<?php
/**
 * Canonical demo dataset + shared constants (weights, categories, bots).
 * Ported from the approved design handoff so the dashboard is populated
 * out-of-the-box for the Pillow Fair demo and as a fallback before the
 * first real scan.
 *
 * @package AICitare
 */

if ( ! defined( 'ABSPATH' ) ) {
	exit;
}

class AICitare_Data {

	/** Weighted contribution of each sub-score to the GEO Score (must sum to 1). */
	public static function weights() {
		return array(
			'schema'    => 0.25,
			'structure' => 0.25,
			'crawl'     => 0.20,
			'authority' => 0.20,
			'freshness' => 0.10,
		);
	}

	/** Category metadata: label + tag colors (per design tokens). */
	public static function categories() {
		return array(
			'schema'    => array( 'label' => 'Structured Data', 'color' => '#5b50e6', 'bg' => '#ecebfd' ),
			'structure' => array( 'label' => 'Content Structure', 'color' => '#0d9488', 'bg' => '#dcf5f1' ),
			'crawl'     => array( 'label' => 'Crawlability', 'color' => '#2563eb', 'bg' => '#e3edfe' ),
			'authority' => array( 'label' => 'Authority', 'color' => '#c026d3', 'bg' => '#f9e6fc' ),
			'freshness' => array( 'label' => 'Freshness', 'color' => '#d97706', 'bg' => '#fdf0e0' ),
		);
	}

	/** The 12 AI crawlers the plugin can allow / block. */
	public static function bot_meta() {
		return array(
			array( 'id' => 'GPTBot', 'owner' => 'OpenAI', 'purpose' => 'ChatGPT' ),
			array( 'id' => 'OAI-SearchBot', 'owner' => 'OpenAI', 'purpose' => 'ChatGPT Search' ),
			array( 'id' => 'ClaudeBot', 'owner' => 'Anthropic', 'purpose' => 'Claude' ),
			array( 'id' => 'PerplexityBot', 'owner' => 'Perplexity', 'purpose' => 'Perplexity' ),
			array( 'id' => 'Google-Extended', 'owner' => 'Google', 'purpose' => 'Gemini & AI Overviews' ),
			array( 'id' => 'Bingbot', 'owner' => 'Microsoft', 'purpose' => 'Copilot' ),
			array( 'id' => 'Applebot-Extended', 'owner' => 'Apple', 'purpose' => 'Apple Intelligence' ),
			array( 'id' => 'Amazonbot', 'owner' => 'Amazon', 'purpose' => 'Alexa' ),
			array( 'id' => 'CCBot', 'owner' => 'Common Crawl', 'purpose' => 'Open training set' ),
			array( 'id' => 'Bytespider', 'owner' => 'ByteDance', 'purpose' => 'Doubao AI' ),
			array( 'id' => 'Meta-ExternalAgent', 'owner' => 'Meta', 'purpose' => 'Meta AI' ),
			array( 'id' => 'DuckAssistBot', 'owner' => 'DuckDuckGo', 'purpose' => 'DuckAssist' ),
		);
	}

	/** Default allowed-crawler map. */
	public static function default_bots() {
		return array(
			'GPTBot' => true, 'OAI-SearchBot' => true, 'ClaudeBot' => true, 'PerplexityBot' => true,
			'Google-Extended' => true, 'Bingbot' => true, 'Applebot-Extended' => true, 'Amazonbot' => true,
			'CCBot' => false, 'Bytespider' => false, 'Meta-ExternalAgent' => false, 'DuckAssistBot' => true,
		);
	}

	/**
	 * Hosts that identify a human visitor arriving from an AI assistant.
	 * Keyed by the display source; values are case-insensitive host substrings.
	 */
	public static function ai_referrers() {
		return array(
			'ChatGPT'    => array( 'chatgpt.com', 'chat.openai.com' ),
			'Perplexity' => array( 'perplexity.ai' ),
			'Gemini'     => array( 'gemini.google.com' ),
			'Claude'     => array( 'claude.ai' ),
			'Copilot'    => array( 'copilot.microsoft.com' ),
			'You.com'    => array( 'you.com' ),
			'Poe'        => array( 'poe.com' ),
		);
	}

