| 1 |
<?php |
| 2 |
/** |
| 3 |
* SEO Score Calculator |
| 4 |
* |
| 5 |
* Advanced SEO scoring system based on 2025 Google ranking factors |
| 6 |
* Implements AI-driven content analysis, searcher engagement signals, and current SEO best practices |
| 7 |
* |
| 8 |
* @package ThinkRank\AI |
| 9 |
* @since 1.0.0 |
| 10 |
*/ |
| 11 |
|
| 12 |
declare(strict_types=1); |
| 13 |
|
| 14 |
namespace ThinkRank\AI; |
| 15 |
|
| 16 |
use ThinkRank\Core\Database; |
| 17 |
|
| 18 |
// Prevent direct access |
| 19 |
if (!defined('ABSPATH')) { |
| 20 |
exit; |
| 21 |
} |
| 22 |
|
| 23 |
/** |
| 24 |
* SEO Score Calculator Class |
| 25 |
* |
| 26 |
* Implements 2025 SEO scoring algorithm based on: |
| 27 |
* - Google's Q1 2025 algorithm updates (First Page Sage research) |
| 28 |
* - Satisfying content as #1 ranking factor (23%) |
| 29 |
* - Searcher engagement and intent satisfaction (12%) |
| 30 |
* - Mobile Experience Score (MES) and Core Web Vitals 2.0 |
| 31 |
* - Content freshness and niche expertise signals |
| 32 |
* |
| 33 |
* @since 1.0.0 |
| 34 |
*/ |
| 35 |
class SEOScoreCalculator { |
| 36 |
|
| 37 |
/** |
| 38 |
* Database instance |
| 39 |
* |
| 40 |
* @var Database |
| 41 |
*/ |
| 42 |
private Database $database; |
| 43 |
|
| 44 |
/** |
| 45 |
* Memoised collected performance measurement, and whether it was resolved. |
| 46 |
* |
| 47 |
* Two factors read it and both may be asked for on every post in a list, so |
| 48 |
* the lookup happens once per calculator. `null` is a real answer here — the |
| 49 |
* separate flag keeps "not looked up yet" distinct from "nothing measured". |
| 50 |
* |
| 51 |
* @since 2.3.1 |
| 52 |
* @var array|null |
| 53 |
*/ |
| 54 |
private ?array $measured_performance = null; |
| 55 |
|
| 56 |
/** |
| 57 |
* @since 2.3.1 |
| 58 |
* @var bool |
| 59 |
*/ |
| 60 |
private bool $measured_performance_resolved = false; |
| 61 |
|
| 62 |
/** |
| 63 |
* Length bands the editor scores against, in characters. |
| 64 |
* |
| 65 |
* Public so every surface that judges a title or description — the editor |
| 66 |
* score and the Bulk Snippets problem filter — reads one set of numbers. |
| 67 |
* Before these existed the bands were literals inside the scoring methods, |
| 68 |
* and a second screen would have had to copy them and drift (#727). |
| 69 |
* |
| 70 |
* @since 2.8.0 |
| 71 |
*/ |
| 72 |
public const TITLE_OPTIMAL_MIN = 35; |
| 73 |
public const TITLE_OPTIMAL_MAX = 60; |
| 74 |
public const DESCRIPTION_OPTIMAL_MIN = 120; |
| 75 |
public const DESCRIPTION_OPTIMAL_MAX = 160; |
| 76 |
|
| 77 |
/** |
| 78 |
* 2025 SEO scoring factors (Q1 2025 Google Algorithm) |
| 79 |
* Based on First Page Sage research and Google's latest updates |
| 80 |
* |
| 81 |
* @var array |
| 82 |
*/ |
| 83 |
private array $scoring_factors = [ |
| 84 |
'satisfying_content' => 23, // #1: Consistent publication of satisfying content |
| 85 |
'title_optimization' => 14, // #2: Keyword in meta title (looser matching) |
| 86 |
'niche_expertise' => 13, // #3: Hub & spoke content clusters |
| 87 |
'searcher_engagement' => 12, // #4: Dwell time, bounce rate, pages/session |
| 88 |
'backlink_authority' => 13, // #5: Quality backlinks (declining but important) |
| 89 |
'content_freshness' => 6, // #6: Quarterly content updates |
| 90 |
'mobile_experience' => 5, // #7: Mobile Experience Score (MES) - NEW 2025 |
| 91 |
'trustworthiness' => 4, // #8: E-E-A-T verification |
| 92 |
'link_diversity' => 3, // #9: Link distribution across multiple pages |
| 93 |
'core_web_vitals' => 3, // #10: Page speed + Interaction Readiness |
| 94 |
'site_security' => 2, // #11: SSL certificate |
| 95 |
'internal_linking' => 1, // #12: Declining importance |
| 96 |
'technical_factors' => 1, // #13: Meta descriptions, schema, etc. |
| 97 |
]; |
| 98 |
|
| 99 |
/** |
| 100 |
* Constructor |
| 101 |
* |
| 102 |
* @param Database $database Database instance |
| 103 |
*/ |
| 104 |
public function __construct(Database $database) { |
| 105 |
$this->database = $database; |
| 106 |
} |
| 107 |
|
| 108 |
/** |
| 109 |
* Calculate comprehensive modern SEO score. |
| 110 |
* |
| 111 |
* Supports multiple focus keywords: when `$options['target_keywords']` holds |
| 112 |
* more than one keyword the score is computed independently for each and the |
| 113 |
* HIGHEST overall score is returned as the final result, with per-keyword |
| 114 |
* results (`keyword_results`) and OR-combined per-check matches |
| 115 |
* (`keyword_checks`) attached. A single keyword (or the legacy |
| 116 |
* `target_keyword` option) falls through to the single-keyword path. |
| 117 |
* |
| 118 |
* @param array $content_data Content analysis data |
| 119 |
* @param array $metadata Post metadata |
| 120 |
* @param array $options Additional options |
| 121 |
* @return array Complete scoring result |
| 122 |
*/ |
| 123 |
public function calculate_score(array $content_data, array $metadata, array $options = []): array { |
| 124 |
$keywords = $this->resolve_target_keywords($options, $metadata); |
| 125 |
|
| 126 |
if (count($keywords) > 1) { |
| 127 |
return $this->calculate_score_multi($content_data, $metadata, $keywords, $options); |
| 128 |
} |
| 129 |
|
| 130 |
$options['target_keyword'] = $keywords[0] ?? ''; |
| 131 |
$result = $this->compute_score($content_data, $metadata, $options); |
| 132 |
|
| 133 |
// Expose the keyword surface uniformly so consumers can rely on it |
| 134 |
// regardless of how many keywords were supplied. |
| 135 |
$result['target_keywords'] = $keywords; |
| 136 |
if (!empty($keywords)) { |
| 137 |
$result['keyword_results'] = [[ |
| 138 |
'keyword' => $keywords[0], |
| 139 |
'overall_score' => $result['overall_score'], |
| 140 |
'grade' => $result['grade'], |
| 141 |
]]; |
| 142 |
$result['keyword_checks'] = $this->analyze_keyword_checks($content_data, $metadata, $keywords); |
| 143 |
} |
| 144 |
$result['keywords'] = $this->keyword_placements($content_data, $metadata, $keywords); |
| 145 |
|
| 146 |
return $result; |
| 147 |
} |
| 148 |
|
| 149 |
/** |
| 150 |
* Resolve the target keyword list from the options array, falling back to |
| 151 |
* the focus keywords carried on the post's metadata. |
| 152 |
* |
| 153 |
* Accepts `target_keywords` (array) or the legacy `target_keyword` (string), |
| 154 |
* trims, drops empties and removes case-insensitive duplicates. |
| 155 |
* |
| 156 |
* Options win over metadata on purpose: the editor scores unsaved keyword |
| 157 |
* edits by passing them explicitly, and that live value must beat whatever |
| 158 |
* is currently persisted. The metadata fallback applies only when the |
| 159 |
* caller mentions no keyword option AT ALL — a caller that passes an empty |
| 160 |
* keyword is deliberately clearing it (the editor does exactly this when |
| 161 |
* the field is emptied), so the stored value must not resurrect it. |
| 162 |
* |
| 163 |
* Without the fallback, every caller that hands over |
| 164 |
* `Metabox_Manager::get_post_metadata()` (the metabox and the MCP scoring |
| 165 |
* abilities) silently scored as if no keyword were set. |
| 166 |
* |
| 167 |
* @param array $options Scoring options. |
| 168 |
* @param array $metadata Post metadata (may carry focus_keyword(s)). |
| 169 |
* @return string[] Normalized keyword list. |
| 170 |
*/ |
| 171 |
private function resolve_target_keywords(array $options, array $metadata = []): array { |
| 172 |
$raw = []; |
| 173 |
if (!empty($options['target_keywords']) && is_array($options['target_keywords'])) { |
| 174 |
$raw = $options['target_keywords']; |
| 175 |
} elseif (isset($options['target_keyword']) && $options['target_keyword'] !== '') { |
| 176 |
$raw = [$options['target_keyword']]; |
| 177 |
} elseif (!$this->options_mention_keywords($options)) { |
| 178 |
if (!empty($metadata['focus_keywords']) && is_array($metadata['focus_keywords'])) { |
| 179 |
$raw = $metadata['focus_keywords']; |
| 180 |
} elseif (isset($metadata['focus_keyword']) && is_string($metadata['focus_keyword']) && $metadata['focus_keyword'] !== '') { |
| 181 |
$raw = [$metadata['focus_keyword']]; |
| 182 |
} |
| 183 |
} |
| 184 |
|
| 185 |
$seen = []; |
| 186 |
$keywords = []; |
| 187 |
foreach ($raw as $keyword) { |
| 188 |
$keyword = trim((string) $keyword); |
| 189 |
if ($keyword === '') { |
| 190 |
continue; |
| 191 |
} |
| 192 |
$key = strtolower($keyword); |
| 193 |
if (isset($seen[$key])) { |
| 194 |
continue; |
| 195 |
} |
| 196 |
$seen[$key] = true; |
| 197 |
$keywords[] = $keyword; |
| 198 |
} |
| 199 |
|
| 200 |
return $keywords; |
| 201 |
} |
| 202 |
|
| 203 |
/** |
| 204 |
* Whether the caller said anything about keywords — including saying |
| 205 |
* "none". Distinguishes an intentional clear (score with no keyword) from |
| 206 |
* silence (fall back to the post's stored focus keywords). |
| 207 |
* |
| 208 |
* @param array $options Scoring options. |
| 209 |
* @return bool True when a keyword option key is present. |
| 210 |
*/ |
| 211 |
private function options_mention_keywords(array $options): bool { |
