FTorch is a production-stable Fortran library that enables direct coupling of PyTorch machine learning models to Fortran scientific codes without requiring a Python runtime. It provides a comprehensive Fortran interface to LibTorch (PyTorch's C++ backend) for tensor operations, model inference, automatic differentiation, optimizers (SGD, Adam, AdamW), and loss functions (MSE, cross-entropy). The architecture uses a three-layer design: Fortran modules with derived types (torch_tensor, torch_model, torch_optim) that call a thin C API (ctorch.h/ctorch.cpp), which in turn wraps LibTorch. This layered approach isolates Fortran code from C++ ABI changes and allows the library to be built as a shared or static library with CMake, supporting CPU, CUDA, HIP, XPU, and MPS backends. The library includes extensive examples covering tensor manipulation, simple neural networks, ResNet-18 inference, batching, multi-I/O, looping, multi-GPU, MPI parallelism, autograd, and full training loops. Unit tests use pFUnit and integration tests are built from examples, with CI testing across multiple compilers (gfortran, ifx, flang) and platforms. A Python utility package (ftorch_utils) provides model conversion (PyTorch to TorchScript) and validation. FTorch is published in the Journal of Open Source Software (JOSS) and is used in several HPC climate and fusion simulation codes (E3SM, CAM, ICON, UKAEA). The codebase demonstrates strong engineering practices: fypp code generation reduces boilerplate for multiple tensor ranks and data types, C interoperability is handled cleanly via iso_c_binding, and finalizers ensure automatic resource cleanup. However, there are some maintenance risks including reliance on fypp preprocessing, a large CI matrix, and the need to track LibTorch releases. A legacy API (torch_tensor_from_array_legacy) is still present but deprecated. Overall, FTorch is a well-crafted, domain-specific library that fills a critical gap for Fortran-HPC users needing ML integration.
This repository implements a modern Automatic Number Plate Recognition (ANPR) pipeline as a production-ready application. It ingests images (via HTTP upload or WebSocket stream), runs detection using the `fast-alpr` library (YOLO-based ONNX model), performs OCR with `fast-plate-ocr`, applies post-processing (confusion-pair correction, Turkish plate format parsing, temporal majority voting), optionally tracks vehicles via an IoU-based multi-object tracker, and stores only HMAC-SHA256 hashed plate texts for privacy compliance. The system is exposed through a FastAPI REST API with WebSocket support and a HTMX web UI, plus a Typer CLI for inference and server management. It is containerized with Docker and publishes to PyPI and GHCR. The repository also contains a legacy directory (`legacy/`) with a complete YOLOv5 training pipeline for license plate detection, but that code is explicitly unmaintained and serves as historical reference only. The modern codebase is actively maintained, well-tested, and demonstrates strong software engineering practices including protocol-based abstractions, structured logging, Prometheus metrics, and privacy-by-design persistence.
GeoWombat is an open-source Python library for geospatial raster data processing and remote sensing analysis at scale. It provides a high-level API built on xarray, rasterio, Dask, and geopandas, enabling lazy, parallel processing of multi-spectral satellite imagery (Landsat, Sentinel-2, NAIP, etc.). The library's architecture centers on a custom xarray accessor (`.gw`) that exposes geospatial properties and methods, and a configuration context manager for setting sensor profiles, reprojection parameters, and other global settings. Key capabilities include raster I/O with on-the-fly warping/mosaicking, vegetation indices, radiometric corrections (BRDF, 6S, topographic), QA masking, time-series analysis, moving-window statistics (Cython-accelerated), and machine/deep learning classification and object detection (YOLO, TorchGeo, SAM). It supports STAC cloud access and distributed computation via Dask. The codebase demonstrates solid engineering with modular backends, extensible patterns, and comprehensive test coverage, though it carries some technical debt from pinned dependencies and large monolithic modules. It is actively maintained, with a broad user community and integrations with wider geospatial ecosystems.
vinta/awesome-python is a curated list of Python frameworks, libraries, and tools, maintained as a single markdown file (README.md) and rendered into a static website via a Python build system. The repository's core function is to generate an opinionated, searchable, and sortable web catalog of ~650 Python projects, enriched with live GitHub star counts and PyPI download statistics. The build pipeline (build.py) parses the markdown using an AST-based parser (readme_parser.py), fetches metrics via dedicated scripts (fetch_github_stars.py, fetch_pypi_downloads_via_clickpy.py), and outputs Jinja2 templates into static HTML/CSS/JS. A GitHub Actions workflow runs daily to refresh data and deploy to GitHub Pages. While the primary output is a static website, the build toolchain and data-fetching scripts are reusable for other awesome-list-style projects. The repository also contains AI-assisted maintenance workflows (Claude skills) for reviewing PRs and auditing the list, reflecting a hybrid of curated content and automation.
