Help with generation.

2.1.0-alpha 2020-04-27 20:22 UTC

This package is auto-updated.

Last update: 2023-11-28 04:20:58 UTC


Help with markup generation.

Build Status


Note: Currently, Schemer generates only the "microdata" format.

Schemer is designed to be pretty simple to use: Just pass a structured array to the ::build() static method on the Scheme you want:

echo AlwaysBlank\Schemer\Scheme\PostalAddress::build([
    ['street', '123 Oak St'],
    ['state', 'OR'],
    ['city', 'Portland'],
    ['zip', '97123'],
    ['pobox', 'P.O. 1234'],
    ['country', 'USA'],

// <span itemscope itemprop=""><span itemprop="streetAddress">123 Oak St</span><span itemprop="addressRegion">OR</span><span itemprop="addressLocality">Portland</span><span itemprop="postalCode">97123</span><span itemprop="postOfficeBoxNumber">P.O. 1234</span><span itemprop="addressCountry">USA</span></span>

You can optionally pass a second argument, which is a array of arguments that will override the defaults. This allows you to do things like change the wrapping element tag, add arbitrary attributes, etc.

echo AlwaysBlank\Schemer\Scheme\PostalAddress::build([
    ['street', '123 Oak St'],
    ['state', 'OR'],
    ['city', 'Portland'],
    ['zip', '97123'],
    ['pobox', 'P.O. 1234'],
    ['country', 'USA'],
], [
    'tag' => 'a',
    'attributes' => [
        'href'      => '',
        'hidden'    => true,
// <a itemscope itemprop="" href="' hidden><span itemprop="streetAddress">123 Oak St</span><span itemprop="addressRegion">OR</span><span itemprop="addressLocality">Portland</span><span itemprop="postalCode">97123</span><span itemprop="postOfficeBoxNumber">P.O. 1234</span><span itemprop="addressCountry">USA</span></a>

The actual HTML output is slightly different: To allow for simple breaks but maximum styling options, each inline element is followed by <span class="spc">&#8203;</span>, a zero-width space. This allows the browser to break at this point, but contributes no formatting apart from that.


Currently Schemer supports the following Schemes in some fashion. More will be added in the future! Please feel free to file an issue with schemes you'd like to see, or—even better!—a pull request adding them. See "How It Works" below for more information on how to create new Schemes and Properties.

  • PostalAddress can understand the following:
    • street
    • state
    • city
    • zip (zip code)
    • pobox (P.O. box)
    • country
  • LocalBusiness can understand the following:
    • name
    • phone (phone segments will be rendered as links with tel:)
    • url (url can take a string, or an array with two items keyed url and content)
    • address (this is an array containing keys that will be understood by PostalAddress)

How It Works

There are three basic parts to Schemer:

  • Node
  • Property
  • Scheme


The Node is the simplest element, and also the only element that actually behaves like a class (Propertys and Schemes, though classes, are really just wrappers for static methods). It provides a sort of wrapper for individual segments of a Propery or Scheme, and is what is used to generate the actual HTML.

In general, you probably won't be interacting with Nodes very much outside of creating them by passing a set of arguments:

$Node = Node::add([
    'itemscope'     => true,
    'itemtype'      => '',
    'itemprop'      => 'description',
    'tag'           => 'div',
    'content'       => "Always Blank",
    'attributes'    => [
        'href'      => '',
        'hidden'    => true,

Understanding this syntax is helpful, because it will be necessary for creating new Propertys, or to modify existing properties or Schemes.


A Property's one purpose in life is to return a Node with the appropriate arguments for the the property it represents. It is also implemented as a Trait so that it can be composed into a Scheme later on.

A Property must meet the following requirements:

  • It implements a public static method that has the same name as whatever you want to use as the key/name for the field in question—i.e. address or phone. In general, these names should be simple rather than directly matching the property they represent. For instance, telephone is represented here as just phone.
  • This method takes only one argument.
  • This method must return a Node.
  • The Class and file names of the Property must be, essentially, CamelCase versions of the method name. i.e. pobox => POBox, name => Name, etc.

A Property can take input of any kind (although most take strings), but you should be aware that there isn't a obvious mechanism (apart from reading the source code) to know what specific format a Property might be expecting its argument to be in, so keep user experience in mind when building these things.


A Scheme is the top dog; the final thing Schemer exists to do. They are surprisingly simple. A Scheme has to do the following:

  • extend the class AlwaysBlank\Schemer\Scheme\Scheme. This actually probably most of the functionality.
  • Implement a public static method called wrap. It takes two arguments: content, which is a string, and args which is an optional array. It returns a Node. Use this method to set the properties (i.e. itemprop) that the element that contains this Scheme will need.
  • use any Property Traits that you want this Scheme to have access to. This defines the arguments it will accept.

Here's an example Scheme:

class Example extends AlwaysBlank\Schemer\Scheme\Scheme
    use City;
    use Name;
    public static function wrap(string $content, array $args = []): Node
        return Node::add(array_merge([
            'itemscope'     => true,
            'itemtype'      => '',
            'content'       => $content,
            'tag'           => 'section',
            'attributes'    => [
                'data-section' => 'example',
        ], $args));


When defining Nodes, you can use abbreviations for the properties. This may be useful if you find yourself in a situation where you need to type up a lot of Nodes.

Supported abbreviations are:

Property Alias(es)
itemscope scope, iscope
itemprop prop, iprop
itemtype type, itype
attributes attr, attrs

This will create a Scheme called Example that will understand city and name, and will use a <section> with an additional attribute of data-section="example".

Going Further

If you have more questions about how these things work, I encourage you to dig into the source code. I've made an effort to document the code clearly inline, and keep everything simple and clear.


To debug Schemer, set the PHP constant ALWAYSBLANK_SCHEMER_DEBUG to true. As with any debug flag, I don't recommend you set this in production.

Currently the debug flag has the following effects (this list may expand with time):

  • Stop suppressing exceptions when calling Propertys with bad arguments.



The output of a Scheme is generally pretty "flat" in the sense that it's just a list of elements that aren't nested within on another. (The exception being if the Scheme in question includes other Schemes as Propertys, as LocalBusiness does with Address.) This is more or less by design: So far as I can tell, building a system that would allow a user to dynamically modify the nesting of arbitrary elements would create an extreme amount of complexity in a library that is supposed to be very simple. In general, Scheme-level content isn't usually heavily nested anyway (apart from the aforementioned exceptions).

If you find yourself needing a more complex structure, Schemer is design to be flexible enough that you can build Schemes and Propertys "on the fly"—or even just use Propertys directly without the need for a Scheme.


This library is currently limited in terms of what it "understands". While it can be easily expanded, I have no intention of ever making it comprehensive with respect to the full specifications. Polite feature requests are always welcome (and good PRs will likely be merged without question), but please keep in mind that the primary purpose of this library is to make a small slice of repetitive tasks less repetitive, not to solve all microdata-related problems.