| 212 |
return array_key_exists('target_keywords', $options) |
| 213 |
|| array_key_exists('target_keyword', $options); |
| 214 |
} |
| 215 |
|
| 216 |
/** |
| 217 |
* Score each keyword independently and return the highest-scoring result. |
| 218 |
* |
| 219 |
* @param array $content_data Content analysis data. |
| 220 |
* @param array $metadata Post metadata. |
| 221 |
* @param string[] $keywords Target keywords (already normalized, 2+). |
| 222 |
* @param array $options Additional options. |
| 223 |
* @return array Best scoring result, with per-keyword data attached. |
| 224 |
*/ |
| 225 |
private function calculate_score_multi(array $content_data, array $metadata, array $keywords, array $options): array { |
| 226 |
$per_keyword = []; |
| 227 |
$best = null; |
| 228 |
$best_keyword = $keywords[0]; |
| 229 |
|
| 230 |
foreach ($keywords as $keyword) { |
| 231 |
$opts = $options; |
| 232 |
unset($opts['target_keywords']); |
| 233 |
$opts['target_keyword'] = $keyword; |
| 234 |
|
| 235 |
$result = $this->compute_score($content_data, $metadata, $opts); |
| 236 |
|
| 237 |
$per_keyword[] = [ |
| 238 |
'keyword' => $keyword, |
| 239 |
'overall_score' => $result['overall_score'], |
| 240 |
'grade' => $result['grade'], |
| 241 |
'score_breakdown' => $result['score_breakdown'], |
| 242 |
]; |
| 243 |
|
| 244 |
if ($best === null || $result['overall_score'] > $best['overall_score']) { |
| 245 |
$best = $result; |
| 246 |
$best_keyword = $keyword; |
| 247 |
} |
| 248 |
} |
| 249 |
|
| 250 |
// Final score = highest individual keyword score. Retain per-keyword |
| 251 |
// results and OR-combined checks so the UI can show both. |
| 252 |
$best['target_keyword'] = $best_keyword; |
| 253 |
$best['target_keywords'] = $keywords; |
| 254 |
$best['keyword_results'] = $per_keyword; |
| 255 |
$best['keyword_checks'] = $this->analyze_keyword_checks($content_data, $metadata, $keywords); |
| 256 |
$best['keywords'] = $this->keyword_placements($content_data, $metadata, $keywords); |
| 257 |
|
| 258 |
return $best; |
| 259 |
} |
| 260 |
|
| 261 |
/** |
| 262 |
* Evaluate per-location keyword checks across ALL focus keywords. |
| 263 |
* |
| 264 |
* Each check (title, meta description, content, image alt, slug) passes when |
| 265 |
* ANY of the focus keywords matches that location. |
| 266 |
* |
| 267 |
* @param array $content_data Content analysis data. |
| 268 |
* @param array $metadata Post metadata. |
| 269 |
* @param string[] $keywords Target keywords. |
| 270 |
* @return array<string,array{passed:bool,matched_keywords:string[]}> |
| 271 |
*/ |
| 272 |
private function analyze_keyword_checks(array $content_data, array $metadata, array $keywords): array { |
| 273 |
$title = self::lower((string) ($metadata['title'] ?? $content_data['title'] ?? '')); |
| 274 |
$description = self::lower((string) ($metadata['description'] ?? '')); |
| 275 |
$content = self::lower(self::plain_text((string) ($content_data['content'] ?? ''))); |
| 276 |
|
| 277 |
$alts = ''; |
| 278 |
foreach ((array) ($content_data['images'] ?? []) as $image) { |
| 279 |
$alts .= ' ' . self::lower((string) ($image['alt'] ?? '')); |
| 280 |
} |
| 281 |
|
| 282 |
// Build a searchable slug haystack from the post's OWN slug — never the |
| 283 |
// full URL path. The path carries ancestors, category bases and date |
| 284 |
// segments, so a child of /clinical-trials/ reported "keyword in slug" |
| 285 |
// for a page actually slugged `contact-us`. It also breaks the other |
| 286 |
// way: an unpublished post has no pretty permalink (get_permalink() |
| 287 |
// returns ?p=123), so the path held no slug at all and every draft |
| 288 |
// scored "no match" until it was published. Hyphens/underscores become |
| 289 |
// spaces so multi-word keywords can match. |
| 290 |
$slug = self::lower(self::slug_haystack($content_data)); |
| 291 |
|
| 292 |
$haystacks = [ |
| 293 |
'title' => trim($title), |
| 294 |
'meta_description' => trim($description), |
| 295 |
'content' => trim($content), |
| 296 |
'image_alt' => trim($alts), |
| 297 |
'slug' => trim($slug), |
| 298 |
]; |
| 299 |
|
| 300 |
$checks = []; |
| 301 |
foreach ($haystacks as $location => $haystack) { |
| 302 |
$matched = []; |
| 303 |
foreach ($keywords as $keyword) { |
| 304 |
if ($this->keyword_matches($haystack, self::lower(trim($keyword)))) { |
| 305 |
$matched[] = $keyword; |
| 306 |
} |
| 307 |
} |
| 308 |
$checks[$location] = [ |
| 309 |
'passed' => !empty($matched), |
| 310 |
'matched_keywords' => $matched, |
| 311 |
]; |
| 312 |
} |
| 313 |
|
| 314 |
return $checks; |
| 315 |
} |
| 316 |
|
| 317 |
/** |
| 318 |
* The keyword placements the editor draws one gauge segment for, in the |
| 319 |
* order a reader meets them (#729). |
| 320 |
* |
| 321 |
* @since 2.11.0 |
| 322 |
* @var string[] |
| 323 |
*/ |
| 324 |
public const PLACEMENTS = ['title', 'meta_description', 'slug', 'first_paragraph', 'subheading', 'content', 'image_alt']; |
| 325 |
|
| 326 |
/** |
| 327 |
* Characters of plain text read as the opening when the content has no |
| 328 |
* paragraph tag. |
| 329 |
* |
| 330 |
* @since 2.11.0 |
| 331 |
*/ |
| 332 |
private const OPENING_CHARS = 300; |
| 333 |
|
| 334 |
/** |
| 335 |
* Where each focus keyword is placed, keyword by keyword (#729). |
| 336 |
* |
| 337 |
* analyze_keyword_checks() answers "does ANY keyword appear here" for |
| 338 |
* five places; this answers "where does THIS keyword appear" for seven, |
| 339 |
* so the editor can show each keyword's own gauge. Same matcher, so a |
| 340 |
* keyword counts as a word (not a fragment) and a keyword in a script |
| 341 |
* written without spaces (Thai, Chinese, Japanese) still matches. |
| 342 |
* |
| 343 |
* `where` names the heading or alt text that matched, so the editor can |
| 344 |
* say which one. |
| 345 |
* |
| 346 |
* @since 2.11.0 |
| 347 |
* |
| 348 |
* @param array $content_data Content analysis data. |
| 349 |
* @param array $metadata Post metadata (title, description). |
| 350 |
* @param string[] $keywords Focus keywords. |
| 351 |
* @return array<int, array{keyword: string, passed: int, total: int, placements: array<string, array{passed: bool, where: string}>}> |
| 352 |
*/ |
| 353 |
public function keyword_placements(array $content_data, array $metadata, array $keywords): array { |
| 354 |
$html = (string) ($content_data['content'] ?? ''); |
| 355 |
$plain = self::lower(self::plain_text($html)); |
| 356 |
|
| 357 |
$headings = []; |
| 358 |
$source = isset($content_data['headings']) && is_array($content_data['headings']) ? $content_data['headings'] : $this->extract_headings($html); |
| 359 |
foreach ($source as $heading) { |
| 360 |
$text = trim((string) ($heading['text'] ?? '')); |
| 361 |
if ((int) ($heading['level'] ?? 0) >= 2 && '' !== $text) { |
| 362 |
$headings[] = $text; |
| 363 |
} |
| 364 |
} |
| 365 |
|
| 366 |
$alts = []; |
| 367 |
foreach ((array) ($content_data['images'] ?? []) as $image) { |
| 368 |
$alt = trim((string) ($image['alt'] ?? '')); |
| 369 |
if ('' !== $alt) { |
| 370 |
$alts[] = $alt; |
| 371 |
} |
| 372 |
} |
| 373 |
|
| 374 |
$single = [ |
| 375 |
'title' => self::lower((string) ($metadata['title'] ?? $content_data['title'] ?? '')), |
| 376 |
'meta_description' => self::lower((string) ($metadata['description'] ?? '')), |
| 377 |
'slug' => self::lower(self::slug_haystack($content_data)), |
| 378 |
'first_paragraph' => self::lower(self::opening($html)), |
| 379 |
'content' => $plain, |
| 380 |
]; |
| 381 |
|
| 382 |
$out = []; |
| 383 |
foreach ($keywords as $keyword) { |
| 384 |
$needle = self::lower(trim((string) $keyword)); |
| 385 |
$placements = []; |
| 386 |
|
| 387 |
foreach (self::PLACEMENTS as $placement) { |
| 388 |
if ('subheading' === $placement || 'image_alt' === $placement) { |
| 389 |
$where = ''; |
| 390 |
foreach ('subheading' === $placement ? $headings : $alts as $text) { |
| 391 |
if ($this->keyword_matches(self::lower($text), $needle)) { |
| 392 |
$where = $text; |
| 393 |
break; |
| 394 |
} |
| 395 |
} |
| 396 |
$placements[$placement] = ['passed' => '' !== $where, 'where' => $where]; |
| 397 |
continue; |
| 398 |
} |
| 399 |
|
| 400 |
$placements[$placement] = ['passed' => $this->keyword_matches($single[$placement], $needle), 'where' => '']; |
| 401 |
} |
| 402 |
|
| 403 |
$out[] = [ |
| 404 |
'keyword' => (string) $keyword, |
| 405 |
'passed' => count(array_filter(array_column($placements, 'passed'))), |
| 406 |
'total' => count(self::PLACEMENTS), |
| 407 |
'placements' => $placements, |
| 408 |
]; |
| 409 |
} |
| 410 |
|
| 411 |
return $out; |
| 412 |
} |
| 413 |
|
| 414 |
/** |
| 415 |
* The post's own slug as searchable text: hyphens and underscores become |
| 416 |
* spaces so a multi-word keyword can match, and a slug WordPress |
| 417 |
* percent-encoded (Thai, Cyrillic, Chinese) is decoded, or it could never |
| 418 |
* match a keyword typed in that script. |
| 419 |
* |
| 420 |
* Never the full URL path: the path carries ancestors, category bases and |
| 421 |
