‹ 首页

ti-hnparse

@eph5xx · 收录于 1 周前

Python scripts for fetching a single HN post (hnparse.py) and searching HN posts via Algolia (hnsearch.py)

场景抓取帖子搜索内容

适合你，如果需要从 Hacker News 获取或搜索帖子数据。

/ 下载安装
ti-hnparse.skill双击,或拖进 Claude 桌面版 / Cowork,即完成安装↓ .skill↓ .zip
用别的 agent?下载 .zip 解压,把文件夹放进它的技能目录
Claude Code~/.claude/skills/(项目级 .claude/skills/)
Codex CLI~/.codex/skills/
Cursor自动读取上面两处目录
其他工具见其文档的「skills」目录;两个下载是同一份文件,只是名字不同
/ 通过 npx 安装 校验哈希
npx oh-my-skill add eph5xx/tweakidea/ti-hnparse
/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- eph5xx/tweakidea/ti-hnparse
/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify eph5xx/tweakidea/ti-hnparse
安装目标可用 --agent / --scope 或 --to 明确指定;省略时只会在唯一已存在的 agent 目录上自动选择,零命中或多命中会停止并提示。content_hash 缺失或不一致均拒装。
42GitHub stars
~460最小装载
~3.2K含声明引用
~3.2K文本包总量
镜像托管

怎么用

技能原文 SKILL.md作者撰写 · MIT · 7275849

Contains two standalone Python scripts that talk to the free Algolia Hacker News API:

  • hnparse.py — fetch one HN post (article + full comment tree) and write structured markdown.
  • hnsearch.py — search/list HN posts by query and time range; emits JSON to stdout.
hnparse.py usage
uv run <path-to>/hnparse.py <hn-url-or-id> [-o <output-directory>]
  • Uses uv run inline script metadata for dependencies (playwright, httpx, trafilatura)
  • Fetches post metadata and comment tree from Algolia HN API
  • Extracts linked article via Playwright headless browser (falls back to plain HTTP)
  • Writes content.md to the output directory
  • Default output: ~/.tweakidea/hn/hn-{item_id}/content.md
  • Prints the absolute path of the output file to stdout (last line)
  • All progress messages go to stderr

For better article extraction from JS-heavy sites, run uv run playwright install chromium once. Without it, the script falls back to plain HTTP which works fine for most sites.

hnsearch.py usage
uv run <path-to>/hnsearch.py --query "<freeform>" --days <N> \
  [--tags story|show_hn|ask_hn|front_page|poll] \
  [--min-points N] [--limit N]
  • Single dependency: httpx (no Playwright / trafilatura needed)
  • Queries https://hn.algolia.com/api/v1/search (popularity-sorted)
  • --query required; --days required (0 = no time filter / all-time)
  • --tags defaults to story (excludes comments); allowed: story, show_hn, ask_hn, front_page, poll; multiple tags comma-separated
  • --min-points defaults to 10; --limit defaults to 20 (max 50)
  • Prints a JSON array of hits to stdout. Each hit has: id, title, url, points, num_comments, author, created_at_i, relative_age, hn_url.
  • Progress and errors go to stderr. Exits non-zero on HTTP failure.
按 MIT 许可原样转载,未经改动 · 在 GitHub 查看 →

评论

登录即可评论;带「已验证安装」的,是发布者名下有本店的安装或持有记录。
…