Content and schemas

Every collection declares the frontmatter it expects. Files that do not match fail the build with the file name and the problem, instead of surfacing as nil in a view.

Declaring attributes

module Content
  class Post < Andromeda::Entry
    collection :posts, base: "app/content/posts", pattern: "**/*.{md,mdx}"

    attribute :title, :string, required: true
    attribute :pub_date, :date, required: true
    attribute :draft, :boolean, default: false
    attribute :tags, :array, of: :string, default: []
    attribute :hero_image, :image
    attribute :author, :reference, collection: :authors

    scope :published, -> { where(draft: false) }

    def reading_time = (body.length / 400.0).ceil
  end
end

Type	Accepts	Becomes
:string :integer :float :boolean	the matching YAML type	itself
:date :datetime	2026-09-01, 'Jul 08 2022', ISO strings	Date / Time
:array	a list, with of: for the element type	Array
:enum	one of values:	the value
:hash	a mapping	Hash
:image	a path relative to the entry	an image, published through the asset pipeline
:reference	an id in another collection	the target entry, resolved lazily

Options: required:, default: (a value or a callable), of:, values:, collection:.

Frontmatter

YAML (---) and TOML (+++) are both supported. YAML is read with 1.2 semantics, matching Astro: yes and no stay strings, while true and false are booleans. Dates and timestamps become Date and Time.

Keys are snake_case. Content copied from an Astro project usually has camelCase keys; bin/rails andromeda:fix rewrites them, and bin/rails andromeda:check reports them.

Querying

Content::Post.all
Content::Post.published.order(pub_date: :desc).limit(10)
Content::Post.where(draft: false, lang: "en")          # every key must match
Content::Post.where { |post| post.tags.include?("rails") }
Content::Post.find("hello-world")   # raises Andromeda::EntryNotFound

where with a hash compares values for equality, so use the block form to look inside an array. order is stable: entries that tie keep the order they were loaded in, as they do in Astro.

Entries expose id, data, body, file_path, digest, html and headings, plus a reader per declared attribute (post.title).

Entry ids

Ids follow Astro’s glob() loader rules, so URLs from an Astro project keep working:

File	Id
hello-world.md	hello-world
Guides/Getting Started.md	guides/getting-started
posts/hello/index.mdx	posts/hello
a slug: in the frontmatter	that value, verbatim