Object Oriented Content Migration – – What is migration?

Object Oriented Content Migration

Matt Johnson / @xmatt / alleyinteractive.com

http://xmattus.github.io/object-oriented-content-migration

• We are a full-service digital agency
• WordPress.com VIP partner
• Hiring!

What is migration?

• Your client has an old site (maybe WordPress, maybe something else).
• You're making them a new site.
• Their old site has content in it. They worked hard on that content!

Content migration can be one of the most fun parts of a project...

• Reverse-engineering weird legacy systems.
• Building code to clean up and format your content real nice.
• The satisfaction of processing thousands (or hundreds of thousands) of posts with a single CLI command.

...or one of the least fun

• Oops, legacy content is Windows-1252 content encoding, WordPress speaks UTF-8.
• Half the authors are mysteriously missing.
• "Oh hey, Matt, we have this microsite we forgot to tell you about until right now, a week before the launch. Can we just merge it into the main site's migration?"

WXR: Usually Not an Option

• Most Alley projects fully transform sites.
• Rare (but not unheard of) for old site to be WP at all.
• Major information architecture changes are typical, e.g.
  • Switching from users-as-authors to Co-Authors Plus
  • Loading custom metadata into Fieldmanager
  • Remapping all the taxonomy
  • Adding several custom post types

Make Your Own WXR?

• Can be unwieldy; need code to hook into the old data and generate XML.
• Still limited by the format of WXR.
• No test to see content on a real site before sending to VIP.

Object-Oriented Approach

• ETL means extract, transform, load.
• Use classes to encapsulate extracting and loading.
• Processor: Extracts data.
• Migrateable: Loads data.
• Transform as needed when instantiating migrateables.
• Deploy via WP-CLI.

Processor class

class Processor {
    public $migrateable_class;

    public function __construct() {
        $this->migrateable_class = 'Migrateable';
    }

    abstract function advance_cursor();

    abstract function get_next_item();

    public function load_migrateable() {
        $item = $this->get_next_item();
        $this->migrateable = new $this->migrateable_class( $item );
        $this->advance_cursor();
    }

    public function save_migrateable() {
    	$this->migrateable->save();
    }

}
						

Processor Class

• Abstract methods for iterating over legacy data objects.
• For specific processor types, will impose related migrateable class.

Migrateable Class

// Example where $item is a simple assoc array, e.g. parsed JSON.
class Migrateable {
    public $item;

    public function __construct( $item ) {
        $this->item = $item;
    }

    public function get_title() {
        return empty( $this->item['title'] ) ? $this->item['title'] : '';
    }

    public function get_content() {
    	return empty( $this->item['content'] ) ? $this->item['content'] : '';
    }

    public function get_legacy_url() {
    	return empty( $this->item['url'] ) ? $this->item['url'] : '';
    }

    public function save() {
        $post = $this->get_post_by_legacy_url();
        if ( ! $post ) {
          $post = array();
        }
        $post['post_title'] = $this->get_title();
        $post['post_content'] = $this->get_content();
        // More fields and other fun transformations here
        $post_id = wp_update_post( $post );
        update_post_meta( $post_id, 'legacy_url', $this->get_legacy_url() );
    }

    public function get_post_by_legacy_url() {
    	$post = big_ugly_meta_query( $this->get_legacy_url() );
    	if ( $post && ! is_wp_error( $post ) ) {
    	    return $post;
    	}
    	return false;
    }
}
						

Migrateable Class

• Specific ways to pull data from legacy format.
• Can transform pieces of data in these methods.
• Built-in idempotent behavior.

Implementation Plan

• Subclass Processor and Migrateable to tailor to your situation.
• Control process from WP-CLI using cursor methods in Processor.

The Importance of Idempotence

• Convenient when migrating thousands (or millions) of posts.
• Potentially life-saving when on a deadline!

Some Migrateable Source Examples

• A MySQL database (any weird schema).
• A pile of XML or JSON files.
• An RSS feed.
• A REST API.

Ask me later (or now) about...

• Dealing with media
• Dealing with Co-Authors Plus
• Packaging code for VIP