标签：#r

GitHub热门项目 | 闲鱼自动回复管理系统是一个基于 Python + FastAPI 开发的自动化客服系统，专为闲鱼平台设计。系统通过 WebSocket 连接闲鱼服务器，实时接收和处理消息，提供智能化的自动回复服务。同时集成闲鱼自动发货，自动评价，自动擦亮等功能，实现闲鱼虚拟商品自动化流程。 | Stars: 5,759 | 42 stars today | 语言: Python

GitHub热门项目 | Biomni: a general-purpose biomedical AI agent | Stars: 3,461 | 32 stars today | 语言: Python

GitHub热门项目 | Lightweight, open-source AI agent for your tools, chats, and workflows. | Stars: 45,639 | 120 stars today | 语言: Python

GitHub热门项目 | DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/. | Stars: 26,041 | 128 stars today | 语言: Python

Spotify is expanding free access to managed accounts for under-13s.

XGIMI's Eflin Flip 4K is a portable, stylish projector that provides outstanding image quality at a great price.

At its peak, around 12 million people subscribed to World of Warcraft so that they could explore the realm of Azeroth together. The audience for PointlessQuest is quite a bit smaller. On launch day, the game hit a peak of 15 concurrent players… and no, that sentence isn't missing a word. Then again, basically everything […]

The app is designed for people who want to create social content, but find traditional video editing tools too complex or time-consuming.

Rime is handling over 100 million calls each month across multiple companies

Gwen Shapira shares how teams are scaling AI features using PostgreSQL for mission-critical apps. She explains how to leverage Postgres's multi-modal capabilities - including JSONB parsing and high-recall HNSW vector indexing - to deliver deterministic and semantic context to LLMs. She also discusses vector quantization to speed up queries by 4x and strategies for managing agentic memory. By Gwen Shapira

How an unexpected regional constraint forced us to deeply understand Azure GPU VM families, naming conventions, and workload fit. Introduction As architects, we often assume that infrastructure decisions are straightforward: "The workload is already running successfully in Region A. Let's deploy the same Kubernetes workload in Region B." That's exactly what we thought. Our workload consisted of a Visual Element Detection (VED) service hosted on Kubernetes. The application uses a PyTorch model to analyze images and detect various visual elements in an image file. The service was already running successfully on a node pool backed by Azure's NVads_A10_v5 GPU VMs. Then we hit an unexpected challenge. The target region did not offer NVads_A10_v5 instances. What looked like a simple deployment exercise became a deep dive into Azure GPU virtual machine families, GPU architectures, VM naming conventions, and workload characteristics. This article shares what I learned in the hope that it helps others who find themselves evaluating Azure GPU SKUs for AI inference workloads. I am relatively new to the world of MLOps, Model deployments, GPU Workloads etc and equally interested and excited to learn more on this front. The Workload Before discussing VM selection, let's understand the workload characteristics: Model Type : PyTorch Model Size : less than 200 MB (.pth) Image Resolution : ~2000 x 2000 Expected Throughput : 5-7 requests/sec Platform : AKS (Kubernetes) Workload Type : Inference only This is important because GPU sizing should always start from the workload and not from the VM catalog. Step 1: Understanding Azure GPU VM Families Many engineers first encounter Azure GPU machines through names like: NV12s_v3 NV6ads_A10_v5 NC4as_T4_v3 ND96isr_H100_v5 The naming can be intimidating. The first breakthrough was understanding that Azure organizes GPU VMs into three primary families: N-Series ├── NV ├── NC └── ND NV Series – Visualization and Graphics NV-series VMs are designe