	/** Default settings. */
	public static function default_settings() {
		return array(
			'schema'     => 'complementary', // complementary | full
			'demo_mode'  => false,           // default OFF: scan real posts. true = use seed dataset.
			'include_drafts' => false,       // also scan/analyze draft & pending posts.
			'check_links'    => false,       // OFF by default: opt-in outbound-link verification (makes external HTTP requests).
			'amazon_enabled' => false,       // Master switch for the Amazon link-verification module.
			'amazon'         => array(       // Amazon PA-API 5.0 credentials (secret stored encrypted).
				'access'  => '',
				'secret'  => '',
				'partner' => '',
			),
			'keys'       => array(
				'openai'     => '',
				'anthropic'  => '',
				'google'     => '',
				'perplexity' => '',
				'grok'       => '',
			),
			'frequency'        => 'daily',   // hourly | twicedaily | daily | weekly
			'visibility_freq'  => 'weekly',  // daily | weekly
			'last_scan'        => '',
			'last_scan_ts'     => 0,
			'competitor_setup' => false,
			'report_email'         => '',
			'reports_enabled'      => false,
			'report_freq'          => 'Monthly',
			'alerts_enabled'       => true,
			'score_drop_threshold' => 5,
			'ollama_enabled'       => false,
			'ollama_url'           => 'http://host.docker.internal:11434',
			'ollama_model'         => 'llama3.2',
			'models'               => array(
				'openai'     => 'gpt-4o-mini',
				'perplexity' => 'sonar',
				'google'     => 'gemini-2.0-flash',
				'anthropic'  => 'claude-3-5-haiku-latest',
				'grok'       => 'grok-3',
			),
		);
	}

	/** Score trend history (5 prior months; current month is computed live). */
	public static function default_history() {
		return array( 54, 57, 56, 61, 65 );
	}

	/** AI citation tracking demo data: position per [ChatGPT, Perplexity, Gemini, Claude, Grok, Ollama] or null. */
	public static function citations() {
		return array(
			array( 'q' => 'best cooling pillow 2026', 'e' => array( null, 2, 4, null, 3, null ) ),
			array( 'q' => 'memory foam vs latex pillow', 'e' => array( 3, 1, null, 5, 2, 4 ) ),
			array( 'q' => 'best pillow for neck pain', 'e' => array( null, null, null, null, null, null ) ),
			array( 'q' => 'side sleeper pillow recommendation', 'e' => array( 2, 3, 2, 4, 3, null ) ),
			array( 'q' => 'is the casper pillow worth it', 'e' => array( 1, 2, null, 3, null, 2 ) ),
			array( 'q' => 'down vs down alternative pillow', 'e' => array( null, 4, null, null, 5, null ) ),
		);
	}

	/** Competitor share-of-voice demo data. */
	public static function competitors() {
		return array(
			array( 'name' => 'Sleep Foundation', 'share' => 34 ),
			array( 'name' => 'Wirecutter', 'share' => 22 ),
			array( 'name' => 'Sleepopolis', 'share' => 15 ),
			array( 'name' => 'Pillow Fair', 'share' => 11, 'me' => true ),
			array( 'name' => "Tom's Guide", 'share' => 10 ),
			array( 'name' => 'Other sites', 'share' => 8 ),
		);
	}

	/** Recent AI crawler hits demo data. */
	public static function bot_log() {
		return array(
			array( 'bot' => 'PerplexityBot', 'page' => '/best-cooling-pillows', 'time' => '4 min ago', 'ok' => true ),
			array( 'bot' => 'GPTBot', 'page' => '/memory-foam-vs-latex', 'time' => '12 min ago', 'ok' => true ),
			array( 'bot' => 'ClaudeBot', 'page' => '/side-sleeper-guide', 'time' => '31 min ago', 'ok' => true ),
			array( 'bot' => 'Bytespider', 'page' => '/neck-pain-pillows', 'time' => '48 min ago', 'ok' => false ),
			array( 'bot' => 'Google-Extended', 'page' => '/casper-original-review', 'time' => '1 hr ago', 'ok' => true ),
			array( 'bot' => 'CCBot', 'page' => '/pillow-loft-guide', 'time' => '2 hr ago', 'ok' => false ),
			array( 'bot' => 'GPTBot', 'page' => '/down-vs-alternative', 'time' => '3 hr ago', 'ok' => true ),
		);
	}

