PRC Schema SEO

Provides schema generation, SEO meta tags, primary term management, preview panels, and site-level template defaults for titles & descriptions on Pew Research Center platform sites.

Features

• JSON-LD schema generation with post-type level overrides
• Meta tag output (title, description, canonical, robots, Open Graph, Twitter)
• Primary term selection per taxonomy
• Editor previews (search, social, chat, internal)
• Site Editor sidebar for global pattern configuration and noindex lists
• Extensible pattern token engine

Pattern Tokens

Patterns can be set for title & description via Site Editor sidebar fields. The following tokens are available:

Static tokens:

%post_title%
%site_name%
%primary_category%
%post_type%
%year%
%author%
%categories%      (comma-separated list of category names)
%tags%            (comma-separated list of tag names)

Dynamic tokens:

%primary_term:taxonomy%   (chosen primary term for the taxonomy, falls back to first term)
%terms:taxonomy%          (comma-separated list of all term names for taxonomy)

Whitespace is normalized after replacement. Empty tokens resolve to an empty string.

Extending Tokens

Use the prc_schema_seo_pattern_tokens filter to register additional simple tokens (key/value pairs). Dynamic patterns should be handled separately (e.g. by also parsing custom placeholders inside a secondary filter if needed).

add_filter( 'prc_schema_seo_pattern_tokens', function( $tokens, $post_id ) {
    $modified = get_post( $post_id ) ? get_post_modified_time( 'Y-m-d', false, $post_id ) : '';
    $tokens['%modified_date%'] = $modified;
    return $tokens;
}, 10, 2 );

Filters Overview

Key filters & actions (with typical signatures):

Hook	Type	Signature	Purpose
prc_schema_seo_allowed_schema_types	filter	array types = apply_filters( hook, default, post_type )	Whitelist schema types shown in UI
prc_schema_seo_schema_type_default	filter	string type = apply_filters( hook, defaultType, post_type )	Override default schema type per post type
prc_schema_seo_schema_data	filter	array graph = apply_filters( hook, graphArray, post_id, seo_data )	Modify array of schema objects before JSON-LD serialization
prc_schema_seo_person_schema	filter	Person obj = apply_filters( hook, personSchema, post_id, seo_data )	Adjust Person schema
prc_schema_seo_article_schema	filter	Article obj = apply_filters( hook, articleSchema, post_id, seo_data, schema_type )	Adjust Article/NewsArticle/etc schema
prc_schema_seo_webpage_schema	filter	WebPage obj = apply_filters( hook, webpageSchema, post_id, seo_data )	Adjust WebPage schema
prc_schema_seo_organization_schema	filter	Organization obj = apply_filters( hook, orgSchema )	Adjust Organization schema
prc_schema_seo_should_output_schema	filter	bool should = apply_filters( hook, true, post_id )	Conditionally disable schema output
prc_schema_seo_meta_tags	filter	array tags = apply_filters( hook, metaArray, post_id, seo_data )	Modify meta tag array pre-render
prc_schema_seo_canonical_url	filter	string url = apply_filters( hook, permalink, post_id, seo_data )	Override canonical URL
prc_schema_seo_noindex	filter	bool noindex = apply_filters( hook, isNoindex, post_id, seo_data )	Override robots noindex flag
prc_schema_seo_og_image_url	filter	string url = apply_filters( hook, ogUrl, post_id, seo_data )	Override Open Graph image URL
prc_schema_seo_twitter_image_url	filter	string url = apply_filters( hook, twitterUrl, post_id, seo_data )	Override Twitter image URL
prc_schema_seo_title	filter	string title = apply_filters( hook, title, post_id )	Final title after pattern resolution
prc_schema_seo_title_pattern	filter	string pattern = apply_filters( hook, rawPattern, post_id )	Adjust raw title pattern before token substitution
prc_schema_seo_description_pattern	filter	string pattern = apply_filters( hook, rawPattern, post_id )	Adjust raw description pattern before token substitution
prc_schema_seo_description_fallback	filter	string desc = apply_filters( hook, fallbackDesc, post_id )	Provide fallback description text
prc_schema_seo_pattern_tokens	filter	array tokens = apply_filters( hook, tokenMap, post_id )	Add/modify pattern tokens
prc_schema_seo_primary_term_id	filter	int termId = apply_filters( hook, termId, taxonomy, post_id )	Override selected primary term per taxonomy
prc_schema_seo_template_defaults	filter	array defaults = apply_filters( hook, defaultsArray )	Override site-level template defaults set via settings
prc_schema_seo_rest_prepare	filter	array data = apply_filters( hook, seoData, post_id )	Modify REST response for seo data
prc_schema_seo_sanitized_primary_terms	filter	array terms = apply_filters( hook, sanitizedTerms, rawData )	Adjust sanitized primary terms map
prc_schema_seo_cache_cleared	action	do_action( hook, post_id )	Fires after SEO cache is invalidated
prc_schema_seo_schema_output	action	do_action( hook, post_id, jsonLdScript )	Fires after schema markup is printed
prc_schema_seo_after_meta_tags	action	do_action( hook, post_id, tagsHtmlOrArray )	Fires after meta tags output/cached