A hospital-network client wanted our system to output patient data in actual FHIR format - the standard interoperability format healthcare systems use to talk to each other - instead of whatever shape we felt like inventing. Made total sense from their side, their EHR software only accepts FHIR resources, not our custom JSON. From my side, it meant I now had to get an LLM to produce a Patient resource that was FHIR R4 compliant, field for field. I opened the FHIR R4 spec page for Patient to see what I was dealing with. Closed the tab about four minutes later. It's not one flat object - names have their own nested structure with use / family / given arrays, telecom is a list of typed contact points, addresses have their own multi-field shape, and half the fields have specific allowed value sets straight out of a separate FHIR terminology spec. This was not going to be a quick z.object({...}) . Two days into hand-writing it, and I wasn't even done I started anyway, because what else was I going to do: const PatientSchema = z . object ({ resourceType : z . literal ( " Patient " ), identifier : z . array ( z . object ({ system : z . string (), value : z . string (), }) ), name : z . array ( z . object ({ use : z . enum ([ " official " , " usual " , " nickname " , " maiden " ]), family : z . string (), given : z . array ( z . string ()), }) ), telecom : z . array ( z . object ({ system : z . enum ([ " phone " , " email " , " fax " ]), value : z . string (), use : z . enum ([ " home " , " work " , " mobile " ]). optional (), }) ), gender : z . enum ([ " male " , " female " , " other " , " unknown " ]), birthDate : z . string (), address : z . array ( z . object ({ use : z . enum ([ " home " , " work " , " temp " ]). optional (), line : z . array ( z . string ()), city : z . string (), state : z . string (), postalCode : z . string (), country : z . string (), }) ), // ...and I still hadn't gotten to maritalStatus, communication, // contact, generalPractitioner, managingOr

If you’ve ever built or interacted with DeFi protocols, you know the mathematical limitations of Constant Product Market Makers ($x \times y = k$). While AMMs are great for retail liquidity, executing a large transaction (e.g., $100,000+) directly against a liquidity pool triggers two major issues: Severe Slippage: The marginal price of the asset degrades exponentially relative to the trade size. MEV Exploitation (Sandwich Attacks): Public mempool transactions are highly vulnerable. Front-running bots will buy the asset ahead of your execution block, push the price up to your maximum slippage limit, and dump it immediately after. To solve this without relying on centralized, custodial desks or risky off-chain escrow setups, we built Typelex —a decentralized, non-custodial P2P OTC protocol. Here is a look at how we bypassed the AMM bonding curve entirely using on-chain atomic swaps. The Architecture of an On-Chain Atomic Swap Instead of routing trades through active liquidity pools, Typelex utilizes isolated smart contracts to execute peer-to-peer trades. The entire swap happens atomically : either all conditions are met within a single block execution, or the entire transaction reverts. Conceptual Smart Contract Logic (Solidity-based) To understand how the trustless escrow works under the hood, here is a simplified mental model of the swap execution logic: struct Order { address maker; address taker; // address(0) if public address tokenA; uint256 amountA; address tokenB; uint256 amountB; bool active; } mapping(uint256 => Order) public orders; function takeOrder(uint256 orderId) external { Order storage order = orders[orderId]; require(order.active, "Order not active"); if (order.taker != address(0)) { require(msg.sender == order.taker, "Unauthorized taker"); } order.active = false; // Pull Token B from Taker to Maker IERC20(order.tokenB).transferFrom(msg.sender, order.maker, order.amountB); // Push Token A from Contract Escrow to Taker IERC20(order.tokenA).transfer

I built a small MCP server a while back — developer-presence , seven tools wrapping the GitHub REST API and the DEV.to API so an agent can check my repo stats, list my articles, or draft a new post without me leaving the chat. It's mine, I wrote every line, there's no third-party package doing anything sketchy under the hood. By the usual "vet your MCP servers before installing them" checklist, it passes clean. I've written that checklist article before. What I hadn't thought carefully about until recently is that vetting the server doesn't vet the data. Two of its tools go straight to the point: @mcp.tool () def get_repo_stats ( repo : str ) -> dict : """ Get stars, forks, watchers, open issues for enjoykumawat/<repo>. """ r = _gh ( f " /repos/ { GITHUB_USERNAME } / { repo } " ) return { " name " : r [ " name " ], " stars " : r [ " stargazers_count " ], " forks " : r [ " forks_count " ], " watchers " : r [ " watchers_count " ], " open_issues " : r [ " open_issues_count " ], " language " : r . get ( " language " ), " description " : r . get ( " description " ), } description is free text. Any repo owner can put anything in it. If I ever point this tool at a repo I don't control — someone else's fork, a dependency, anything — that field lands in my agent's context exactly the same way a trusted instruction would: as text in a tool result, with no marker distinguishing "this came from GitHub's database, unfiltered" from "this is something I told the agent to do." The server is safe. The channel is safe. The payload was never vetted at all, because there was nothing to vet — it's just whatever a stranger typed into a form. I only really felt this because of a task I run on a schedule: check dev.to for trending posts in a few tags, score them, and use the highest scorers as source material for what to write about next. Step one of that job is a loop over tag pages: for tag in [ " ai " , " llm " , " mcp " , " claudecode " , " agents " , " productivity " ]: url = f " http

