npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@lizychy0329/pi-ax

v0.0.6

Published

ax integration for pi: structured web extraction via CSS selectors and HTTP API calls

Readme

hero

pi-ax

ax 集成包 for pi — 结构化网页提取,无需写任何解析脚本。

ax is a Rust CLI built for fast, token-cheap web extraction. This package wraps it as a pi extension — giving your agent a web_extract tool with 7 operation modes, plus a skill that tells the agent when to use it vs fetch_content.


项目简介与技术栈

| 项目 | 说明 | |---|---| | ax CLI | Rust — 底层提取引擎,支持 7 种模式、CSS 选择器、TSV 输出、URL 缓存 | | web_extract 工具 | TypeScript — 将 ax CLI 封装为 pi 结构化参数工具,注册为 web_extract | | SKILL | Markdown — 指导 agent 何时用 ax、何时用 fetch_content | | 自动安装 | 首次使用时自动检测环境,缺失则从 ax 官方 下载二进制到 ~/.pi/agent/bin/(无需 sudo) | | 测试 | Vitest — 33 单元测试 + 7 集成测试,覆盖全部 mode 分支和可选参数 |

架构

pi Agent (LLM)
    │
    ├── web_extract 工具 ─── buildAxArgs(params) ─── spawnSync("ax", args)
    │         │                              │
    │         └── TypeBox schema             └── Rust ax CLI (7 modes)
    │
    └── SKILL ─── 决策指南:何时 ax、何时 fetch_content

快速开始

安装

# 从 npm 安装(推荐)
pi install npm:@lizychy0329/pi-ax

# 或从 GitHub 安装
pi install git:github.com/lizyChy0329/pi-ax

安装后,pi 会自动:

  1. 检测 ax 是否已安装
  2. 若未安装,自动下载 ax 二进制到 ~/.pi/agent/bin/(无需 sudo)
  3. 注册 web_extract 工具和 ax skill

第一个请求

告诉 agent:

"提取 https://github.com 页面上所有导航链接的标题"

agent 会自动调用:

{
  "url": "https://github.com",
  "mode": "extract",
  "selector": "a[href]",
  "fields": "title=@innerText"
}

本地开发

git clone https://github.com/lizyChy0329/pi-ax
cd pi-ax
npm install

# 运行测试
npm test

# 临时加载(开发模式)
pi -e ./src/index.ts

为什么需要他?

  • 不用写解析脚本:过去要从 HTML 提取数据得写正则或 DOM 解析,现在只要一个 web_extract 调用
  • Token 便宜:默认 TSV 输出,约为 JSON 的 1/3 token
  • 7 种模式,一个工具:HTTP 报告、结构发现、文本定位、结构化提取、表格提取、Markdown 转换、原始解析
  • 自动安装:不用手动装 ax,缺失时自动从官网下载
  • Skill 决策:告诉 agent 什么场景用 ax、什么场景用 fetch_content,避免工具选择混乱

能做什么?不能做什么?

tools-compare

✅ web_extract 能做

| 模式 | 场景 | 示例 | |---|---|---| | fetch | REST API 调用、HTTP 调试 | fetch(url, method="POST", body='{"a":1}') | | outline | 页面结构发现(提取前先探查) | outline(url) → tag.class 模式计数 | | locate | 找某段文本的 CSS 选择器 | locate(url, query="Pricing") → .nav-pricing | | extract | 结构化行提取(列表、卡片、搜索结果) | extract(selector=".card", fields="title=a, href=a@href") | | table | HTML <table> → 键值行 | table(selector="table") → 列名来自 <th> | | markdown | 文档页 → 可读 Markdown(支持 token 预算) | markdown(url, budget=800) | | parse | 原始页面内容(带 ~2min URL 缓存) | parse(url, fresh=true) 绕过缓存 |

❌ 不该用 web_extract

| 场景 | 改用 | |---|---| | YouTube 视频内容 | fetch_content | | PDF 文档 | fetch_content | | GitHub 仓库分析 | fetch_content | | 纯文章阅读(无需 CSS 选择器) | fetch_content |


工作流:从未知页面提取数据

workflow

步骤 1:发现结构 → web_extract mode: outline

$ ax https://example.com --outline
 210  path
  90  li
  82  a.Link
  37  li.item

步骤 2:定位选择器 → web_extract mode: locate

$ ax https://example.com --locate "Pricing"
→ .nav-pricing

步骤 3:结构化提取 → web_extract mode: extract

{ "url": "https://example.com", "mode": "extract",
  "selector": ".item", "fields": "name=a, price=.price" }

步骤 4:分页继续 → envelope: true + offset

{ "url": "https://example.com", "mode": "extract",
  "selector": ".item", "fields": "name=a",
  "limit": 20, "envelope": true }

Aim for ≤3 tool calls. The 2-minute URL cache makes repeated probes free.

常见陷阱

  • <a> 自身做 row selector:字段 a 在 <a> 内找不到子元素 → 空。用父容器作 row selector。
  • HTTPS→HTTP 跳转:百度等站点强制降级,直接用 http://。
  • 单行压缩 HTML:--outline 效果差,用 --fresh 或 Python 格式化。

相关资源链接

| 资源 | 链接 | |---|---| | ax 官方项目 | github.com/yusukebe/ax | | ax 安装脚本 | ax.yusuke.run/install | | ax 最新发布 | github.com/yusukebe/ax/releases/latest | | 本包源码 | github.com/lizyChy0329/pi-ax | | pi-coding-agent | pi-coding-agent | | TypeBox | github.com/sinclairzx81/typebox | | Vitest | vitest.dev |


License

MIT