	/** The 7 canonical demo pages with sub-scores and issues. */
	public static function seed_pages() {
		return array(
			array(
				'id' => 'p1', 'title' => 'Best Cooling Pillows for Hot Sleepers (2026)', 'url' => '/best-cooling-pillows',
				'updated' => '8 days ago', 'stale' => false,
				'sub' => array( 'schema' => 40, 'structure' => 55, 'crawl' => 80, 'authority' => 50, 'freshness' => 72 ),
				'issues' => array(
					array( 'id' => 'p1a', 'cat' => 'schema', 'sev' => 'high', 'gain' => 35, 'title' => 'No ItemList or Review schema', 'fixLabel' => 'Add schema', 'desc' => 'This is a product roundup but emits no structured data — AI engines can’t parse which pillows you recommend or their ratings.', 'after' => 'JSON-LD: ItemList · 7 × Review — validated ✓' ),
					array( 'id' => 'p1b', 'cat' => 'structure', 'sev' => 'high', 'gain' => 15, 'title' => 'No quick-answer summary', 'fixLabel' => 'Insert block', 'desc' => 'There is no concise answer block near the top. AI models favor a 40–60 word direct answer they can lift verbatim.', 'after' => 'Quick-answer block inserted above the fold (52 words)' ),
					array( 'id' => 'p1c', 'cat' => 'structure', 'sev' => 'med', 'gain' => 12, 'title' => 'Comparison is an image, not an HTML table', 'fixLabel' => 'Convert table', 'desc' => 'Your spec comparison is a JPG. Crawlers cannot read pricing, loft or firmness from pixels.', 'after' => 'HTML table with 5 columns × 7 rows — machine-readable' ),
					array( 'id' => 'p1d', 'cat' => 'authority', 'sev' => 'high', 'gain' => 25, 'title' => 'No author bio or Person schema', 'fixLabel' => 'Add author box', 'desc' => 'No byline, credentials or Person markup. E-E-A-T signals are a strong driver of AI citation.', 'after' => 'Author box + Person schema (sleep writer, 6 yrs) added' ),
					array( 'id' => 'p1e', 'cat' => 'crawl', 'sev' => 'low', 'gain' => 10, 'title' => 'Not listed in llms.txt', 'fixLabel' => 'Add to llms.txt', 'desc' => 'This high-value page is missing from your llms.txt index, lowering its discovery priority.', 'after' => 'Added to llms.txt under "Buying guides"' ),
				),
			),
			array(
				'id' => 'p2', 'title' => 'Memory Foam vs Latex Pillows: Which Is Right?', 'url' => '/memory-foam-vs-latex',
				'updated' => '11 months ago', 'stale' => true,
				'sub' => array( 'schema' => 60, 'structure' => 65, 'crawl' => 85, 'authority' => 60, 'freshness' => 50 ),
				'issues' => array(
					array( 'id' => 'p2a', 'cat' => 'freshness', 'sev' => 'high', 'gain' => 20, 'title' => 'Content is 11 months old', 'fixLabel' => 'Refresh', 'desc' => 'Last meaningful update was 11 months ago. Update the date and refresh 2026 model references.', 'after' => 'Republished · dateModified set to today' ),
					array( 'id' => 'p2b', 'cat' => 'structure', 'sev' => 'med', 'gain' => 12, 'title' => 'No side-by-side comparison table', 'fixLabel' => 'Add table', 'desc' => 'A direct comparison table strongly boosts pickup for "X vs Y" queries.', 'after' => 'HTML table: Memory foam vs Latex — 9 attributes' ),
					array( 'id' => 'p2c', 'cat' => 'authority', 'sev' => 'med', 'gain' => 15, 'title' => 'No sources cited', 'fixLabel' => 'Add sources', 'desc' => 'No outbound citations or methodology. Adding referenced studies raises trust signals.', 'after' => '3 referenced sources + testing methodology added' ),
				),
			),
			array(
				'id' => 'p3', 'title' => 'Best Pillows for Neck Pain', 'url' => '/neck-pain-pillows',
				'updated' => '16 months ago', 'stale' => true,
				'sub' => array( 'schema' => 35, 'structure' => 48, 'crawl' => 78, 'authority' => 45, 'freshness' => 40 ),
				'issues' => array(
					array( 'id' => 'p3a', 'cat' => 'schema', 'sev' => 'high', 'gain' => 30, 'title' => 'No ItemList / Review schema', 'fixLabel' => 'Add schema', 'desc' => 'A health-adjacent roundup with zero structured data — the biggest gap on the site.', 'after' => 'JSON-LD: ItemList · 6 × Review — added & validated' ),
					array( 'id' => 'p3b', 'cat' => 'structure', 'sev' => 'high', 'gain' => 18, 'title' => 'No FAQ section', 'fixLabel' => 'Generate FAQ', 'desc' => 'Neck-pain queries are highly question-driven. A FAQ block with FAQPage schema captures them.', 'after' => 'FAQ section (5 Q&A) + FAQPage schema added' ),
					array( 'id' => 'p3c', 'cat' => 'structure', 'sev' => 'med', 'gain' => 10, 'title' => 'No quick-answer summary', 'fixLabel' => 'Insert block', 'desc' => 'Add a concise answer near the top for AI lift.', 'after' => 'Quick-answer block inserted (48 words)' ),