* date segments, so a child of /clinical-trials/ reported "keyword in slug" |
| 422 |
* for a page actually slugged `contact-us`. An unpublished post has no |
| 423 |
* pretty permalink either, so the path held no slug at all. |
| 424 |
* |
| 425 |
* @param array $content_data Content analysis data. |
| 426 |
* @return string |
| 427 |
*/ |
| 428 |
private static function slug_haystack(array $content_data): string { |
| 429 |
$slug = (string) ($content_data['slug'] ?? ''); |
| 430 |
if ('' === $slug) { |
| 431 |
// Draft with no slug assigned yet: score what WordPress would |
| 432 |
// generate from the title, which is what the editor shows as the |
| 433 |
// proposed URL — so the check reads the same before and after |
| 434 |
// publishing instead of flipping. |
| 435 |
$slug = sanitize_title((string) ($content_data['title'] ?? '')); |
| 436 |
} |
| 437 |
|
| 438 |
return trim(str_replace(['-', '_'], ' ', rawurldecode($slug))); |
| 439 |
} |
| 440 |
|
| 441 |
/** |
| 442 |
* The opening of the content: its first paragraph with text, or the first |
| 443 |
* few hundred characters when it has none. Counted in characters, not |
| 444 |
* words, so a language written without spaces is not read as one word. |
| 445 |
* |
| 446 |
* @param string $html Content HTML. |
| 447 |
* @return string Plain text. |
| 448 |
*/ |
| 449 |
private static function opening(string $html): string { |
| 450 |
if (preg_match_all('/<p\b[^>]*>(.*?)<\/p>/isu', $html, $matches)) { |
| 451 |
foreach ($matches[1] as $paragraph) { |
| 452 |
$text = self::collapse_whitespace(wp_strip_all_tags($paragraph)); |
| 453 |
if ('' !== $text) { |
| 454 |
return $text; |
| 455 |
} |
| 456 |
} |
| 457 |
} |
| 458 |
|
| 459 |
$text = self::plain_text($html); |
| 460 |
|
| 461 |
return function_exists('mb_substr') ? mb_substr($text, 0, self::OPENING_CHARS) : substr($text, 0, self::OPENING_CHARS); |
| 462 |
} |
| 463 |
|
| 464 |
/** |
| 465 |
* Content as plain text, with a space where each tag was. wp_strip_all_tags() |
| 466 |
* alone joins neighbouring blocks — "…coffee grinder</h3><p>A good…" became |
| 467 |
* "coffee grinderA good" — so a keyword at the end of a heading or a |
| 468 |
* paragraph was no longer a word and did not match. |
| 469 |
* |
| 470 |
* @param string $html Content HTML. |
| 471 |
* @return string |
| 472 |
*/ |
| 473 |
private static function plain_text(string $html): string { |
| 474 |
$spaced = preg_replace('/<[^>]+>/', ' $0 ', $html); |
| 475 |
|
| 476 |
return self::collapse_whitespace(wp_strip_all_tags(null === $spaced ? $html : (string) $spaced)); |
| 477 |
} |
| 478 |
|
| 479 |
/** |
| 480 |
* Runs of whitespace down to one space. |
| 481 |
* |
| 482 |
* The `/u` pass is the one that understands a multibyte space, but |
| 483 |
* preg_replace() answers null on bytes that are not valid UTF-8 rather than |
| 484 |
* throwing — and casting that null to a string blanked the haystack, so a |
| 485 |
* post carrying one mojibake byte (a Latin-1 paste, an old import) reported |
| 486 |
* every keyword as missing from its body, its opening, and every |
| 487 |
* subheading. The gauge said 0/7 and told the author to add a keyword that |
| 488 |
* was already there. |
| 489 |
* |
| 490 |
* Falls back to the byte-wise collapse, which is what this did before the |
| 491 |
* multibyte work added the modifier. Same reasoning keyword_matches() |
| 492 |
* already records for its own PCRE failure: a pattern PCRE refuses must not |
| 493 |
* be reported as a confident "no match". |
| 494 |
* |
| 495 |
* @param string $text Text to collapse. |
| 496 |
* @return string |
| 497 |
*/ |
| 498 |
private static function collapse_whitespace(string $text): string { |
| 499 |
$collapsed = preg_replace('/\s+/u', ' ', $text); |
| 500 |
|
| 501 |
if (null === $collapsed) { |
| 502 |
$collapsed = preg_replace('/\s+/', ' ', $text); |
| 503 |
} |
| 504 |
|
| 505 |
return trim(null === $collapsed ? $text : (string) $collapsed); |
| 506 |
} |
| 507 |
|
| 508 |
/** |
| 509 |
* Lowercase in any script. strtolower() only folds ASCII, so "Кофе" never |
| 510 |
* matched "кофе". |
| 511 |
* |
| 512 |
* @param string $text Text. |
| 513 |
* @return string |
| 514 |
*/ |
| 515 |
private static function lower(string $text): string { |
| 516 |
return function_exists('mb_strtolower') ? mb_strtolower($text, 'UTF-8') : strtolower($text); |
| 517 |
} |
| 518 |
|
| 519 |
/** |
| 520 |
* Scripts written without spaces between words. |
| 521 |
* |
| 522 |
* @since 2.1.0 |
| 523 |
* @var string |
| 524 |
*/ |
| 525 |
private const SCRIPTIO_CONTINUA = '/[\p{Han}\p{Hiragana}\p{Katakana}\p{Thai}\p{Lao}\p{Khmer}\p{Myanmar}]/u'; |
| 526 |
|
| 527 |
/** |
| 528 |
* Whether a keyword appears in a haystack as a word rather than as a |
| 529 |
* fragment of a longer one. |
| 530 |
* |
| 531 |
* The five keyword checks used a plain strpos(), so any substring hit |
| 532 |
* counted: "test coronavirus" matched "la|test coronavirus|news", "art" |
| 533 |
* matched "start", "cat" matched "category". The panel then confidently |
| 534 |
* reported a keyword placement that does not exist (#416). Same class of |
| 535 |
* problem #71 fixed in the Image SEO rewriter, and the same remedy. |
| 536 |
* |
| 537 |
* Both arguments are expected lowercased already. |
| 538 |
* |
| 539 |
* @since 2.1.0 |
| 540 |
* |
| 541 |
* @param string $haystack Text to search. |
| 542 |
* @param string $needle Keyword, lowercased and trimmed. |
| 543 |
* @return bool |
| 544 |
*/ |
| 545 |
private function keyword_matches(string $haystack, string $needle): bool { |
| 546 |
if ($needle === '' || $haystack === '') { |
| 547 |
return false; |
| 548 |
} |
| 549 |
|
| 550 |
if (!$this->supports_word_boundaries($needle)) { |
| 551 |
return strpos($haystack, $needle) !== false; |
| 552 |
} |
| 553 |
|
| 554 |
$matched = preg_match('/\b' . preg_quote($needle, '/') . '\b/u', $haystack); |
| 555 |
|
| 556 |
// PCRE refusing the pattern — invalid UTF-8 in the keyword, a |
| 557 |
// backtrack limit — must not be reported as a confident "no match". |
| 558 |
// Fall back to the behaviour this replaced rather than invent a |
| 559 |
// negative the user cannot explain. |
| 560 |
if ($matched === false) { |
| 561 |
return strpos($haystack, $needle) !== false; |
| 562 |
} |
| 563 |
|
| 564 |
return $matched === 1; |
| 565 |
} |
| 566 |
|
| 567 |
/** |
| 568 |
* Whether \b can express "this keyword, as a word" for this keyword. |
| 569 |
* |
| 570 |
* It asserts a transition between a word and a non-word character, which |
| 571 |
* only means something where words are separated. Two cases where it is |
| 572 |
* not, both verified against PCRE rather than assumed: |
| 573 |
* |
| 574 |
* - the keyword's own edges are not word characters ("c++", "#seo"), so |
| 575 |
* no boundary can assert there and a real match is lost; |
| 576 |
* - scripts written without spaces, where the neighbouring characters |
| 577 |
* are word characters too — "冠状� |
| 578 |
毒" inside "最新冠状� |
| 579 |
毒新闻" is a |
| 580 |
* legitimate match that \b never sees. |
| 581 |
* |
| 582 |
* Accented Latin and Cyrillic need no special handling: PHP's /u modifier |
| 583 |
* turns on Unicode character properties, so "café" correctly does not |
| 584 |
* match "cafés" and "коронавирус" does not match "коронавирусный". |
| 585 |
* |
| 586 |
* @since 2.1.0 |
| 587 |
* |
| 588 |
* @param string $needle Keyword, lowercased and trimmed. |
| 589 |
* @return bool |
| 590 |
*/ |
| 591 |
private function supports_word_boundaries(string $needle): bool { |
| 592 |
if (preg_match(self::SCRIPTIO_CONTINUA, $needle)) { |
| 593 |
return false; |
| 594 |
} |
| 595 |
|
| 596 |
return preg_match('/^\w/u', $needle) === 1 && preg_match('/\w$/u', $needle) === 1; |
| 597 |
} |
| 598 |
|
| 599 |
/** |
| 600 |
* Compute the SEO score for a single target keyword. |
| 601 |
* |
| 602 |
* @param array $content_data Content analysis data |
| 603 |
* @param array $metadata Post metadata |
| 604 |
* @param array $options Additional options (expects scalar target_keyword) |
| 605 |
* @return array Complete scoring result |
| 606 |
* |
| 607 |
* @throws \Exception On failure. |
| 608 |
*/ |
| 609 |
private function compute_score(array $content_data, array $metadata, array $options = []): array { |
| 610 |
$scores = []; |
| 611 |
$suggestions = []; |
| 612 |
$total_score = 0; |
| 613 |
|
| 614 |
try { |
| 615 |
// 1. Satisfying Content (23 points) - #1 factor in 2025 |
| 616 |
$satisfying_result = $this->score_satisfying_content($content_data, $options['target_keyword'] ?? '', $metadata); |
| 617 |
$scores['satisfying_content'] = $satisfying_result; |
| 618 |
$total_score += $satisfying_result['score']; |
| 619 |
$suggestions = array_merge($suggestions, $satisfying_result['suggestions']); |
| 620 |
} catch (\Exception $e) { |
| 621 |
throw $e; |
| 622 |
} |
| 623 |
|
| 624 |
try { |
| 625 |
// 2. Title Optimization (14 points) - Looser keyword matching in 2025 |
| 626 |
$title_result = $this->score_2025_title_optimization($metadata['title'] ?? '', $options['target_keyword'] ?? ''); |
