Breakdance data & ACF fields Content Handlers for Analysis and workflows.

Breakdance pages store their content in WordPress post meta as breakdance_data. When Dynamic Fields or ACF fields are used on a page, only the variable reference is stored in breakdance_data — not the resolved value.

This means that to get an accurate word count, image count, or link count — or to send your page content to AI for analysis, repurposing, or SEO assistance — the tool needs to read what actually renders on the page, not what’s stored in the database.

The PHP lives in E:\GIT_REPOS\wp-scos-strategic-content-operating-system, not in the SEO Command Center repo. There are two layers: a legacy extractor in brighter-core, and a newer CA module in site-essentials that delegates to it.

Core content extraction — BW_Content_Analysis

File: brighter-core/includes/class-content-analysis.php

This is the single source of truth for “full page content” (Breakdance + editor + ACF).

Aggregation priority

private static function aggregate_content($post_id, $post) {
        $bd_content = self::get_breakdance_content($post_id);
        if ($bd_content !== '') {
            // Breakdance data exists: use it as primary content (do not double-count with post_content)
            $content = $bd_content;
        } else {
            $content = $post->post_content;
        }
        // ACF fields (if ACF is active)
        if (function_exists('get_fields')) {
            $fields = get_fields($post_id);
            if ($fields) {
                $content .= ' ' . self::extract_acf_content($fields);
            }
        }
        return $content;
    }

Breakdance: reads _breakdance_data (fallback breakdance_data), decodes tree_json_string, walks the tree via scan_content_props() — no per-element whitelist. Handles headings (text + tags), rich HTML, links, and image keys — but not dynamic elements (Query Loop, Post Repeater).

ACF: uses get_fields($post_id) for all fields on the post. extract_acf_content():

• Skips keys starting with _ (ACF internal refs)
• Concatenates string values
• Recurses arrays (repeaters, flexible content, groups)
• Does not resolve relationship/post object fields to linked post content — only stored scalar/meta values

Post types are gated by bw_cs_post_types() in brighter-core/includes/bw-content-strategy.php (all public CPTs except attachments, nav items, WooCommerce types, etc.).

Word count / CAR analysis — Content_Analysis (site-essentials)

File: site-essentials/Modules/ContentArchitecture/Content_Analysis.php

	private static function get_content( $post_id, $post ) {
		if ( class_exists( '\BW_Content_Analysis' )
			&& method_exists( '\BW_Content_Analysis', 'get_aggregated_content' ) ) {
			$content = \BW_Content_Analysis::get_aggregated_content( $post_id );
			if ( $content ) {
				return $content;
			}
		}
		return $post->post_content;
	}

So both bw_word_count and scos_ca_word_count should come from the same aggregated source — but they’re written by separate save hooks and can drift if one path fails (permissions, skip-on-unchanged, etc.).

Important: gather_content_inventory.py in the command center only reads pre-computed scos_ca_word_count meta — it does not call get_aggregated_content() live.

Social amplification — how content gets to Make.com

Two-step flow:

1. Webhook trigger (metadata only)

File: brighter-core/includes/social-amplification/class-webhook-trigger.php

manual_trigger($post_id) POSTs JSON to Make.com with post ID, URL, title, excerpt, dates, breadcrumb, content_type, featured image URLs — no body content in the manual payload.

Automatic trigger (currently disabled) would include image_optimization with full content via get_image_optimization_data().

2. REST callback (full content for AI)

File: brighter-core/includes/social-amplification/class-social-amplification-api.php

The Make.com scenario is designed to call generate-prompt after the webhook fires — that’s where Breakdance + ACF body text becomes source_material.

        // Use same aggregated content as Content Analysis (post_content + ACF + Breakdance)
        $raw_content = class_exists('BW_Content_Analysis') ? BW_Content_Analysis::get_aggregated_content($post_id) : $post->post_content;
        $source_material = self::sanitize_content_for_prompt($raw_content);

Known bug vectors (matches your “content still buggy” note)

1. Breakdance images often count as 0 — BD JSON stores from/id, not resolved URLs; <img src=""> tags may be empty at analysis time (confirmed on some productions site pages).
2. Dynamic BD elements invisible — Query Loops / Post Repeaters have no static text in the tree; word count will be low or zero on pages that are mostly dynamic (e.g. word_count: "7" on the QR generator post in your inventory).
3. ACF relationship fields — only raw stored values are included, not linked post content.
4. Skip-on-unchanged — both engines skip if last_analyzed === post_modified. Breakdance saves via REST may update _breakdance_data without changing post_modified; the on_breakdance_data_saved() hooks exist specifically to fix that, but anything that bypasses those hooks will leave stale counts.
5. Dual meta prefixes — legacy bw_* vs scos_ca_*; inventory uses whichever prefix it detects first.
6. Social webhook vs generate-prompt split — if Make.com isn’t calling generate-prompt, the automation only gets excerpt/title, not aggregated body content.

Quick reference — what to call when

# Live aggregated content (Breakdance + ACF + editor)

wp eval ‘echo BW_Content_Analysis::get_aggregated_content(POST_ID);’

# Pre-computed word count (may be stale)

wp post meta get POST_ID scos_ca_word_count

# Force re-analysis batch

# Admin → CA Overview → Run All, or wp_ajax scos_run_analysis_batch