP check.php script offers standalone server environment diagnostics. Code quality is average overall. The Bash scripts are functional but contain significant duplication across Serv00 variants and lack modularity. The Node.js and Python HTTP servers are well-structured and demonstrate clean asynchronous patterns, while the PHP diagnostic script is a polished, self-contained tool. There is no test suite or CI/CD, and external binary URLs are hardcoded, posing medium technical drift risk. The repository is actively maintained by a large community, making it a practical solution for proxy deployment despite its architectural limitations.
WebViewTvLive is an Android TV live streaming application that uses the Tencent X5 WebView to display official TV channel web pages. It loads channel playlists from remote JSON sources, caches them locally, and provides a TV-optimized UI with remote control and touch input support. The app employs a per-site adapter pattern (WebpageAdapter) to inject JavaScript and trigger fullscreen playback for different streaming websites (e.g., CCTV, YouTube, MGTV). It features a channel bar, waiting indicator, settings for aspect ratio and source selection, and automatic playlist updates. The codebase is well-structured with separation of playlist management, settings, and UI, but the project is officially paused as open source, with the last released version being 1.6. Dependencies are reasonably up-to-date, though some deprecated Android APIs are used.
This repository is an educational collection of minimal, from-scratch implementations of core LLM and generative AI systems, maintained by DataWhale. It is not a single production application; rather, it is a set of self-contained 'Tiny' sub-projects covering a small LLM trained on TinyStories, a RAG pipeline, a ReAct agent, an LLM evaluation harness, a DDPM diffusion model, a Transformer, a GraphRAG system, and an image-based RAG pipeline. Each sub-project has its own README, requirements, and runnable code or notebooks, with heavy Chinese-language comments explaining the underlying theory. The code is intentionally simple and readable rather than production-hardened, and it serves as a teaching resource for understanding how these systems work internally.
This is a research repository from the SCIR-HI lab at Harbin Institute of Technology, focused on instruction-tuning large language models (LLaMA, Alpaca, Bloom, Huozi) with Chinese medical knowledge. The core contribution is the construction of Chinese medical instruction datasets using structured knowledge bases (cMeKG) and medical literature via GPT3.5, followed by LoRA-based fine-tuning. The repository provides scripts for fine-tuning, inference, and model export, along with pre-trained LoRA weights. While it demonstrates a complete research pipeline (data generation, training, evaluation), it is primarily a research artifact rather than a production application. The code is functional and includes a clean, reusable Prompter class for template management, but lacks software engineering rigor (no tests, no CI). Reproducibility is moderate: dependencies are pinned, but the data generation process is not fully automated, and external base models and LoRA weights must be obtained separately.
This repository contains a fully functional Android application named 'Bluetooth LE Spam' that transmits crafted Bluetooth Low Energy advertisements to spoof a wide range of devices (Apple Continuity, Microsoft Swift Pair, Google Fast Pair, Samsung Easy Setup, Lovespouse) on nearby phones and tablets, triggering unwanted pop-ups and notifications. It also includes a spam detector that identifies BLE spam originating from Flipper Zeros and similar tools, running as a foreground service even on the lock screen. The application is built with Kotlin, Jetpack Compose, Material 3, Room database for persistence, and uses both legacy and modern (BLE 5) advertising modes. The codebase is well-structured around an interface-based advertisement set generator pattern and a queue handler for cycling through attacks, but it is explicitly unmaintained by the author and many exploits are patched on modern OS versions, limiting practical impact. The architecture is a standard Android single-activity app with multiple fragments and foreground services; the theme system (dynamic color, custom nav bar, status badges) demonstrates reasonable front-end craftsmanship.
EasyMocap is a multi-view 3D human motion capture toolbox designed for research and prototyping. It provides end-to-end pipelines from video input to 3D body mesh recovery, including camera calibration, 2D keypoint detection (via OpenPose, HRNet, MediaPipe, YOLO), multi-view association and triangulation, SMPL/SMPL-X body model fitting, and neural body rendering. The architecture is modular: a config-driven factory pattern (via YAML and `load_object`) instantiates pipeline stages (detection, triangulation, fitting, visualization), and a command-line entry point `emc` orchestrates the workflow. The repository also contains annotation tools, a real-time 3D visualization server using Open3D, and neural body research code (NeRF-based) for novel view synthesis. Despite its research character, the core pipeline is structured as a reusable application framework with separate packages for estimators, operations, and visualization. However, the codebase exhibits significant technical debt: hardcoded paths, duplicated modules (e.g., two HRNet implementations), incomplete error handling, mixed Chinese/English comments, and stale vendored dependencies (Eigen 3.3.7, pybind11). There are no unit tests, and many scripts rely on external model checkpoints that may not be accessible. The repository is actively maintained with recent commits, but the code quality reflects a fast-moving research project rather than a production-grade system.