| 627 |
$scores['title_optimization'] = $title_result; |
| 628 |
$total_score += $title_result['score']; |
| 629 |
$suggestions = array_merge($suggestions, $title_result['suggestions']); |
| 630 |
} catch (\Exception $e) { |
| 631 |
throw $e; |
| 632 |
} |
| 633 |
|
| 634 |
try { |
| 635 |
// 3. Niche Expertise (13 points) - Hub & spoke content clusters |
| 636 |
$expertise_result = $this->score_niche_expertise($content_data, $options['target_keyword'] ?? ''); |
| 637 |
$scores['niche_expertise'] = $expertise_result; |
| 638 |
$total_score += $expertise_result['score']; |
| 639 |
$suggestions = array_merge($suggestions, $expertise_result['suggestions']); |
| 640 |
} catch (\Exception $e) { |
| 641 |
throw $e; |
| 642 |
} |
| 643 |
|
| 644 |
try { |
| 645 |
// 4. Searcher Engagement (12 points) - Dwell time, bounce rate, pages/session |
| 646 |
$engagement_result = $this->score_searcher_engagement($content_data); |
| 647 |
$scores['searcher_engagement'] = $engagement_result; |
| 648 |
$total_score += $engagement_result['score']; |
| 649 |
$suggestions = array_merge($suggestions, $engagement_result['suggestions']); |
| 650 |
} catch (\Exception $e) { |
| 651 |
throw $e; |
| 652 |
} |
| 653 |
|
| 654 |
try { |
| 655 |
// 5. Backlink Authority (13 points) - Quality backlinks |
| 656 |
$backlink_result = $this->score_backlink_authority($content_data); |
| 657 |
$scores['backlink_authority'] = $backlink_result; |
| 658 |
$total_score += $backlink_result['score']; |
| 659 |
$suggestions = array_merge($suggestions, $backlink_result['suggestions']); |
| 660 |
} catch (\Exception $e) { |
| 661 |
throw $e; |
| 662 |
} |
| 663 |
|
| 664 |
// 6. Content Freshness (6 points) - Quarterly updates priority |
| 665 |
$freshness_result = $this->score_content_freshness($content_data); |
| 666 |
$scores['content_freshness'] = $freshness_result; |
| 667 |
$total_score += $freshness_result['score']; |
| 668 |
$suggestions = array_merge($suggestions, $freshness_result['suggestions']); |
| 669 |
|
| 670 |
// 7. Mobile Experience (5 points) - NEW: Mobile Experience Score (MES) |
| 671 |
$mobile_result = $this->score_mobile_experience($content_data); |
| 672 |
$scores['mobile_experience'] = $mobile_result; |
| 673 |
$total_score += $mobile_result['score']; |
| 674 |
$suggestions = array_merge($suggestions, $mobile_result['suggestions']); |
| 675 |
|
| 676 |
// 8. Trustworthiness (4 points) - E-E-A-T verification |
| 677 |
$trust_result = $this->score_trustworthiness($content_data, $metadata); |
| 678 |
$scores['trustworthiness'] = $trust_result; |
| 679 |
$total_score += $trust_result['score']; |
| 680 |
$suggestions = array_merge($suggestions, $trust_result['suggestions']); |
| 681 |
|
| 682 |
// 9. Link Diversity (3 points) - Multiple pages with backlinks |
| 683 |
$diversity_result = $this->score_link_diversity($content_data); |
| 684 |
$scores['link_diversity'] = $diversity_result; |
| 685 |
$total_score += $diversity_result['score']; |
| 686 |
$suggestions = array_merge($suggestions, $diversity_result['suggestions']); |
| 687 |
|
| 688 |
// 10. Core Web Vitals (3 points) - Interaction Readiness + CLS 2.0 |
| 689 |
$vitals_result = $this->score_core_web_vitals($content_data); |
| 690 |
$scores['core_web_vitals'] = $vitals_result; |
| 691 |
$total_score += $vitals_result['score']; |
| 692 |
$suggestions = array_merge($suggestions, $vitals_result['suggestions']); |
| 693 |
|
| 694 |
// 11. Site Security (2 points) - SSL certificate |
| 695 |
$security_result = $this->score_site_security($content_data); |
| 696 |
$scores['site_security'] = $security_result; |
| 697 |
$total_score += $security_result['score']; |
| 698 |
$suggestions = array_merge($suggestions, $security_result['suggestions']); |
| 699 |
|
| 700 |
// 12. Internal Linking (1 point) - Declining importance |
| 701 |
$internal_result = $this->score_internal_linking($content_data); |
| 702 |
$scores['internal_linking'] = $internal_result; |
| 703 |
$total_score += $internal_result['score']; |
| 704 |
$suggestions = array_merge($suggestions, $internal_result['suggestions']); |
| 705 |
|
| 706 |
// 13. Technical Factors (1 point) - Meta descriptions, schema, etc. |
| 707 |
$technical_result = $this->score_technical_factors($content_data, $metadata, $options); |
| 708 |
$scores['technical_factors'] = $technical_result; |
| 709 |
$total_score += $technical_result['score']; |
| 710 |
$suggestions = array_merge($suggestions, $technical_result['suggestions']); |
| 711 |
|
| 712 |
try { |
| 713 |
$prioritized_suggestions = $this->prioritize_suggestions($suggestions, $scores); |
| 714 |
$grade = $this->get_grade_from_score($total_score); |
| 715 |
|
| 716 |
return [ |
| 717 |
'overall_score' => min(100, $total_score), |
| 718 |
'score_breakdown' => $scores, |
| 719 |
'suggestions' => $prioritized_suggestions, |
| 720 |
'grade' => $grade, |
| 721 |
// Readability + content-quality labels so persisted scores |
| 722 |
// (e.g. bulk-analyzed on import) populate the post-list columns |
| 723 |
// without a manual re-analyze. The REST endpoint still overrides |
| 724 |
// these with live-editor values when the metabox provides them. |
| 725 |
'readability_score' => $this->format_readability_label($content_data), |
| 726 |
'content_quality' => $this->derive_content_quality($content_data), |
| 727 |
'calculated_at' => current_time('mysql'), |
| 728 |
'algorithm_version' => '2025.2', |
| 729 |
'algorithm_source' => 'First Page Sage Q1 2025 Research', |
| 730 |
'factors_count' => count($scores), |
| 731 |
]; |
| 732 |
} catch (\Exception $e) { |
| 733 |
throw $e; |
| 734 |
} |
| 735 |
} |
| 736 |
|
| 737 |
/** |
| 738 |
* Score satisfying content - #1 factor in 2025 (23 points) |
| 739 |
* Google tests content to see if it satisfies search intent |
| 740 |
* |
| 741 |
* @param array $content_data Content analysis data |
| 742 |
* @param string $target_keyword Target keyword |
| 743 |
* @param array $metadata Post metadata (title, description) |
| 744 |
* @return array Scoring result |
| 745 |
*/ |
| 746 |
private function score_satisfying_content(array $content_data, string $target_keyword, array $metadata = []): array { |
| 747 |
$score = 0; |
| 748 |
$max_score = $this->scoring_factors['satisfying_content']; |
| 749 |
$suggestions = []; |
| 750 |
|
| 751 |
$content = $content_data['content'] ?? ''; |
| 752 |
$word_count = $content_data['word_count'] ?? 0; |
| 753 |
$meta_description = (string) ($metadata['description'] ?? ''); |
| 754 |
|
| 755 |
// Content depth and comprehensiveness (8 points). Tiers softened so a |
| 756 |
// genuinely useful post is not capped the way the old 2000-word gate did |
| 757 |
// (Rank Math awards full content credit well below 2000 words). |
| 758 |
if ($word_count >= 1500) { |
| 759 |
$score += 8; |
| 760 |
} elseif ($word_count >= 1000) { |
| 761 |
$score += 7; |
| 762 |
$suggestions[] = 'Consider expanding content to 1500+ words for more comprehensive coverage'; |
| 763 |
} elseif ($word_count >= 600) { |
| 764 |
$score += 5; |
| 765 |
$suggestions[] = 'Content is adequate - aim for 1000+ words for stronger topic depth'; |
| 766 |
} elseif ($word_count >= 300) { |
| 767 |
$score += 3; |
| 768 |
$suggestions[] = 'Content is thin - aim for 600+ words minimum'; |
| 769 |
} else { |
| 770 |
$score++; |
| 771 |
$suggestions[] = 'Content too shallow - Google prioritizes comprehensive, satisfying content'; |
| 772 |
} |
| 773 |
|
| 774 |
// Keyword presence & placement (8 points) - deterministic, replaces the |
| 775 |
// old literal-phrase intent heuristic ("what is"/"because"). Measures |
| 776 |
// signals the editor actually controls: body (3), first paragraph (3), |
| 777 |
// meta description (2) - the last mirrors Rank Math's "keyword in meta |
| 778 |
// description" basic-SEO check. |
| 779 |
if (empty($target_keyword)) { |
| 780 |
$score += 4; // Benefit of the doubt when no focus keyword is set. |
| 781 |
$suggestions[] = 'Set a focus keyword so content relevance can be measured'; |
| 782 |
} else { |
| 783 |
if ($this->keyword_in_content($content, $target_keyword)) { |
| 784 |
$score += 3; |
| 785 |
} else { |
| 786 |
$suggestions[] = "Use the focus keyword '{$target_keyword}' in the body content"; |
| 787 |
} |
| 788 |
if ($this->keyword_in_first_paragraph($content, $target_keyword)) { |
| 789 |
$score += 3; |
| 790 |
} else { |
| 791 |
$suggestions[] = "Mention '{$target_keyword}' near the start of the content (first paragraph)"; |
| 792 |
} |
| 793 |
if ($this->keyword_in_meta($meta_description, $target_keyword)) { |
| 794 |
$score += 2; |
| 795 |
} else { |
| 796 |
$suggestions[] = "Include the focus keyword '{$target_keyword}' in the meta description"; |
| 797 |
} |
| 798 |
} |
| 799 |
|
| 800 |
// Content structure & value (7 points) - reuses the deterministic |
| 801 |
// content-quality signal (length, paragraph length, subheading |
| 802 |
// distribution) instead of the noisy sentence-length variety heuristic. |
| 803 |