TrendRadar is a configurable news aggregation and trend analysis system that crawls hot lists from 11+ Chinese platforms (Zhihu, Weibo, Douyin, Bilibili, etc.) and RSS/Atom feeds, filters content via keyword DSL or AI-based interest matching, performs AI-powered analysis (sentiment, controversy, topic trends), generates rich HTML reports, and pushes notifications to 9+ channels (Feishu, DingTalk, Telegram, Email, ntfy, Bark, Slack, WeCom, generic webhook). It is primarily an application (with CLI entry points, scheduling via timeline.yaml, Docker/GitHub Actions deployment) but incorporates data-pipeline patterns such as multi-stage ETL (crawl → store → filter → analyze → report), SQLite storage with optional S3 remote sync, and AI batch processing. The architecture is a modular monolith: the core `trendradar` package handles all orchestration, while a separate `mcp_server` package exposes 20+ tools via Model Context Protocol (MCP) for LLM integration. The system is heavily config-driven (YAML and text files) and includes a standalone web-based config editor. External dependencies include the third-party newsnow API (for hot lists), LiteLLM (for AI model abstraction), and various notification service APIs. The codebase demonstrates good craftsmanship with clean abstractions (storage backends, notification dispatcher, AI client), though some duplication exists between the core and MCP notification logic, and the extensive configuration surface requires careful maintenance.
This repository is a curated collection of 177+ domain-specific scientific 'skills' for AI agents, following the Agent Skills standard. Each skill is a self-contained module providing a CLI tool, documentation, and tests for a specific scientific or technical task—ranging from single-cell RNA-seq analysis (scanpy) and molecular docking (diffdock) to clinical report validation and ISO standards readiness. The architecture is a monorepo of independent skills, each with its own scripts/, references/, assets/, and tests/. Skills are designed to be safe, bounded, and reproducible: they validate inputs before execution, enforce resource limits, avoid network access unless explicitly required, and produce structured JSON reports. The repository also includes a shared test infrastructure (skill_contract) and a meta-skill (autoskill) that can automatically propose new skills from screen activity. While primarily an application library for AI agents, it has strong data/research characteristics due to its heavy use of scientific data formats and validation of research artifacts.
CloakBrowser is a commercial stealth browser automation library that wraps Playwright (Python, JavaScript, .NET) with a custom patched Chromium binary. Its core purpose is to evade bot detection by spoofing browser fingerprints, providing human-like interaction (Bezier mouse moves, realistic typing with typos, smooth scrolling), and managing licensed binary downloads with cryptographic verification. The architecture is a multi-language SDK sharing identical logic across Python, TypeScript, and C#, with a common human interaction engine, proxy/GeoIP integration, and license enforcement via denial files and server validation. Key entry points are the `launch()` and `humanize()` APIs, CLI diagnostics (`info`, `login`, `install`), and the `cloakserve` CDP multiplexer. External dependencies include Playwright/Puppeteer, MaxMind GeoLite2 database, Ed25519 signature verification (BouncyCastle), and a license validation server. The codebase demonstrates high engineering discipline with extensive cross-language parity tests and security-conscious binary distribution, but carries drift risks from triple-language duplication and reliance on Playwright internal APIs.