This repository is a comprehensive MySQL tutorial written in Chinese (with some English terms). It covers everything from installation and configuration to advanced topics like stored procedures, triggers, views, performance optimization, and transaction/locking basics. The content is structured as a series of Markdown files organized into chapters, plus a quick reference guide ('21 minutes') and a curated list of MySQL resources ('awesome-mysql'). There is no executable code, data pipeline, or application logic; it is purely a documentation and learning resource. The repository uses the 'idoc' static site generator to produce a website, but the primary value is the tutorial content. It is well-maintained with recent updates and has a high star count (4751).
This is the official Microsoft .NET documentation repository, containing conceptual articles, tutorials, and thousands of compilable code snippets. It is not a software application, data pipeline, or research project. The primary purpose is to document .NET APIs, frameworks, and best practices through markdown and illustrative C# (and some VB, F#, C++) examples. The code snippets demonstrate realistic usage of .NET libraries, Azure SDKs, Orleans, AI/ML extensions, and more, but are designed for education and reference, not production deployment. The repository is actively maintained with frequent commits, automated build verification, and strict quality controls (markdown linting, link checking, snippet compilation). While the snippets themselves vary in modernity and some legacy examples exist, the overall drift is low because updates keep pace with .NET releases. Reusable components are limited to isolated patterns (custom JSON converters, middleware pipelines, testing utilities) that could be extracted, but most code is tightly coupled to the documentation context.
LibreraReader is a full-featured, multi-format e-book reader for Android (with an experimental iOS/desktop port via Kotlin Multiplatform). It supports PDF, EPUB, MOBI, DjVu, FB2, CBR/CBZ, DOC/DOCX, ODT, RTF, and many other formats through a combination of native C/C++ libraries (MuPDF, DjVuLibre, libmobi, libantiword, libwebp, liblame) and Java-based document translators (ODF/SVM/WMF to HTML/SVG). The app provides a rich reading experience with customizable themes, text-to-speech, OPDS catalog browsing, cloud sync (Google Drive, Dropbox), annotations, bookmarks, and a smart reflow feature for scanned documents. The codebase is large and monolithic, with significant third-party code vendored into the source tree. Development has been frozen (as noted in the README), so the project is in a maintenance-only state. Despite its age and complexity, it contains several well-crafted, reusable utility libraries (e.g., LWXML streaming XML framework, SVM-to-SVG translator, and general-purpose Android utilities) that could be extracted for other projects.
This repository is the public release of the SLAM Handbook, a comprehensive book on Simultaneous Localization and Mapping (SLAM) to be published by Cambridge University Press. It contains the book manuscript (main.pdf), LaTeX style files for acronyms and bibliography strings, and a bug report template. The repository is primarily a document project rather than a software application, but the LaTeX infrastructure (SH_acronyms.sty and SH_string.bib) serves as a reusable library for typesetting SLAM-related documents. The book covers foundations, practical implementations, and spatial AI, with contributions from numerous experts. The repository is actively maintained with incremental releases and community feedback mechanisms via issues and discussions.
This is a PHP library that implements a personal WeChat bot using the web WeChat protocol. It allows users to log in via QR code, monitor incoming messages, and respond automatically through a plugin system. The architecture uses a Laravel-inspired service container with service providers, observers, and a custom extension mechanism. The library provides a CLI entry point and optionally runs as a Swoole-based server for REST API interaction. The code is well-structured but relies on outdated dependencies (Illuminate 5.4, Symfony 3, Guzzle 6.2, PHP 7.0+) and lacks automated tests. The web WeChat protocol is unstable and may be deprecated, limiting production use. The project appears to be maintained minimally, with no recent commits addressing dependency updates or modern PHP practices.
This is a leaked internal source of 'Claude Code', a terminal-based AI coding assistant CLI built by Anthropic. The application is a monolith written in TypeScript with React/Ink for terminal UIs, running on the Bun runtime. In production, it provides an interactive REPL that connects to Claude LLMs to perform software engineering tasks: reading/writing files, executing bash/PowerShell commands, searching code with grep/glob, managing git operations, spawning sub-agents, running MCP servers, and more. It features a sophisticated permission engine, plugin ecosystem, skill system, IDE bridge (VS Code/JetBrains), and remote session support. The architecture is centered around a QueryEngine that orchestrates LLM API calls with streaming, tool loops, and retries. There are ~40+ tools (BashTool, FileEditTool, MCPTool, etc.) and ~50+ slash commands (/commit, /review, /config, /doctor). The codebase is extremely large (~1900 files, 512K+ lines) with significant internal complexity. However, it contains many stubs for removed internal packages (@ant/*, @anthropic-ai/mcpb, etc.), heavy use of bun:bundle feature flags, and countless ant-only conditionals (process.env.USER_TYPE === 'ant') that gate internal-only features. This means the source cannot be run as-is without Anthropic's infrastructure. While the engineering quality is high in places (clean TypeScript with Zod schemas, well-structured tool pipelines, a custom Ink renderer with Yoga layout), the codebase is not designed for external reuse or portability. It requires Bun, has many hardcoded internal service URLs, and the stubbed packages would need real implementations. For learning and inspiration, however, it offers excellent examples of: terminal UI architecture, permission engines, plugin systems, and AI agent tool frameworks.