					array( 'id' => 'p3d', 'cat' => 'authority', 'sev' => 'high', 'gain' => 22, 'title' => 'No author or medical reviewer', 'fixLabel' => 'Add reviewer', 'desc' => 'Health content needs visible expertise. Add an author plus a reviewer with Person schema.', 'after' => 'Author + medical reviewer credentials added' ),
					array( 'id' => 'p3e', 'cat' => 'freshness', 'sev' => 'high', 'gain' => 25, 'title' => 'Content is 16 months old', 'fixLabel' => 'Refresh', 'desc' => 'Oldest page on the site. Refresh content and dateModified.', 'after' => 'Republished · dateModified set to today' ),
				),
			),
			array(
				'id' => 'p4', 'title' => 'The Complete Side Sleeper Pillow Guide', 'url' => '/side-sleeper-guide',
				'updated' => '3 weeks ago', 'stale' => false,
				'sub' => array( 'schema' => 70, 'structure' => 75, 'crawl' => 88, 'authority' => 65, 'freshness' => 80 ),
				'issues' => array(
					array( 'id' => 'p4a', 'cat' => 'authority', 'sev' => 'med', 'gain' => 12, 'title' => 'No testing methodology section', 'fixLabel' => 'Add methodology', 'desc' => 'You rank pillows but don’t explain how. A short methodology block lifts trust.', 'after' => '"How we tested" section added' ),
					array( 'id' => 'p4b', 'cat' => 'schema', 'sev' => 'low', 'gain' => 10, 'title' => 'Missing FAQ schema', 'fixLabel' => 'Add schema', 'desc' => 'You have a FAQ section but no FAQPage markup wrapping it.', 'after' => 'FAQPage schema attached to existing FAQ' ),
				),
			),
			array(
				'id' => 'p5', 'title' => 'Casper Original Pillow Review', 'url' => '/casper-original-review',
				'updated' => '5 days ago', 'stale' => false,
				'sub' => array( 'schema' => 80, 'structure' => 78, 'crawl' => 90, 'authority' => 72, 'freshness' => 85 ),
				'issues' => array(
					array( 'id' => 'p5a', 'cat' => 'structure', 'sev' => 'low', 'gain' => 8, 'title' => 'No pros / cons block', 'fixLabel' => 'Add block', 'desc' => 'A structured pros/cons list is highly liftable for review queries.', 'after' => 'Pros / cons block added (4 + 3)' ),
					array( 'id' => 'p5b', 'cat' => 'authority', 'sev' => 'med', 'gain' => 10, 'title' => 'Reviewer credentials not shown', 'fixLabel' => 'Add credentials', 'desc' => 'Surface who tested the pillow and for how long.', 'after' => 'Reviewer credentials + tested-for duration shown' ),
				),
			),
			array(
				'id' => 'p6', 'title' => 'Down vs Down-Alternative Pillows', 'url' => '/down-vs-alternative',
				'updated' => '2 months ago', 'stale' => false,
				'sub' => array( 'schema' => 55, 'structure' => 60, 'crawl' => 82, 'authority' => 55, 'freshness' => 62 ),
				'issues' => array(
					array( 'id' => 'p6a', 'cat' => 'schema', 'sev' => 'high', 'gain' => 20, 'title' => 'No Review schema', 'fixLabel' => 'Add schema', 'desc' => 'Comparison verdicts have no structured rating for AI to cite.', 'after' => 'Review schema with ratings added' ),
					array( 'id' => 'p6b', 'cat' => 'structure', 'sev' => 'med', 'gain' => 12, 'title' => 'No comparison table', 'fixLabel' => 'Add table', 'desc' => 'Add an HTML comparison table for the two fill types.', 'after' => 'HTML table: Down vs Alternative — 8 attributes' ),
					array( 'id' => 'p6c', 'cat' => 'crawl', 'sev' => 'low', 'gain' => 8, 'title' => 'Not listed in llms.txt', 'fixLabel' => 'Add to llms.txt', 'desc' => 'Missing from the llms.txt index.', 'after' => 'Added to llms.txt' ),
				),
			),
			array(
				'id' => 'p7', 'title' => 'How to Choose the Right Pillow Loft', 'url' => '/pillow-loft-guide',
				'updated' => '6 weeks ago', 'stale' => false,
				'sub' => array( 'schema' => 50, 'structure' => 70, 'crawl' => 85, 'authority' => 58, 'freshness' => 68 ),
				'issues' => array(
					array( 'id' => 'p7a', 'cat' => 'schema', 'sev' => 'high', 'gain' => 18, 'title' => 'No HowTo or FAQ schema', 'fixLabel' => 'Add schema', 'desc' => 'A how-to guide with no HowTo markup. Add it so steps surface in AI answers.', 'after' => 'HowTo schema (4 steps) + FAQPage added' ),
					array( 'id' => 'p7b', 'cat' => 'freshness', 'sev' => 'med', 'gain' => 12, 'title' => 'Could use a content refresh', 'fixLabel' => 'Refresh', 'desc' => 'Update examples and dateModified to stay current.', 'after' => 'Refreshed · dateModified updated' ),
				),
			),
		);
	}