Dolphin is a document image parsing tool that uses a two-stage analyze-then-parse approach with a VLM (Qwen2.5-VL). It takes images or PDFs, performs layout analysis to detect elements (text, tables, formulas, code, figures) with reading order, then parses each element in parallel batches. Outputs structured JSON and Markdown with visualized layout. The codebase is well-structured, offering CLI entry points for page-level, element-level, and layout-only parsing. It demonstrates good engineering practices with batching, coordinate mapping, and post-processing for formula/special characters. The core implementation wraps the Hugging Face Qwen2.5-VL model and includes utility modules for layout string parsing, PDF conversion, and markdown generation. The architecture is a CLI application with minimal coupling, making the utility modules potentially reusable across document parsing projects.
google-images-download is a Python application for searching and downloading images from Google Images. It functions as both a command-line tool and a library that can be imported into other Python scripts. The system uses Selenium with Chrome to handle JavaScript-heavy Google Images pages, automatically manages chromedriver via webdriver-manager, and supports extensive filtering options (size, color, format, usage rights, etc.). Key features include filename sanitization for cross-platform compatibility, metadata extraction, similar image search, and robust error handling for network issues. The architecture is monolithic with a clear separation between argument parsing, URL building, browser automation, and download execution.
Helium is a Python library that provides a higher-level, more intuitive API for browser automation, built on top of Selenium. It simplifies common web automation tasks such as clicking buttons, filling forms, handling alerts, switching windows, and dragging elements by allowing developers to reference page elements by their visible labels rather than low-level CSS selectors or XPath expressions. Under the hood, Helium translates each high-level command into Selenium calls, managing implicit waits, iframe traversal, and window focus automatically. The library is designed for both interactive use (e.g., testing) and scripting (e.g., web scraping), and it seamlessly integrates with raw Selenium for advanced use cases. The architecture follows a layered pattern: a public API module (`helium/__init__.py`) exposes functions and GUI element classes; an implementation layer (`helium/_impl/__init__.py`) contains the core logic including element location via XPath construction, frame iteration, and error handling; and utility modules provide geometry, string, and system helpers. The codebase is well-organized with clear separation between API and implementation, though the main implementation file is large and tightly coupled. The library is actively used (over 8,300 stars) and though maintenance is low-effort, it remains functional with recent Selenium compatibility.
pytorch-forecasting is a deep learning library for time series forecasting, providing implementations of state-of-the-art models such as Temporal Fusion Transformer, N-BEATS, N-HiTS, DeepAR, PatchTST, TiDE, TimeXer, and many others. Built on PyTorch and PyTorch Lightning, it offers a complete pipeline from data preprocessing (TimeSeriesDataSet, encoders, samplers) through model training (with GPU support, callbacks, hyperparameter tuning via optuna) to prediction and evaluation (point, quantile, and distributional metrics). The library supports multivariate forecasting, static and time-varying covariates, categorical features, and multiple targets. A major v2 refactoring is underway, introducing a new data module layer and simplified package interface, though the v2 API is explicitly marked as experimental and unstable. The codebase is well-tested with automated integration tests across all models and metrics, and employs a composition pattern for metrics (supporting weighting, aggregation, and distribution losses) and a layered architecture for models (base classes, package containers, building blocks).
Flask-AppBuilder is a rapid application development framework built on top of Flask. It provides automatic CRUD generation from SQLAlchemy models, a comprehensive role-based security system with multiple authentication backends (database, LDAP, OAuth, SAML, OpenID, remote user), REST API generation with OpenAPI/Swagger, chart views, i18n, and a widget-based UI system. The architecture follows a Flask extension pattern with an AppBuilder class that registers views, menus, and permissions. It includes a security manager that handles authentication, authorization, and user management. The framework is used by major projects like Apache Superset and Apache Airflow. Key components: The security manager provides a pluggable authentication system with role-based access control, permission discovery, and API key management. The model views (ModelView, ModelRestApi) automatically generate CRUD interfaces and REST endpoints from SQLAlchemy models with filtering, sorting, and pagination. The framework includes a widget system for customizing list/show/form rendering, and a chart view system for Google Charts integration. The codebase is well-structured with clear separation of concerns, but shows some technical drift: outdated Bootstrap 3 UI, vendored JavaScript libraries, deprecated chart classes, and old translation files. The core security and CRUD functionality is actively maintained and production-ready.