Database branching is one of the best ideas serverless Postgres brought to the mainstream. Fork the database at a point in time, get an isolated copy with all the data, run something risky against it, throw it away. It made preview databases and safe migrations feel routine. But a real application is not just a database. It is a database, plus the files it stores in object storage, plus the backend code that serves it, plus, increasingly, the model and gateway config it calls for AI. When you branch only the database, those other three stay shared. Your "branch" points at the same S3 bucket, the same deployed backend, and the same AI configuration as everything else. So it is half a copy, and the half it leaves out is where a lot of the interesting bugs and the scary migrations live. Neon's platform preview changes what a branch contains. A branch now forks the database and its data, the object storage and its files, the functions that run your backend, and the AI gateway config, all at the same point in time, all isolated. A branch stops being a database copy and becomes a whole environment. To make sure that is a real claim and not a diagram, I took a full-stack project, branched it, and checked every layer. Here is what happened. TL;DR Elsewhere, "branch" means the database only. Object storage, backend deploys, and AI config stay shared, so you bolt on scripts to fake per-branch versions of them. A Neon branch forks all four together: Postgres + data, object storage + files, functions (each branch gets its own URL), and the AI gateway. I proved it: branched a project with a DB, a bucket of files, a function, and the gateway. The branch came up with a copy of the rows, a copy of the files on its own storage endpoint, its own function URL, and the gateway. A write to the branch left main untouched, and deleting the branch removed all of it. That makes a branch a real environment: true preview stacks, whole-state bug reproduction, and disposable sandboxes for agent

Hey everyone, I recently had an "aha!" moment regarding how React handles updates under the hood, and I wanted to share it because I realize a ton of developers (including myself, until recently) trip over this exact concept. The common mental model is that React Reconciliation compares the Virtual DOM directly to the Real Browser DOM and surgically updates only what changed. But that’s fundamentally incorrect. React never reads or directly compares the real DOM during the diffing process. It actually splits the process into two entirely separate phases —The Render Phase and The Commit Phase —which creates a massive distinction between Re-rendering and Re-painting. Here is the exact breakdown of what happens when a single state change affects just 1 out of 100 divs in a component: The Render Phase (Pure JavaScript) When state changes, React calls your component function. It doesn't know which of your 100 divs changed yet, so it has to evaluate the entire JSX block. The Scope: React re-renders all 100 virtual divs in memory. The Process: It builds a brand-new Virtual DOM tree and compares it to the previous Virtual DOM tree (JavaScript object vs. JavaScript object). The Outcome: It spots that 99divs are identical, but 1 div has an update. It flags that single virtual node with an "Update" tag. Because this happens purely in-memory as JavaScript, it is incredibly fast and cheap. The Commit Phase (The Real DOM Update) This is where Reconciliation does its primary job. It acts as a shield to protect the browser from doing unnecessary work. The Scope: React completely ignores the 99 unchanged elements. The Process: It surgically targets the single real browser div associated with the flagged Virtual DOM element and updates only its modified property (e.g., element.textContent = "New Value"). The Outcome: The browser repaints only 1 single div on the screen. The Conclusion: Reconciliation isn't about stopping React from re-rendering (re-running JS to calculate the UI). It