This is a curated 'awesome list' repository for the LangGraph and LangChain ecosystem, not a functional application or data pipeline. It aggregates links to frameworks, tools, projects, tutorials, and learning resources organized by category and domain. The repository is maintained with GitHub Actions for markdown linting, stale issue management, and automated repo health checks that verify community projects are not archived or stale. While the README is extensively detailed and well-structured, the actual codebase is limited to maintenance scripts (repo-checks.mjs) and configuration files. The system provides no runtime service or data processing—it serves solely as a static resource index. Its primary value is as a community-maintained discovery portal for LangGraph-related projects.
This repository is primarily a static personal website for author Li Xiaolai, hosting multiple original books and educational resources (e.g., 'Everyone Can Use English', Bitcoin whitepaper translation, TOEFL essay collection) as well as two interactive web applications: a Monte Carlo retirement projection tool and a historical ETF retirement strategy demo. The Monte Carlo tool is a well-crafted client-side financial simulator using Geometric Brownian Motion, Box-Muller transform, and Chart.js, complete with a standalone Node.js test suite. The site is built with docsify-style navigation and deployed via GitHub Actions to Alibaba Cloud OSS. While the content dominates the repository, the Monte Carlo application demonstrates solid engineering practices (modular JS, test coverage, performance-aware Float64Array usage) and is the most reusable software component. The repository is actively maintained (latest commit 2026-04-14, CI/CD active), but the surrounding static content has no build system or dependency management, and the Monte Carlo app relies on CDN-hosted Chart.js and a CORS proxy, posing medium technical drift risk.
GoBelieveIO/im_service is an instant messaging (IM) server infrastructure implemented in Go. It provides real-time peer-to-peer messaging, group chat (including super groups for large audiences), and chat rooms. The system is designed for clustered deployment and consists of three separate processes: `im` (access server handling client TCP/WebSocket connections), `ims` (storage server using append-only file logs with in-memory indexes), and `imr` (router server coordinating message delivery between access servers). Authentication is token-based (Redis or JWT). Message storage is per-user file-based with efficient offline message synchronization via sync keys. The codebase includes a custom binary protocol over TCP/WebSocket, a Redis-backed pub/sub mechanism for group events and relationship changes, and MySQL for group and friend data. This is a production-grade system (as indicated by its star count) with comprehensive benchmark tools and Python integration tests. However, the repository appears to be in a maintenance-stale state with no recent commits and some deprecated message types.
This repository is a curated collection of Fritzing part definitions for Adafruit electronic components, including sensors, displays, breakouts, power modules, and development boards. Each part is packaged as a .fzpz archive (ZIP) containing an XML metadata file (.fzp) and SVG graphics for breadboard, schematic, PCB, and icon views. The library is not a software application or data pipeline; it is a static design asset library for use with the Fritzing electronics design tool. It includes both current and retired parts, as well as community-contributed components. The repository's primary value is enabling rapid prototyping and documentation of circuits using Adafruit hardware. While well-organized, it does not contain executable code, and some parts may be outdated or superseded by newer hardware revisions.
Tieba-Cloud-Sign is a PHP web application designed to automate the sign-in process for Baidu Tieba forums. It provides a web interface for managing multiple Tieba accounts (via BDUSS cookies) and scheduling automatic sign-in tasks through a built-in cron system. The application features a plugin architecture that allows extending functionality, such as pushing sign-in results to QQ via go-cqhttp or backing up the database to QQ. The core system handles user authentication, Tieba API interactions (including retrieving TBS tokens and performing sign-ins), and provides an administrative interface for configuration. Plugins are integrated via hooks and can add their own settings, cron jobs, and UI components. Data is stored in a MySQL database, with tables for users, account credentials, plugin configurations, and operational logs. Key technical components include a custom WebSocket client for real-time communication with go-cqhttp (used by QQ-related plugins), a SQL dump utility for database backups, and an AJAX-driven settings interface. The frontend relies on Bootstrap 3.3.2 and jQuery for responsive UI elements, including modal dialogs for plugin configuration.
This repository is an educational Docker tutorial centered on a functional Django REST API application. The core implementation uses Docker Compose to orchestrate two services: a Django web container (DRF) and a PostgreSQL container, with explicit health checks, JSON-file logging limitations, named volumes, and user-defined networks. The Django API itself is a basic music CRUD service with endpoints under /api/musics, secured behind authentication, and housed in a straightforward project structure that demonstrates the typical pattern of a containerized Python microservice. Beyond the main tutorial, the repo provides stand-alone Docker Compose examples covering advanced patterns: YAML anchors to reduce configuration repetition, docker-compose override to batch operate multiple databases, environment variable substitution with .env files, profiles for selective service boot, and cAdvisor for container resource monitoring. Each subfolder includes a README and ready-to-run configuration, making the entire repository a one-stop teaching resource for Docker Compose patterns and best practices. The code is not intended as a production tool but rather as a self-contained, reproducible learning blueprint that shows how to build and manage containerized stacks with modern features. While the application layer is deliberately simple, the Docker-compose layer embodies several good practices around health and logging, and the advanced usage of YAML anchors and overrides exemplifies high-quality configuration craftsmanship that is directly reusable in other projects.