pycdc/pycdas is a C++11 command-line toolchain for reverse engineering CPython bytecode. pycdc decompiles compiled .pyc files back into human-readable Python source; pycdas emits a detailed bytecode disassembly (with options like --show-caches and --pycode-extra). Version coverage spans Python 1.0 (1994) through 3.13, with the correct opcode semantics selected from the .pyc magic number at load time. The entry points are pycdc.cpp and pycdas.cpp, which are thin CLIs over the pycxx static library that owns marshal-format parsing (pyc_object.cpp, pyc_numeric.cpp, pyc_string.cpp, pyc_sequence.cpp), code-object extraction (pyc_code.cpp), and interpreter version detection (pyc_module.cpp). It also supports decompiling raw marshalled code objects via -c -v <version>. Decompilation loads the code object and interprets its bytecode on a virtual stack machine (FastStack with history snapshots) inside ASTree.cpp, reconstructing structured control flow (loops, try/except, comprehensions, with-blocks) and lowering it into an ASTNode tree for source emission. Opcode semantics are maintained as an X-macro include (bytecode_ops.inl) with explicit version ranges, and per-version opcode-name maps live in bytes/python_*_*.cpp. The module boundaries cleanly separate binary I/O (endian-aware PycBuffer/PycFile streams), object marshalling, disassembly, AST construction, and printing - although the decompiler core itself is a dense, version-conditional megafile. QA is the project's strongest asset. Roughly 200 .pyc fixtures spanning every Python major release (plus an xfail set for known-unfixed cases) are checked by tests/run_tests.py, which compares decompiler output to expected output as token streams rather than exact text - tolerating cosmetic whitespace differences while still catching semantic regressions. Linux and Windows CI run the suite, and scripts/pymultic, Dockerfile.pybuild, and a collection of Autoconf modernization patches make it possible to rebuild ancient CPython releases to regenerate those fixtures. Recent commits cover Python 3.13 opcode support, and the repository is actively maintained. Caveats: the 25-year version matrix forces heavily conditional decompilation logic, some mutable global state exists in ASTree, and unknown or future-modified opcodes degrade to partial output with an 'Unsupported opcode' warning. For its niche, it is the most complete open-source CPython decompiler available.
This repository implements a Model Context Protocol (MCP) server and CLI for searching, downloading, and reading academic papers across ~20 scholarly platforms (arXiv, PubMed, Semantic Scholar, CrossRef, OpenAlex, Google Scholar, etc.). It is primarily an application—a Python package distributed via PyPI—that exposes two entry points: an MCP server (stdio-based, compatible with Claude Desktop and other MCP clients) and a command-line interface. The architecture follows a clean two-layer design: a unified tool layer (search_papers, download_with_fallback) that orchestrates concurrent multi-source searches with deduplication, and a modular connector layer where each platform is implemented as a self-contained class inheriting from a common PaperSource abstract base. The system emphasizes a 'free-first' strategy, relying on open public APIs and optional API keys for enhanced access. It includes a fallback chain for PDF downloads (native source → OA repositories → Unpaywall → Sci-Hub) and supports full-text extraction via pypdf. The codebase is well-structured with consistent patterns (Paper dataclass, env config, retry/backoff), comprehensive test coverage (unit and e2e), and CI/CD publishing. However, some connectors are skeletons (IEEE, ACM) or not wired (ChemRxiv), and HTML scraping components are fragile to upstream changes. The last commit date is anomalously in the future (2026-08-17), but the repository shows active issue triage and community engagement.
This repository implements an AI-agent skill that converts images, PDFs, and image-based PPT/PPTX files into fully editable PowerPoint (.pptx) documents. The system uses a deterministic CLI state machine (`editppt`) with a four-phase workflow: normalize input into per-page jobs, dispatch to subagents for page reconstruction, record page manifests, and finalize by rebuilding the PPTX from manifests. It integrates OCR (PaddleOCR or offline detector) for text bounding boxes, image generation/editing backends (built-in, OpenAI-compatible, or Codex OAuth), and LaTeX rendering for formulas. The output preserves native text boxes, shapes (including custom paths and polygons), tables, images, and speaker notes, with strict provenance tracking for all assets. The codebase is well-structured with clear separation of concerns, a manifest-driven assembly pattern, and comprehensive tests. However, the last commit date is anomalously in the future (2026-09-16), suggesting the repository may be abandoned or the timestamp is erroneous; combined with only 1 open issue and no recent activity, it is classified as stale. The technical drift is low as the code uses modern Python and standard libraries, but external dependencies (OCR, image APIs, LaTeX) may become incompatible over time.