$quality = $this->derive_content_quality_score($content_data); // 0-100 |
| 804 |
$score += (int) round(($quality / 100) * 7); |
| 805 |
|
| 806 |
if ($quality < 60) { |
| 807 |
$suggestions[] = 'Improve content structure - break up long paragraphs and add subheadings'; |
| 808 |
} |
| 809 |
|
| 810 |
return [ |
| 811 |
'score' => $score, |
| 812 |
'max_score' => $max_score, |
| 813 |
'suggestions' => $suggestions, |
| 814 |
'details' => [ |
| 815 |
'word_count' => $word_count, |
| 816 |
'keyword_in_content' => !empty($target_keyword) && $this->keyword_in_content($content, $target_keyword), |
| 817 |
'keyword_in_first_paragraph' => !empty($target_keyword) && $this->keyword_in_first_paragraph($content, $target_keyword), |
| 818 |
'keyword_in_meta_description' => !empty($target_keyword) && $this->keyword_in_meta($meta_description, $target_keyword), |
| 819 |
'content_quality_score' => $quality, |
| 820 |
'content_depth' => $this->assess_content_depth_2025($word_count), |
| 821 |
] |
| 822 |
]; |
| 823 |
} |
| 824 |
|
| 825 |
/** |
| 826 |
* Assess content depth for 2025 standards |
| 827 |
* |
| 828 |
* @param int $word_count Word count |
| 829 |
* @return string Depth assessment |
| 830 |
*/ |
| 831 |
private function assess_content_depth_2025(int $word_count): string { |
| 832 |
if ($word_count >= 3000) { return 'Comprehensive'; |
| 833 |
} |
| 834 |
if ($word_count >= 2000) { return 'Detailed'; |
| 835 |
} |
| 836 |
if ($word_count >= 1200) { return 'Adequate'; |
| 837 |
} |
| 838 |
if ($word_count >= 800) { return 'Basic'; |
| 839 |
} |
| 840 |
return 'Insufficient'; |
| 841 |
} |
| 842 |
|
| 843 |
/** |
| 844 |
* Score 2025 title optimization with looser keyword matching |
| 845 |
* |
| 846 |
* @param string $title Post title |
| 847 |
* @param string $target_keyword Target keyword |
| 848 |
* @return array Scoring result |
| 849 |
*/ |
| 850 |
private function score_2025_title_optimization(string $title, string $target_keyword): array { |
| 851 |
$score = 0; |
| 852 |
$max_score = $this->scoring_factors['title_optimization']; |
| 853 |
$suggestions = []; |
| 854 |
|
| 855 |
if (empty($title)) { |
| 856 |
$suggestions[] = 'Add a compelling, click-worthy title that matches search intent'; |
| 857 |
return ['score' => 0, 'max_score' => $max_score, 'suggestions' => $suggestions]; |
| 858 |
} |
| 859 |
|
| 860 |
$title_length = mb_strlen($title); |
| 861 |
|
| 862 |
// 2025 length optimization (6 points). 60 characters is the recommended |
| 863 |
// maximum for best SERP visibility before Google truncates the title. |
| 864 |
if ($title_length >= self::TITLE_OPTIMAL_MIN && $title_length <= self::TITLE_OPTIMAL_MAX) { |
| 865 |
$score += 6; |
| 866 |
} elseif ($title_length >= 25 && $title_length <= 75) { |
| 867 |
$score += 4; |
| 868 |
$suggestions[] = 'Optimize title length to 35-60 characters for better SERP visibility'; |
| 869 |
} else { |
| 870 |
$score++; |
| 871 |
$suggestions[] = $title_length < 25 ? |
| 872 |
'Title too short - aim for 35-60 characters' : |
| 873 |
'Title too long - risk truncation in search results'; |
| 874 |
} |
| 875 |
|
| 876 |
// Looser keyword matching (6 points) - 2025 update |
| 877 |
if (!empty($target_keyword)) { |
| 878 |
$title_lower = strtolower($title); |
| 879 |
$keyword_lower = strtolower($target_keyword); |
| 880 |
|
| 881 |
// Exact match |
| 882 |
if (strpos($title_lower, $keyword_lower) !== false) { |
| 883 |
$score += 6; |
| 884 |
} else { |
| 885 |
// Check for semantic variations (2025 improvement) |
| 886 |
$semantic_match = $this->check_semantic_keyword_match($title, $target_keyword); |
| 887 |
if ($semantic_match) { |
| 888 |
$score += 5; // Almost full credit for semantic match |
| 889 |
$suggestions[] = 'Good semantic keyword usage - Google now recognizes keyword variations'; |
| 890 |
} else { |
| 891 |
// Check for partial keyword match |
| 892 |
$keyword_parts = explode(' ', $keyword_lower); |
| 893 |
$partial_matches = 0; |
| 894 |
foreach ($keyword_parts as $part) { |
| 895 |
if (strpos($title_lower, $part) !== false) { |
| 896 |
$partial_matches++; |
| 897 |
} |
| 898 |
} |
| 899 |
|
| 900 |
if ($partial_matches > 0) { |
| 901 |
$score += round(($partial_matches / count($keyword_parts)) * 4); |
| 902 |
$suggestions[] = "Include more parts of target keyword '{$target_keyword}' in title"; |
| 903 |
} else { |
| 904 |
$suggestions[] = "Include target keyword '{$target_keyword}' or related terms in title"; |
| 905 |
} |
| 906 |
} |
| 907 |
} |
| 908 |
} else { |
| 909 |
$score += 2; // Partial credit |
| 910 |
$suggestions[] = 'Set a target keyword to optimize title effectiveness'; |
| 911 |
} |
| 912 |
|
| 913 |
// Title readability (2 points) - mirrors Rank Math's title checks for a |
| 914 |
// number/power word (drives CTR) and emotional sentiment. |
| 915 |
$has_number = (bool) preg_match('/\d/', $title); |
| 916 |
$has_power_word = $this->title_has_power_word($title); |
| 917 |
$has_sentiment = $this->title_has_sentiment_word($title); |
| 918 |
|
| 919 |
if ($has_number || $has_power_word) { |
| 920 |
$score++; |
| 921 |
} else { |
| 922 |
$suggestions[] = 'Add a number or a power word to the title to boost click-through rate'; |
| 923 |
} |
| 924 |
if ($has_sentiment) { |
| 925 |
$score++; |
| 926 |
} else { |
| 927 |
$suggestions[] = 'Use an emotional/sentiment word in the title to make it more compelling'; |
| 928 |
} |
| 929 |
|
| 930 |
return [ |
| 931 |
'score' => $score, |
| 932 |
'max_score' => $max_score, |
| 933 |
'suggestions' => $suggestions, |
| 934 |
'details' => [ |
| 935 |
'title_length' => $title_length, |
| 936 |
'optimal_range' => '35-60 characters', |
| 937 |
'keyword_present' => !empty($target_keyword) && strpos(strtolower($title), strtolower($target_keyword)) !== false, |
| 938 |
'semantic_match' => !empty($target_keyword) ? $this->check_semantic_keyword_match($title, $target_keyword) : false, |
| 939 |
'has_number_or_power_word' => $has_number || $has_power_word, |
| 940 |
'has_sentiment_word' => $has_sentiment, |
| 941 |
] |
| 942 |
]; |
| 943 |
} |
| 944 |
|
| 945 |
// Placeholder methods for remaining 2025 factors |
| 946 |
|
| 947 |
private function score_niche_expertise(array $content_data, string $target_keyword): array { |
| 948 |
$max_score = $this->scoring_factors['niche_expertise']; |
| 949 |
$suggestions = []; |
| 950 |
|
| 951 |
// Without a focus keyword we cannot measure topical coverage; award |
| 952 |
// partial credit rather than capping the ceiling with a placeholder. |
| 953 |
if (empty($target_keyword)) { |
| 954 |
return [ |
| 955 |
'score' => 7, |
| 956 |
'max_score' => $max_score, |
| 957 |
'suggestions' => ['Set a focus keyword and use it in subheadings and the URL for stronger topical signals'], |
| 958 |
'details' => ['expertise_level' => 'Unmeasured (no focus keyword)'], |
| 959 |
]; |
| 960 |
} |
| 961 |
|
| 962 |
$score = 0; |
| 963 |
$content = (string) ($content_data['content'] ?? ''); |
| 964 |
$headings = (array) ($content_data['headings'] ?? []); |
| 965 |
$images = (array) ($content_data['images'] ?? []); |
| 966 |
$slug = strtolower((string) ($content_data['slug'] ?? '')); |
| 967 |
|
| 968 |
// Keyword in a subheading (4 points). |
| 969 |
if ($this->keyword_in_subheadings($headings, $target_keyword)) { |
| 970 |
$score += 4; |
| 971 |
} else { |
| 972 |
$suggestions[] = "Include '{$target_keyword}' in at least one subheading (H2-H6)"; |
| 973 |
} |
| 974 |
|
| 975 |
// Keyword density in a healthy band (4 points). Rank Math treats |
| 976 |
// ~0.5%-2.5% as optimal; reward in-band, partial when present but thin. |
| 977 |
$density = $this->keyword_density($content, $target_keyword); |
| 978 |
if ($density >= 0.5 && $density <= 2.5) { |
| 979 |
$score += 4; |
| 980 |
} elseif ($density > 0) { |
| 981 |
$score += 2; |
| 982 |
$suggestions[] = $density > 2.5 |
| 983 |
? 'Keyword density is high - reduce repetition to avoid over-optimization' |
| 984 |
: 'Keyword density is low - use the focus keyword a little more often'; |
| 985 |
} else { |
| 986 |
$suggestions[] = "Use the focus keyword '{$target_keyword}' in the content"; |
| 987 |
} |
| 988 |
|
| 989 |
// URL optimization (3 points): keyword in slug (2) + a reasonably short |
| 990 |
// URL (1). Rank Math flags overly long URLs, so reward concise slugs. |
| 991 |
$keyword_slug = str_replace(' ', '-', strtolower($target_keyword)); |
| 992 |
$keyword_in_slug = $slug !== '' && (strpos($slug, $keyword_slug) !== false || strpos(str_replace('-', '', $slug), str_replace('-', '', $keyword_slug)) !== false); |
| 993 |
if ($keyword_in_slug) { |
| 994 |
$score += 2; |
| 995 |
} else { |
| 996 |
$suggestions[] = 'Include the focus keyword in the URL slug'; |
| 997 |
} |
| 998 |
|
| 999 |
// A slug under ~75 chars keeps the URL clean and fully visible in SERPs. |
| 1000 |
$slug_length = strlen($slug); |
| 1001 |
if ($slug === '' || $slug_length <= 75) { |
| 1002 |
$score++; |
| 1003 |
} else { |
| 1004 |
$suggestions[] = 'Shorten the URL slug - long URLs are harder to read and share'; |