	/** Get the active page dataset (scanned or seeded). */
	public static function get_pages() {
		$settings = self::get_settings();

		// Demo mode ON: always show the bundled sample dataset.
		if ( ! empty( $settings['demo_mode'] ) ) {
			return self::seed_pages();
		}

		// Demo mode OFF (default): surface only real scanned posts (ids like
		// "wp123"). Never the bundled demo seed (ids "p1"…"p7") — so a fresh
		// install shows an empty "run a scan" state instead of fake pages.
		$pages = get_option( AICITARE_OPT_PAGES );
		if ( ! is_array( $pages ) ) {
			return array();
		}
		return array_values( array_filter( $pages, function ( $p ) {
			return isset( $p['id'] ) && 0 === strpos( (string) $p['id'], 'wp' );
		} ) );
	}

	public static function get_fixed() {
		$f = get_option( AICITARE_OPT_FIXED );
		return is_array( $f ) ? $f : array();
	}

	public static function get_bots() {
		$b = get_option( AICITARE_OPT_BOTS );
		return ( is_array( $b ) && ! empty( $b ) ) ? $b : self::default_bots();
	}

	public static function get_settings() {
		$defaults = self::default_settings();
		$s        = get_option( AICITARE_OPT_SETTINGS );
		if ( ! is_array( $s ) || empty( $s ) ) {
			return $defaults;
		}
		$s = wp_parse_args( $s, $defaults );
		// Deep-merge the keys sub-array so newly added providers (e.g. Grok)
		// appear on installs that saved settings before the provider existed.
		$s['keys']   = wp_parse_args( is_array( $s['keys'] ) ? $s['keys'] : array(), $defaults['keys'] );
		$s['models'] = wp_parse_args( ( isset( $s['models'] ) && is_array( $s['models'] ) ) ? $s['models'] : array(), $defaults['models'] );
		$s['amazon'] = wp_parse_args( ( isset( $s['amazon'] ) && is_array( $s['amazon'] ) ) ? $s['amazon'] : array(), $defaults['amazon'] );
		return $s;
	}

	/** Decrypted API keys for server-side use (never sent to the browser). */
	public static function get_keys_plain() {
		$keys = self::get_settings()['keys'];
		$out  = array();
		foreach ( $keys as $k => $v ) {
			$out[ $k ] = AICitare_Crypto::decrypt( $v );
		}
		return $out;
	}

	public static function get_history() {
		$settings = self::get_settings();
		if ( ! empty( $settings['demo_mode'] ) ) {
			return self::default_history();
		}
		// A score trend is only meaningful once real posts have been scanned.
		// This also hides any leftover demo-era history on upgraded installs.
		if ( ! self::get_pages() ) {
			return array();
		}
		$h = get_option( AICITARE_OPT_HISTORY );
		return ( is_array( $h ) && ! empty( $h ) ) ? $h : array();
	}
}