Return types should match expected types; invalid types may be ignored and logged (see validation in assets loader for schema types).

Building

From monorepo root:

npm run build -w @prc/schema-seo

Internationalization

All user-facing strings are wrapped with __(), _e(), or esc_html_e() using the prc-schema-seo text domain. A POT file is maintained in languages/prc-schema-seo.pot.

To regenerate the POT file (requires WP-CLI):

wp i18n make-pot . languages/prc-schema-seo.pot --domain=prc-schema-seo

Or use the npm script for a reminder:

npm run i18n:pot -w @prc/schema-seo

Translators can use the POT file to create PO/MO files for specific locales (e.g., prc-schema-seo-es_ES.po).

Performance

The plugin uses WordPress object caching extensively to ensure fast page loads:

Benchmarked Metrics (Typical Post)

• Schema Generation:
  • Cold (cache miss): ~130ms + 2.4 MB memory
  • Warm (cached): <0.1ms (13,000x faster)
  • Cache TTL: 1 hour
• Meta Tags:
  • Cold: ~8-12ms
  • Warm (cached): <1ms
  • Cache TTL: 1 hour

Cache Strategy

• All schema and meta tag output is cached using wp_cache_* functions
• Cache keys are post/term-specific: schema_{post_id}, meta_tags_{post_id}
• Cache automatically invalidates on post/term updates
• Memory overhead is minimal on warm requests (~0 KB incremental)

Performance Characteristics

• Normal operation: Warm cache ensures <1ms overhead per request
• Cache miss scenarios: Post update, cache flush, or first view after TTL expiry
• Cold timing acceptable: 130ms amortized over 1 hour (3600 requests) = 0.036ms average per request
• VIP-optimized: Uses memcached object cache on WordPress VIP for distributed caching
• Optimization applied: Batched taxonomy term queries reduce cold generation time

Benchmarking

Use the included WP-CLI command to measure performance on your content:

wp prc-schema-seo benchmark --post=123
wp prc-schema-seo benchmark --post=123 --iterations=5
wp prc-schema-seo benchmark --term=55 --taxonomy=category

Output includes cold/warm timings, memory usage, cache ratios, and output sizes in JSON format.

Stress Test (Heavy Taxonomy Load)

Benchmark on a deliberately heavy post (≈50 categories + 40 tags) to validate worst-case performance:

Scenario	Schema Cold	Schema Warm	Meta Cold	Meta Warm Avg	Cold Memory (Schema)	Cold Memory (Meta)	Output Size (Schema)	Output Size (Meta)
Heavy Post	119.25ms	0.01ms	1.01ms	0.80ms	~1.62MB	~1.5KB	1400 bytes	1436 bytes

Observations:

• Warm cache retrieval remains near-zero even under large taxonomy cardinality.
• One-time cold schema cost (≈120ms) is acceptable given 1h TTL and amortization.
• Memory footprint for cold build reflects Spatie object graph construction; warm delta is effectively 0KB.
• Meta tag generation remains sub-millisecond warm even when schema includes many term references.

Planned refinements (optional): switch to hrtime() for higher precision warm timings and static caching of invariant Organization/Publisher nodes to reduce cold build time further.

Notes

• Title/description patterns only apply when explicit values are missing.
• Primary term taxonomy map stored in _prc_seo_data meta key.
• Caching uses dedicated cache group for schema & meta tags.
• Use prc_schema_seo_allowed_schema_types to restrict UI options; invalid filter return (non-array or non-string entries) will be logged and ignored.
• Pattern engine supports both static and dynamic tokens; use prc_schema_seo_pattern_tokens for simple replacements and dynamic hooks for custom complex parsing.