| 1005 |
} |
| 1006 |
|
| 1007 |
// Keyword in image alt text (2 points) - mirrors Rank Math's |
| 1008 |
// "keyword in image alt" check. When the post has no images the check |
| 1009 |
// does not apply, so award the points (benefit of the doubt) rather |
| 1010 |
// than capping the ceiling for legitimately image-less posts. |
| 1011 |
if (empty($images)) { |
| 1012 |
$score += 2; |
| 1013 |
$suggestions[] = 'Add a relevant image with the focus keyword in its alt text'; |
| 1014 |
} elseif ($this->keyword_in_alt($images, $target_keyword)) { |
| 1015 |
$score += 2; |
| 1016 |
} else { |
| 1017 |
$suggestions[] = 'Include the focus keyword in at least one image alt attribute'; |
| 1018 |
} |
| 1019 |
|
| 1020 |
return [ |
| 1021 |
'score' => $score, |
| 1022 |
'max_score' => $max_score, |
| 1023 |
'suggestions' => $suggestions, |
| 1024 |
'details' => [ |
| 1025 |
'keyword_in_subheading' => $this->keyword_in_subheadings($headings, $target_keyword), |
| 1026 |
'keyword_density' => round($density, 2), |
| 1027 |
'keyword_in_slug' => $keyword_in_slug, |
| 1028 |
'slug_length' => $slug_length, |
| 1029 |
'keyword_in_image_alt' => $this->keyword_in_alt($images, $target_keyword), |
| 1030 |
], |
| 1031 |
]; |
| 1032 |
} |
| 1033 |
|
| 1034 |
private function score_searcher_engagement(array $content_data): array { |
| 1035 |
$max_score = $this->scoring_factors['searcher_engagement']; |
| 1036 |
|
| 1037 |
// Engagement (dwell time / bounce) is off-page, so estimate it from the |
| 1038 |
// on-page signals that drive it: readability (half) + content structure |
| 1039 |
// (half). This raises the old readability-only floor. |
| 1040 |
$readability = (float) ($content_data['readability_score'] ?? 50); |
| 1041 |
$structure = (float) $this->derive_content_quality_score($content_data); |
| 1042 |
|
| 1043 |
$readability_pts = ($readability / 100) * ($max_score / 2); |
| 1044 |
$structure_pts = ($structure / 100) * ($max_score / 2); |
| 1045 |
$score = (int) round($readability_pts + $structure_pts); |
| 1046 |
|
| 1047 |
return [ |
| 1048 |
'score' => $score, |
| 1049 |
'max_score' => $max_score, |
| 1050 |
'suggestions' => $score < ($max_score * 0.7) |
| 1051 |
? ['Improve readability and structure (shorter sentences, subheadings, shorter paragraphs)'] |
| 1052 |
: [], |
| 1053 |
'details' => ['engagement_estimate' => round(($score / $max_score) * 100, 1) . '%'], |
| 1054 |
]; |
| 1055 |
} |
| 1056 |
|
| 1057 |
private function score_backlink_authority(array $content_data): array { |
| 1058 |
$max_score = $this->scoring_factors['backlink_authority']; |
| 1059 |
$external_links = (int) ($content_data['external_links'] ?? 0); |
| 1060 |
$dofollow_links = (int) ($content_data['external_dofollow_links'] ?? 0); |
| 1061 |
|
| 1062 |
// Backlinks are off-page and cannot be read from post content. We use |
| 1063 |
// the only on-page proxy available - whether the content cites external |
| 1064 |
// sources - and avoid hard-penalizing posts for something outside the |
| 1065 |
// editor's control (the old external_links*3 formula needed 5 outbound |
| 1066 |
// links just to reach full marks, dragging nearly every post down). |
| 1067 |
// Dofollow links pass equity, so they earn full credit (Rank Math's |
| 1068 |
// "external dofollow link" check); nofollow-only citations earn less. |
| 1069 |
$suggestions = []; |
| 1070 |
if ($dofollow_links >= 2) { |
| 1071 |
$score = $max_score; |
| 1072 |
} elseif ($dofollow_links === 1) { |
| 1073 |
$score = (int) round($max_score * 0.85); |
| 1074 |
} elseif ($external_links > 0) { |
| 1075 |
// Cites sources but every external link is nofollow. |
| 1076 |
$score = (int) round($max_score * 0.75); |
| 1077 |
$suggestions[] = 'Add at least one dofollow link to an authoritative external source'; |
| 1078 |
} else { |
| 1079 |
$score = (int) round($max_score * 0.6); |
| 1080 |
$suggestions[] = 'Cite authoritative external sources, and build quality backlinks to this page'; |
| 1081 |
} |
| 1082 |
|
| 1083 |
return [ |
| 1084 |
'score' => $score, |
| 1085 |
'max_score' => $max_score, |
| 1086 |
'suggestions' => $suggestions, |
| 1087 |
'details' => [ |
| 1088 |
'external_links' => $external_links, |
| 1089 |
'external_dofollow_links' => $dofollow_links, |
| 1090 |
], |
| 1091 |
]; |
| 1092 |
} |
| 1093 |
private function score_trustworthiness(array $content_data, array $metadata): array { |
| 1094 |
// E-E-A-T is an off-page/site-wide signal we cannot reliably measure |
| 1095 |
// from a single post. Award full credit (benefit of the doubt) rather |
| 1096 |
// than a fixed partial that silently caps every post's ceiling. |
| 1097 |
return [ |
| 1098 |
'score' => $this->scoring_factors['trustworthiness'], |
| 1099 |
'max_score' => $this->scoring_factors['trustworthiness'], |
| 1100 |
'suggestions' => ['Add author credentials, citations, and contact information to reinforce trustworthiness'], |
| 1101 |
'details' => ['trust_level' => 'Assumed adequate'], |
| 1102 |
]; |
| 1103 |
} |
| 1104 |
|
| 1105 |
private function score_link_diversity(array $content_data): array { |
| 1106 |
// Off-page link distribution; not measurable per post. Full credit. |
| 1107 |
return [ |
| 1108 |
'score' => $this->scoring_factors['link_diversity'], |
| 1109 |
'max_score' => $this->scoring_factors['link_diversity'], |
| 1110 |
'suggestions' => [], |
| 1111 |
'details' => ['diversity_level' => 'Assumed adequate'], |
| 1112 |
]; |
| 1113 |
} |
| 1114 |
private function score_site_security(array $content_data): array { |
| 1115 |
$max_score = $this->scoring_factors['site_security']; |
| 1116 |
$ssl = function_exists('is_ssl') ? is_ssl() : true; |
| 1117 |
|
| 1118 |
return [ |
| 1119 |
'score' => $ssl ? $max_score : 0, |
| 1120 |
'max_score' => $max_score, |
| 1121 |
'suggestions' => $ssl ? [] : ['Serve the site over HTTPS (install an SSL certificate)'], |
| 1122 |
'details' => ['ssl_enabled' => $ssl, 'security_level' => $ssl ? 'Good' : 'Insecure'], |
| 1123 |
]; |
| 1124 |
} |
| 1125 |
private function check_semantic_keyword_match(string $text, string $keyword): bool { |
| 1126 |
// Simple semantic matching - can be enhanced with AI/NLP |
| 1127 |
$keyword_parts = explode(' ', strtolower($keyword)); |
| 1128 |
$text_lower = strtolower($text); |
| 1129 |
|
| 1130 |
$matches = 0; |
| 1131 |
foreach ($keyword_parts as $part) { |
| 1132 |
if (strpos($text_lower, $part) !== false) { |
| 1133 |
$matches++; |
| 1134 |
} |
| 1135 |
} |
| 1136 |
|
| 1137 |
// Consider it a semantic match if 70% of keyword parts are present |
| 1138 |
return ($matches / count($keyword_parts)) >= 0.7; |
| 1139 |
} |
| 1140 |
private function prioritize_suggestions(array $suggestions, array $scores = []): array { |
| 1141 |
// Map each suggestion back to the factor that emitted it, so priority |
| 1142 |
// can rank by the points the factor actually lost instead of keyword- |
| 1143 |
// matching the advice text — which sorted a 2-point title tweak above |
| 1144 |
// a 6-point thin-content loss and contradicted the row's own impact |
| 1145 |
// tag (#408). |
| 1146 |
$by_text = []; |
| 1147 |
foreach ($scores as $factor => $result) { |
| 1148 |
if (!is_array($result) || empty($result['suggestions']) || !is_array($result['suggestions'])) { |
| 1149 |
continue; |
| 1150 |
} |
| 1151 |
$lost = max(0, (float) ($result['max_score'] ?? 0) - (float) ($result['score'] ?? 0)); |
| 1152 |
foreach ($result['suggestions'] as $text) { |
| 1153 |
if (is_string($text) && !isset($by_text[$text])) { |
| 1154 |
$by_text[$text] = ['factor' => (string) $factor, 'lost' => $lost]; |
| 1155 |
} |
| 1156 |
} |
| 1157 |
} |
| 1158 |
|
| 1159 |
$prioritized = []; |
| 1160 |
|
| 1161 |
foreach ($suggestions as $suggestion) { |
| 1162 |
$origin = $by_text[$suggestion] ?? null; |
| 1163 |
|
| 1164 |
// A factor already at full marks loses nothing to this advice — |
| 1165 |
// it was occupying list positions (sometimes at "High") while |
| 1166 |
// recovering zero points. Dropped rather than sorted last. |
| 1167 |
if (null !== $origin && $origin['lost'] <= 0) { |
| 1168 |
continue; |
| 1169 |
} |
| 1170 |
|
| 1171 |
if (null !== $origin) { |
| 1172 |
$priority = $origin['lost'] >= 4 ? 'High' : ($origin['lost'] >= 2 ? 'Medium' : 'Low'); |
| 1173 |
} else { |
| 1174 |
// No factor attached (defensive: a filter-added or legacy |
| 1175 |
// suggestion) — the old keyword map is the fallback. |
| 1176 |
$priority = $this->determine_suggestion_priority($suggestion); |
| 1177 |
} |
| 1178 |
|
| 1179 |
$prioritized[] = [ |
| 1180 |
'text' => $suggestion, |
| 1181 |
'priority' => $priority, |
| 1182 |
'impact' => $this->estimate_impact($suggestion), |
| 1183 |
'effort' => $this->estimate_effort($suggestion), |
| 1184 |
'factor' => $origin['factor'] ?? null, |
| 1185 |
'points_recoverable' => $origin['lost'] ?? null, |
| 1186 |
]; |
| 1187 |
} |
| 1188 |
|
| 1189 |
// Biggest recoverable loss first; keyword-mapped stragglers (no |
| 1190 |
// factor) sort within their priority band after the measured rows. |