OWTF (Offensive Web Testing Framework) is a full-stack web application for automated and manual penetration testing. The Python/Tornado backend exposes a JWT-protected REST API for managing targets, sessions, plugin worklists, workers, reports, and proxy history, while the React/Redux frontend provides dashboards, data tables, plugin marketplace, and proxy controls. The application also includes a customizable intercepting HTTP/HTTPS proxy with SSL certificate generation, an interceptor chain for header/body/URL/delay modification, transaction history, a repeater, and live interception - comparable in scope to a Burp Suite-style tool. Execution is plugin-driven: built-in Python plugins covering web, network, and auxiliary tests, plus wrappers around external tools like Nikto and Arachni, are run by multiprocessing workers. Community plugins can be uploaded after AST-based static validation that blocks dangerous imports and execution patterns. The system stores plugin outputs, deduplicates them via SHA-256 fingerprints, tracks execution metrics, and supports HTML reporting. User authentication includes JWT, email verification, OTP, and admin role management. Architecturally, this is a monolith with modular managers and a REST API, not microservices. Deployment targets Docker Compose, Kubernetes, and cloud Terraform. The codebase contains genuinely reusable components: AST-based plugin validation, cross-platform file locking, output fingerprint deduplication, structured logging, and SQLAlchemy query helpers. However, it also carries significant legacy debt: old frontend patterns (React 16 class components, Redux 3, Redux-Saga 0.16), Python 2 remnants in scripts, hardcoded secrets in compose files, and several stub or placeholder external plugins. The repository is fundamentally an application rather than a data research project; bundled dictionaries and execution metrics serve the security workflow but do not define its primary character.
django-wiki is a full-featured wiki system built as a reusable Django application. It provides hierarchical wiki pages with URL slug paths, revision control, attachments, images, notifications, and granular permissions. The architecture is modular, with a plugin system that allows extensions like macros, image galleries, and notification digests. Content is authored in Markdown with custom extensions for wiki links, tables of contents, and code highlighting. The system integrates with Django's authentication, sites framework, and admin interface. It is designed for production use with caching, spam protection, and i18n support. The codebase is well-structured with clear separation of concerns, though it carries some technical debt from vendored frontend assets and deprecated patterns.
python-o365 is a mature Python library that provides a comprehensive, object-oriented wrapper around the Microsoft Graph API. It enables developers to programmatically access and manage Microsoft 365 services including Outlook email and calendar, OneDrive and SharePoint files, Teams chats and presence, Planner tasks, directory users and groups, and more. The library abstracts OAuth2 authentication (supporting authorization code, public client, and client credentials flows) and handles token refresh, pagination, and API versioning internally. It is designed as a client-side SDK, not a server application, and can be embedded into scripts, web applications (e.g., Flask), or automation workflows.
SpiffWorkflow is a Python library for executing BPMN 2.0 and DMN 1.x workflows. It provides a complete workflow engine with parsing, execution, serialization, and migration capabilities. The library supports standard BPMN constructs (gateways, events, subprocesses, multi-instance loops, call activities) and DMN decision tables with UNIQUE and COLLECT hit policies. It includes a pluggable Python script engine for evaluating expressions and scripts within tasks, and a comprehensive serialization framework with version migration for workflow persistence. The architecture is modular, with separate packages for core workflow specs, BPMN specifics, DMN integration, Camunda-compatible extensions, and Spiff-specific customizations. This is not a standalone application but a library intended for embedding in larger systems; it is used as the backend for the SpiffWorkflow frontend tools.
Shed Skin is a Python-to-C++ transpiler that translates a restricted, implicitly statically typed subset of Python 3 into optimized C++. It employs type inference to resolve types and automatically generates C++ code, which is then compiled into native executables or shared library extensions. The architecture consists of a frontend (parser, type inference, virtual analysis), a C++ code generator, and a comprehensive runtime library (implementing Python built-in types, standard library modules, and bundling Boost headers for portability). The project includes over 80 example programs covering games, simulations, compression, AI, and graphics, serving as both showcases and benchmarks. A pytest unit test suite and a CMake/ctest integration test suite validate compiler correctness. The tool is primarily a command-line application invoked via `shedskin build <script>`, producing binaries that typically achieve 1-100x speedups over CPython.