Load Testing APIs with wrk (Detailed Guide) This article explains how to use wrk on Ubuntu to stress-test backend APIs (Flask, FastAPI, Spring Boot, etc.) and interpret the results. 1. What is wrk? wrk is a modern, high-performance HTTP benchmarking tool written in C. Key features: High concurrency: thousands of concurrent connections Multi-threaded: uses multiple CPU cores Lua scripting: for custom headers, bodies, tokens Faster than Apache Benchmark (ab): lighter and more stable 2. Install wrk On Ubuntu/Debian: ...

Access a Git Bare Repo on Windows WSL2 from the LAN In development, you often need to share Git repositories across multiple machines. If you use WSL2 on Windows and want other LAN machines to access a Git bare repo inside WSL2, this guide walks you through the setup. 1. Create a Git bare repo in WSL2 In WSL2, go to the target directory: git init --bare my_project.git my_project.git is a bare repo with no working tree, only Git data. A bare repo behaves like a remote and can be cloned and pushed. 2. Enable SSH in WSL2 Other machines will access via SSH. ...

Simplified Git Branch Workflow (Solo / Small Team) This workflow is a simplified version of Git Flow. It is suitable for personal projects or small teams: structured but not heavy. 1. Main branch (long-lived) main Always stable and release-ready Production deployments come from here For small teams, main is usually enough; no need for develop. 2. Feature development (feature branch) Naming: feature/<feature-name> Purpose: build new features, then merge back to main Examples: ...

Use a Local Git Bare Repo to Separate Dev and Test Environments In full-stack work, a common problem is how to isolate dev and test environments. Many people host on GitHub or GitLab, but private projects may not be suitable for public hosting. Git is distributed. You can set up a local bare repo as a remote to move code from dev -> test in one machine. What is a bare repository? A normal repo (git init) has a working tree + .git metadata and can be edited directly. A bare repo (git init --bare) has only Git data, no working tree. It is usually used as a remote. In short: ...

Introduction For TypeScript files with the .ts extension, we cannot run them directly. We need to transpile TypeScript to JavaScript and then run the JavaScript output. There are two common approaches: upload .ts to the server and compile via CI, or transpile locally and upload the .js build to production. If you want to run and test locally during development, you can use ts-node, but the project still needs a build step for production.

Introduction I want to build an AI system that supports tree-shaped or graph-shaped Q&A, instead of a traditional single-thread chat flow. Exploration Open-source framework research flowise

Introduction Mermaid is a framework for creating diagrams using code. This post shows how to install the tooling on your server and render Mermaid code into images. Steps Install the renderer Run: npm install -g @mermaid-js/mermaid-cli Note: the CLI requires npm version >= 20. It is recommended to manage npm versions with nvm. If you do not have nvm, install it with: curl -o https://raw.githubusercontent.com/nvm-sh/nvim/v0.39.4/install.sh | bash Restart your shell, then run: ...

How to Truly Master a Paper Conclusion To truly master a paper, reading once is not enough. You need to decompose, verify, and reconstruct it, and then express the key points in your own words or implementation. The goal: explain the core contribution in 5 minutes, derive key formulas by hand, and reproduce a core experiment. Principles and background A paper is a compressed expression of a problem. It omits background, intuition, failed attempts, and many details. Mastery requires “decompressing” that information into your own knowledge network: assumptions, derivations, engineering steps, and the limits of the results. Only then can you judge when to use it and when not to. ...

What problem does this paper solve, and what are the results? We know AI systems are expanding and can solve general tasks. But many AI agent applications today target small tasks. NVIDIA argues that small language models (SLMs) are capable, more suitable, and cheaper, and should be a main direction for future agents. The paper discusses: What tasks current SLMs can handle Where general language ability matters The limits of SLMs as agents Conclusion: moving from LLMs to SLMs has advantages in both capability and cost.