| 1191 |
usort($prioritized, function($a, $b) { |
| 1192 |
$al = $a['points_recoverable'] ?? -1; |
| 1193 |
$bl = $b['points_recoverable'] ?? -1; |
| 1194 |
if ($al !== $bl) { |
| 1195 |
return $bl <=> $al; |
| 1196 |
} |
| 1197 |
$priority_order = ['High' => 3, 'Medium' => 2, 'Low' => 1]; |
| 1198 |
return $priority_order[$b['priority']] - $priority_order[$a['priority']]; |
| 1199 |
}); |
| 1200 |
|
| 1201 |
return $prioritized; |
| 1202 |
} |
| 1203 |
|
| 1204 |
/** |
| 1205 |
* Determine suggestion priority based on content |
| 1206 |
* |
| 1207 |
* @param string $suggestion Suggestion text |
| 1208 |
* @return string Priority level |
| 1209 |
*/ |
| 1210 |
private function determine_suggestion_priority(string $suggestion): string { |
| 1211 |
$high_priority_keywords = ['title', 'keyword', 'content quality', 'heading']; |
| 1212 |
$medium_priority_keywords = ['meta description', 'internal link', 'readability']; |
| 1213 |
|
| 1214 |
$suggestion_lower = strtolower($suggestion); |
| 1215 |
|
| 1216 |
foreach ($high_priority_keywords as $keyword) { |
| 1217 |
if (strpos($suggestion_lower, $keyword) !== false) { |
| 1218 |
return 'High'; |
| 1219 |
} |
| 1220 |
} |
| 1221 |
|
| 1222 |
foreach ($medium_priority_keywords as $keyword) { |
| 1223 |
if (strpos($suggestion_lower, $keyword) !== false) { |
| 1224 |
return 'Medium'; |
| 1225 |
} |
| 1226 |
} |
| 1227 |
|
| 1228 |
return 'Low'; |
| 1229 |
} |
| 1230 |
|
| 1231 |
/** |
| 1232 |
* Estimate impact of implementing suggestion |
| 1233 |
* |
| 1234 |
* @param string $suggestion Suggestion text |
| 1235 |
* @return string Impact level |
| 1236 |
*/ |
| 1237 |
private function estimate_impact(string $suggestion): string { |
| 1238 |
// Simple heuristic - can be enhanced with ML |
| 1239 |
if (strpos(strtolower($suggestion), 'title') !== false) { return 'High'; |
| 1240 |
} |
| 1241 |
if (strpos(strtolower($suggestion), 'content') !== false) { return 'High'; |
| 1242 |
} |
| 1243 |
if (strpos(strtolower($suggestion), 'keyword') !== false) { return 'Medium'; |
| 1244 |
} |
| 1245 |
return 'Low'; |
| 1246 |
} |
| 1247 |
|
| 1248 |
/** |
| 1249 |
* Estimate effort required to implement suggestion |
| 1250 |
* |
| 1251 |
* @param string $suggestion Suggestion text |
| 1252 |
* @return string Effort level |
| 1253 |
*/ |
| 1254 |
private function estimate_effort(string $suggestion): string { |
| 1255 |
// Simple heuristic - can be enhanced with ML |
| 1256 |
if (strpos(strtolower($suggestion), 'rewrite') !== false) { return 'High'; |
| 1257 |
} |
| 1258 |
if (strpos(strtolower($suggestion), 'add') !== false) { return 'Medium'; |
| 1259 |
} |
| 1260 |
if (strpos(strtolower($suggestion), 'optimize') !== false) { return 'Medium'; |
| 1261 |
} |
| 1262 |
return 'Low'; |
| 1263 |
} |
| 1264 |
private function calculate_topic_relevance(string $content, string $target_keyword): float { |
| 1265 |
if (empty($content) || empty($target_keyword)) { |
| 1266 |
return 0.0; |
| 1267 |
} |
| 1268 |
|
| 1269 |
// Occurrences and the word count both come from the reading text, so |
| 1270 |
// shortcode syntax is in neither. |
| 1271 |
$content_lower = strtolower(self::reading_text_of($content)); |
| 1272 |
$keyword_lower = strtolower($target_keyword); |
| 1273 |
|
| 1274 |
// Calculate keyword and semantic term frequency |
| 1275 |
$keyword_count = substr_count($content_lower, $keyword_lower); |
| 1276 |
$word_count = \ThinkRank\SEO\Word_Count_Index::count_words($content_lower); |
| 1277 |
|
| 1278 |
if ($word_count === 0) { |
| 1279 |
return 0.0; |
| 1280 |
} |
| 1281 |
|
| 1282 |
// Base relevance from keyword presence |
| 1283 |
$keyword_density = ($keyword_count / $word_count) * 100; |
| 1284 |
$base_relevance = min(1.0, $keyword_density / 2.0); // Optimal around 1-2% |
| 1285 |
|
| 1286 |
// Boost for semantic variations |
| 1287 |
$semantic_boost = $this->calculate_semantic_boost($content_lower, $keyword_lower); |
| 1288 |
|
| 1289 |
return min(1.0, $base_relevance + $semantic_boost); |
| 1290 |
} |
| 1291 |
|
| 1292 |
/** |
| 1293 |
* Calculate semantic boost for related terms |
| 1294 |
* |
| 1295 |
* @param string $content Content text (lowercase) |
| 1296 |
* @param string $keyword Target keyword (lowercase) |
| 1297 |
* @return float Semantic boost (0-0.3) |
| 1298 |
*/ |
| 1299 |
private function calculate_semantic_boost(string $content, string $keyword): float { |
| 1300 |
// Simple semantic term detection - can be enhanced with NLP |
| 1301 |
$semantic_terms = $this->get_semantic_terms($keyword); |
| 1302 |
$boost = 0.0; |
| 1303 |
|
| 1304 |
foreach ($semantic_terms as $term) { |
| 1305 |
if (strpos($content, $term) !== false) { |
| 1306 |
$boost += 0.05; // Small boost per semantic term |
| 1307 |
} |
| 1308 |
} |
| 1309 |
|
| 1310 |
return min(0.3, $boost); // Cap at 30% boost |
| 1311 |
} |
| 1312 |
|
| 1313 |
/** |
| 1314 |
* Get semantic terms for a keyword |
| 1315 |
* |
| 1316 |
* @param string $keyword Target keyword |
| 1317 |
* @return array Semantic terms |
| 1318 |
*/ |
| 1319 |
private function get_semantic_terms(string $keyword): array { |
| 1320 |
// Simple semantic term generation - can be enhanced with AI/NLP |
| 1321 |
$terms = []; |
| 1322 |
|
| 1323 |
// Add plural/singular variations |
| 1324 |
if (substr($keyword, -1) === 's') { |
| 1325 |
$terms[] = rtrim($keyword, 's'); |
| 1326 |
} else { |
| 1327 |
$terms[] = $keyword . 's'; |
| 1328 |
} |
| 1329 |
|
| 1330 |
// Add common related terms based on keyword |
| 1331 |
$keyword_lower = strtolower($keyword); |
| 1332 |
|
| 1333 |
// SEO-related terms |
| 1334 |
if (strpos($keyword_lower, 'seo') !== false) { |
| 1335 |
$terms = array_merge($terms, ['optimization', 'search engine', 'ranking', 'visibility']); |
| 1336 |
} |
| 1337 |
|
| 1338 |
// WordPress-related terms |
| 1339 |
if (strpos($keyword_lower, 'wordpress') !== false) { // phpcs:ignore WordPress.WP.CapitalPDangit.MisspelledInText -- lowercase on purpose: the haystack is strtolower()ed. |
| 1340 |
$terms = array_merge($terms, ['wp', 'plugin', 'theme', 'cms']); |
| 1341 |
} |
| 1342 |
|
| 1343 |
return $terms; |
| 1344 |
} |
| 1345 |
|
| 1346 |
/** |
| 1347 |
* Keyword density (%) of the target keyword across the plain-text body. |
| 1348 |
* |
| 1349 |
* @param string $content Raw/HTML content. |
| 1350 |
* @param string $target_keyword Target keyword. |
| 1351 |
* @return float Density percentage (0 when no keyword/content). |
| 1352 |
*/ |
| 1353 |
private function keyword_density(string $content, string $target_keyword): float { |
| 1354 |
if (empty($content) || empty($target_keyword)) { |
| 1355 |
return 0.0; |
| 1356 |
} |
| 1357 |
// Numerator and denominator from the same reading text. Counting the |
| 1358 |
// keyword in wp_strip_all_tags() output found it inside shortcode |
| 1359 |
// attributes the word count no longer includes. |
| 1360 |
$plain = strtolower(self::reading_text_of($content)); |
| 1361 |
$word_count = \ThinkRank\SEO\Word_Count_Index::count_words($plain); |
| 1362 |
if ($word_count === 0) { |
| 1363 |
return 0.0; |
| 1364 |
} |
| 1365 |
$occurrences = substr_count($plain, strtolower($target_keyword)); |
| 1366 |
return ($occurrences / $word_count) * 100; |
| 1367 |
} |
| 1368 |
|
| 1369 |
/** |
| 1370 |
* Whether the target keyword (or a semantic variation) appears in the body. |
| 1371 |
* |
| 1372 |
* @param string $content Raw/HTML content. |
| 1373 |
* @param string $target_keyword Target keyword. |
| 1374 |
* @return bool |
| 1375 |
*/ |
| 1376 |
private function keyword_in_content(string $content, string $target_keyword): bool { |
| 1377 |
if (empty($content) || empty($target_keyword)) { |
| 1378 |
return false; |
| 1379 |
} |
| 1380 |
$plain = strtolower(wp_strip_all_tags($content)); |
| 1381 |
if (strpos($plain, strtolower($target_keyword)) !== false) { |
| 1382 |
return true; |
| 1383 |
} |
| 1384 |
return $this->check_semantic_keyword_match($plain, $target_keyword); |
| 1385 |
} |
| 1386 |
|
| 1387 |
/** |
| 1388 |
* Whether the keyword appears early (first paragraph / first ~10% of words). |
| 1389 |
* |
| 1390 |
* @param string $content Raw/HTML content. |
| 1391 |
* @param string $target_keyword Target keyword. |
| 1392 |
* @return bool |
| 1393 |
*/ |
| 1394 |
private function keyword_in_first_paragraph(string $content, string $target_keyword): bool { |
| 1395 |
if (empty($content) || empty($target_keyword)) { |
| 1396 |
return false; |
| 1397 |
} |
| 1398 |
$plain = strtolower(self::reading_text_of($content)); |
| 1399 |
$words = preg_split('/\s+/', $plain, -1, PREG_SPLIT_NO_EMPTY) ?: []; |
| 1400 |
$window = array_slice($words, 0, max(50, (int) ceil(count($words) * 0.1))); |
| 1401 |
return strpos(implode(' ', $window), strtolower($target_keyword)) !== false; |
| 1402 |
} |
| 1403 |
|
| 1404 |
/** |
| 1405 |
* Whether the keyword appears in any subheading (H2–H6). |
| 1406 |
* |
| 1407 |
* @param array $headings Extracted headings (each with 'level' + 'text'). |
| 1408 |
* @param string $target_keyword Target keyword. |
