What challenges arise when you want to use decorators on both synchronous and asynchronous functions? The article introduces how to create decorators that work with both sync and async functions using generators. This article is from Patreon's tech blog.
TL;DR
Agent-S teaches agents to use computers like humans while LangGraph backs a coding agent, PyTorch's internals get opened up, and the lihil async framework chases extreme speed.
Articles & Tutorials
× 11The author shares the process of completely refactoring a small Python project to serverless architecture, moving from VPS to AWS Lambda, reducing costs from $350 per year to $2 per month.
The article introduces implementations of common load balancing algorithms with Python code examples, including random selection, round-robin, least connections, and consistent hashing algorithms.
Python has a famous xkcd comic easter egg: import antigravity. Beyond that, the author accidentally discovered another xkcd easter egg about "unreachable state": when the impossible happens, anything becomes possible...
An author with eight years of Python experience vents about two major pain points: types and exceptions.
The article shares Qodo's reasons and process for choosing the LangGraph framework to develop their AI coding assistant, introducing LangGraph's state machine model, clean API design, node-based architecture, and built-in state management features.
Introduces a Python tool called Unvibe that can generate correct code implementations through unit tests. Its principle is using unit tests as reward functions, calling large models through Monte Carlo tree search algorithms to generate code until all tests pass.
Shares how to optimize a Python program through garbage collection, reducing runtime from 40 seconds to 10 seconds and memory usage from nearly 60 GB to 1 GB. Additionally, using CuPy instead of NumPy reduced runtime to 1.5 seconds.
A very long article that deeply introduces PyTorch's internal mechanisms, including tensor data structures, automatic differentiation mechanisms, memory management, and operation implementations.
The third part of the tech article sharing series by the author of "Python Craftsman", focusing on programmer fundamentals, software engineering, backend development, technical writing, Python, and Go languages.
The author shares years of thinking about programming technology and workplace life experiences. "A programmer's competitiveness isn't writing code, nor any particular language or framework - the core competitiveness is the ability to solve problems through technology."
Projects & Resources
× 12A modular and scalable computer-use framework designed to achieve autonomous interaction with computers and automatically execute complex tasks. (1.4K stars)
A fork of pyright with type checking improvements, enhanced vscode support, and pylance features built into the language server. (1.7K stars)
Open source RAG system supporting multiple data sources including web pages, PDFs, videos, and GitHub repositories, for quickly developing website widgets, Slack bots, and Discord bots.
Manipulate compressed JSON files, supporting gzip, bz2, and lzma (or xz) formats.
Pure Python web framework with features including dependency injection, OpenAPI docs & error response generator, problem pages, data validation & parameter parsing, built-in messaging system, and powerful AI support.
Automatically discovers all FastAPI endpoints and converts them to MCP tools, directly mounting MCP servers to FastAPI applications.
Develop modular command-line applications using a declarative, class-based approach.
Enables MCP interaction with Obsidian through Obsidian's REST API community plugin, including listing directories and files, reading file content, searching files, and editing notes.
Utilizes LLMs from Ollama or LMStudio to search the web, summarize, and reflect, ultimately generating Markdown summaries. (6.7K stars)
Batch download Douyin content with watermark removal, supporting videos, image sets, collections, and music (original audio). Supports command-line parameters or YAML configuration file operation. (2.3K stars)
Generate and integrate OpenAPI documentation using Pydantic v2 with various frameworks (Falcon, Flask, Quart, Sanic, Starlette, Tornado).
A multimodal vision-language model for Scalable Vector Graphics (SVG) generation, capable of performing image-to-SVG and text-to-SVG generation. (2.7K stars)