@core / blog

1.0.1 ▾

1.0.1

vérifiéMIT

GitHub

Blog sur @core/content : articles, auteurs, catégories, tags, séries, articles liés, archives et flux RSS/Atom/JSON

Code7 fichiersContexte~554 tokensanalyse réussie

~554 tokens→ src/lib/blog/AGENTS.md→ .cursor/rules/genpm-core-blog.mdc

Voici exactement ce que lit votre IA quand elle travaille dans src/lib/blog. Rien d’autre n’est ajouté à son contexte.

@core/blog — rules for AI agents

Purpose

Blog domain built only from @core/content collections (posts, authors, categories, tags, series; no tables of its own): filtered and paginated listings, a post with its relations resolved, related posts, monthly archive, series order, RSS 2.0 / Atom / JSON Feed, sitemap and search sources and admin resources. No routes or templates (those come from the app or @core/kit-blog), no comments or newsletter (separate modules).

Map

index.ts — public API: listPosts, getPost, relatedPosts, archive, postsInMonth, rssFeed, atomFeed, jsonFeed, blogSitemapSource, blogSearchSource, blogAdminResources.
collections.ts — schemas (PostSchema uses RichTextSchema). posts.ts — reads. feeds.ts — feeds. sources.ts — sitemap/search/admin.

Integration

Import this module at startup (it registers the collections) and set SITE_URL.
Routes (Next.js example): /blog → listPosts({ page, locale }); /blog/[slug] → getPost(slug, { locale, draft }) (404 when null) + relatedPosts(post); /blog/category/[slug] → listPosts({ category }); same for tag, author, series.
Feeds: /blog/feed.xml → rssFeed({ title, description }) (application/rss+xml), /blog/atom.xml, /blog/feed.json.
Register blogSitemapSource() in src/genpm/seo.ts, blogSearchSource() in src/genpm/search.ts and ...blogAdminResources() in src/genpm/admin.ts.
Render the body with <RichText doc={post.data.body} /> (@core/rich-text) and add jsonLd.article(...) (@core/seo).
Verify: publish a post, it appears in /blog, the feed and the sitemap; a scheduled post does not.

Conventions

Relations are slugs: create the author/category/tag entries with the same slug.
membersOnly posts: the page must hide the body unless the user is entitled; feeds and search only expose the excerpt.
Paths default to /blog/<slug>; pass your own path/postPath if the blog lives elsewhere.

Don't

Don't query content_entries directly for posts; use listPosts (it respects publication and locale fallback).
Don't put the full body of members-only posts in feeds, search or meta descriptions.

# @core/blog — rules for AI agents

## Purpose
Blog domain built only from @core/content collections (`posts`, `authors`, `categories`, `tags`, `series`; no tables of
its own): filtered and paginated listings, a post with its relations resolved, related posts, monthly archive, series
order, RSS 2.0 / Atom / JSON Feed, sitemap and search sources and admin resources. No routes or templates (those
come from the app or @core/kit-blog), no comments or newsletter (separate modules).

## Map
- `index.ts` — public API: `listPosts`, `getPost`, `relatedPosts`, `archive`, `postsInMonth`, `rssFeed`, `atomFeed`, `jsonFeed`, `blogSitemapSource`, `blogSearchSource`, `blogAdminResources`.
- `collections.ts` — schemas (`PostSchema` uses `RichTextSchema`). `posts.ts` — reads. `feeds.ts` — feeds. `sources.ts` — sitemap/search/admin.

## Integration
1. Import this module at startup (it registers the collections) and set `SITE_URL`.
2. Routes (Next.js example): `/blog` → `listPosts({ page, locale })`; `/blog/[slug]` → `getPost(slug, { locale, draft })`
   (404 when null) + `relatedPosts(post)`; `/blog/category/[slug]` → `listPosts({ category })`; same for `tag`, `author`, `series`.
3. Feeds: `/blog/feed.xml` → `rssFeed({ title, description })` (`application/rss+xml`), `/blog/atom.xml`, `/blog/feed.json`.
4. Register `blogSitemapSource()` in `src/genpm/seo.ts`, `blogSearchSource()` in `src/genpm/search.ts` and `...blogAdminResources()` in `src/genpm/admin.ts`.
5. Render the body with `<RichText doc={post.data.body} />` (@core/rich-text) and add `jsonLd.article(...)` (@core/seo).
6. Verify: publish a post, it appears in `/blog`, the feed and the sitemap; a scheduled post does not.

## Conventions
- Relations are slugs: create the author/category/tag entries with the same slug.
- `membersOnly` posts: the page must hide the body unless the user is entitled; feeds and search only expose the excerpt.
- Paths default to `/blog/<slug>`; pass your own `path`/`postPath` if the blog lives elsewhere.

## Don't
- Don't query `content_entries` directly for posts; use `listPosts` (it respects publication and locale fallback).
- Don't put the full body of members-only posts in feeds, search or meta descriptions.

L’arborescence exacte qui sera injectée, après .genpmignore. Épinglée à

src/lib/blog/sources.tslecture seule · 8aaf3ee

// Fuentes de sitemap y buscador del blog, y recursos del panel.
import { contentAdminResource, contentSearchSource, contentSitemapSource } from '../content/index.ts';
import { type Doc, toPlainText } from '../rich-text/index.ts';
import { BLOG_COLLECTIONS } from './collections.ts';

type PathFn = (p: { slug: string; locale: string }) => string;
const defaultPath: PathFn = ({ slug }) => `/blog/${slug}`;

export const blogSitemapSource = (path: PathFn = defaultPath) => contentSitemapSource('posts', path);

/** Documentos de búsqueda con el cuerpo en texto plano. Los posts `membersOnly` solo indexan título y extracto. */
export const blogSearchSource = (path: PathFn = defaultPath) =>
  contentSearchSource('posts', {
    type: 'post',
    path,
    title: (d) => String(d.title ?? ''),
    text: (d) => (d.membersOnly ? String(d.excerpt ?? '') : `${d.excerpt ?? ''}\n${toPlainText(d.body as Doc)}`),
  });

/** Recursos del panel: posts (con enlaces "Ver" y "Vista previa"), autores, categorías, etiquetas y series. */
export const blogAdminResources = (path: PathFn = defaultPath) =>
  BLOG_COLLECTIONS.map((c) => ({ ...contentAdminResource(c, c === 'posts' ? { path } : {}), group: 'Blog' }));