| 1409 |
* @return bool |
| 1410 |
*/ |
| 1411 |
private function keyword_in_subheadings(array $headings, string $target_keyword): bool { |
| 1412 |
if (empty($target_keyword)) { |
| 1413 |
return false; |
| 1414 |
} |
| 1415 |
$keyword_lower = strtolower($target_keyword); |
| 1416 |
foreach ($headings as $heading) { |
| 1417 |
if ((int) ($heading['level'] ?? 0) < 2) { |
| 1418 |
continue; |
| 1419 |
} |
| 1420 |
$text = strtolower((string) ($heading['text'] ?? '')); |
| 1421 |
if ($text !== '' && strpos($text, $keyword_lower) !== false) { |
| 1422 |
return true; |
| 1423 |
} |
| 1424 |
} |
| 1425 |
return false; |
| 1426 |
} |
| 1427 |
|
| 1428 |
/** |
| 1429 |
* Days elapsed since a MySQL datetime string, or null when unparseable. |
| 1430 |
* |
| 1431 |
* @param string $datetime MySQL datetime (e.g. post_modified). |
| 1432 |
* @return int|null |
| 1433 |
*/ |
| 1434 |
private function days_since(string $datetime): ?int { |
| 1435 |
$datetime = trim($datetime); |
| 1436 |
if ($datetime === '' || strpos($datetime, '0000-00-00') === 0) { |
| 1437 |
return null; |
| 1438 |
} |
| 1439 |
$ts = strtotime($datetime); |
| 1440 |
if ($ts === false) { |
| 1441 |
return null; |
| 1442 |
} |
| 1443 |
// strtotime() returns a real unix timestamp, so this must compare against |
| 1444 |
// one: current_time('timestamp') is offset by the site timezone and made |
| 1445 |
// every "days ago" figure wrong by that offset. |
| 1446 |
$now = time(); |
| 1447 |
return (int) floor(($now - $ts) / 86400); |
| 1448 |
} |
| 1449 |
|
| 1450 |
/** |
| 1451 |
* Whether the keyword appears in the meta description. |
| 1452 |
* |
| 1453 |
* @param string $meta_description Meta description text. |
| 1454 |
* @param string $target_keyword Target keyword. |
| 1455 |
* @return bool |
| 1456 |
*/ |
| 1457 |
private function keyword_in_meta(string $meta_description, string $target_keyword): bool { |
| 1458 |
if ($meta_description === '' || $target_keyword === '') { |
| 1459 |
return false; |
| 1460 |
} |
| 1461 |
return strpos(strtolower($meta_description), strtolower($target_keyword)) !== false; |
| 1462 |
} |
| 1463 |
|
| 1464 |
/** |
| 1465 |
* Whether the keyword appears in any image alt text. |
| 1466 |
* |
| 1467 |
* @param array $images Images (each with an 'alt' key). |
| 1468 |
* @param string $target_keyword Target keyword. |
| 1469 |
* @return bool |
| 1470 |
*/ |
| 1471 |
private function keyword_in_alt(array $images, string $target_keyword): bool { |
| 1472 |
if ($target_keyword === '') { |
| 1473 |
return false; |
| 1474 |
} |
| 1475 |
$keyword_lower = strtolower($target_keyword); |
| 1476 |
foreach ($images as $image) { |
| 1477 |
$alt = strtolower((string) ($image['alt'] ?? '')); |
| 1478 |
if ($alt !== '' && strpos($alt, $keyword_lower) !== false) { |
| 1479 |
return true; |
| 1480 |
} |
| 1481 |
} |
| 1482 |
return false; |
| 1483 |
} |
| 1484 |
|
| 1485 |
/** |
| 1486 |
* Whether the title contains a common power word (CTR booster). |
| 1487 |
* |
| 1488 |
* @param string $title Post title. |
| 1489 |
* @return bool |
| 1490 |
*/ |
| 1491 |
private function title_has_power_word(string $title): bool { |
| 1492 |
$title_lower = strtolower($title); |
| 1493 |
foreach (self::get_title_power_words() as $word) { |
| 1494 |
if (strpos($title_lower, $word) !== false) { |
| 1495 |
return true; |
| 1496 |
} |
| 1497 |
} |
| 1498 |
return false; |
| 1499 |
} |
| 1500 |
|
| 1501 |
/** |
| 1502 |
* The power words rewarded by the title check. Single source of truth so the |
| 1503 |
* AI title improver can require the generated title to actually contain one. |
| 1504 |
* |
| 1505 |
* @return string[] |
| 1506 |
*/ |
| 1507 |
public static function get_title_power_words(): array { |
| 1508 |
return [ |
| 1509 |
'ultimate', 'essential', 'complete', 'proven', 'guide', 'best', 'top', |
| 1510 |
'free', 'easy', 'simple', 'quick', 'fast', 'powerful', 'secret', 'expert', |
| 1511 |
'effective', 'amazing', 'incredible', 'exclusive', 'definitive', 'step-by-step', |
| 1512 |
]; |
| 1513 |
} |
| 1514 |
|
| 1515 |
/** |
| 1516 |
* The emotion/sentiment words rewarded by the title check. Single source of |
| 1517 |
* truth shared with the AI title improver so a generated title satisfies the |
| 1518 |
* same validation. |
| 1519 |
* |
| 1520 |
* @return string[] |
| 1521 |
*/ |
| 1522 |
public static function get_title_sentiment_words(): array { |
| 1523 |
return [ |
| 1524 |
// Positive |
| 1525 |
'great', 'good', 'better', 'awesome', 'love', 'win', 'boost', 'improve', |
| 1526 |
'success', 'smart', 'brilliant', 'perfect', 'happy', 'beautiful', |
| 1527 |
// Negative (drives clicks too) |
| 1528 |
'avoid', 'mistake', 'worst', 'stop', 'never', 'bad', 'wrong', 'fail', |
| 1529 |
'danger', 'warning', 'painful', 'ugly', |
| 1530 |
]; |
| 1531 |
} |
| 1532 |
|
| 1533 |
/** |
| 1534 |
* Whether the title carries an emotional/sentiment word (positive or |
| 1535 |
* negative), which Rank Math rewards for higher engagement. |
| 1536 |
* |
| 1537 |
* @param string $title Post title. |
| 1538 |
* @return bool |
| 1539 |
*/ |
| 1540 |
private function title_has_sentiment_word(string $title): bool { |
| 1541 |
$title_lower = strtolower($title); |
| 1542 |
foreach (self::get_title_sentiment_words() as $word) { |
| 1543 |
if (strpos($title_lower, $word) !== false) { |
| 1544 |
return true; |
| 1545 |
} |
| 1546 |
} |
| 1547 |
return false; |
| 1548 |
} |
| 1549 |
|
| 1550 |
/** |
| 1551 |
* Assess content depth based on word count |
| 1552 |
* |
| 1553 |
* @param int $word_count Word count |
| 1554 |
* @return string Depth assessment |
| 1555 |
*/ |
| 1556 |
private function assess_content_depth(int $word_count): string { |
| 1557 |
if ($word_count >= 2000) { return 'Comprehensive'; |
| 1558 |
} |
| 1559 |
if ($word_count >= 1000) { return 'Detailed'; |
| 1560 |
} |
| 1561 |
if ($word_count >= 500) { return 'Moderate'; |
| 1562 |
} |
| 1563 |
if ($word_count >= 300) { return 'Basic'; |
| 1564 |
} |
| 1565 |
return 'Insufficient'; |
| 1566 |
} |
| 1567 |
private function score_content_freshness(array $content_data): array { |
| 1568 |
$max_score = $this->scoring_factors['content_freshness']; |
| 1569 |
$suggestions = []; |
| 1570 |
|
| 1571 |
// Derive freshness from the real last-modified date when available |
| 1572 |
// (live editing has no stored date yet -> treat as fresh). |
| 1573 |
$days = $this->days_since((string) ($content_data['post_modified'] ?? '')); |
| 1574 |
|
| 1575 |
if ($days === null || $days <= 180) { |
| 1576 |
$score = $max_score; |
| 1577 |
$status = 'Current'; |
| 1578 |
} elseif ($days <= 365) { |
| 1579 |
$score = (int) round($max_score * 0.66); |
| 1580 |
$status = 'Aging'; |
| 1581 |
$suggestions[] = 'Content is 6-12 months old - review and refresh it for better freshness signals'; |
| 1582 |
} else { |
| 1583 |
$score = (int) round($max_score * 0.33); |
| 1584 |
$status = 'Stale'; |
| 1585 |
$suggestions[] = 'Content is over a year old - update it to maintain freshness signals'; |
| 1586 |
} |
| 1587 |
|
| 1588 |
return [ |
| 1589 |
'score' => $score, |
| 1590 |
'max_score' => $max_score, |
| 1591 |
'suggestions' => $suggestions, |
| 1592 |
'details' => [ |
| 1593 |
'freshness_status' => $status, |
| 1594 |
'days_since_modified' => $days, |
| 1595 |
], |
| 1596 |
]; |
| 1597 |
} |
| 1598 |
/** |
| 1599 |
* Resolve a post's content into something worth analyzing. |
| 1600 |
* |
| 1601 |
* Delegates to Builder_Content, which knows where each page builder keeps |
| 1602 |
* its text. Kept as the historical entry point for existing callers. |
| 1603 |
* |
| 1604 |
* @since 1.23.0 |
| 1605 |
* |
| 1606 |
* @param \WP_Post $post Post being analyzed. |
| 1607 |
* @return string Content to analyze. |
| 1608 |
*/ |
| 1609 |
public static function resolve_analyzable_content(\WP_Post $post): string { |
| 1610 |
if (!class_exists('\ThinkRank\SEO\Builder_Content')) { |
| 1611 |
require_once THINKRANK_PLUGIN_DIR . 'includes/seo/class-builder-content.php'; |
| 1612 |
} |
| 1613 |
|
| 1614 |
return \ThinkRank\SEO\Builder_Content::resolve($post); |
| 1615 |
} |
| 1616 |
|
| 1617 |
/** |
| 1618 |
* Resolve editor-supplied live content into something worth analyzing. |
| 1619 |
* |
| 1620 |
* @since 1.23.0 |
| 1621 |
* |
| 1622 |
* @param string $live_content Markup supplied by the editor. |
| 1623 |
* @param \WP_Post $post Post the markup belongs to. |
| 1624 |
* @return string Content to analyze. |
| 1625 |
*/ |
| 1626 |
public static function resolve_live_content(string $live_content, \WP_Post $post): string { |
| 1627 |
if (!class_exists('\ThinkRank\SEO\Builder_Content')) { |
| 1628 |
require_once THINKRANK_PLUGIN_DIR . 'includes/seo/class-builder-content.php'; |
| 1629 |
} |
| 1630 |
|
| 1631 |
// Bind the live markup to the post the render makes current. Since #862 |
| 1632 |
// the resolver runs setup_postdata() on this post, so a shortcode that |
| 1633 |
// builds its output from the current post's content — get_the_content(), |
| 1634 |
// get_post()->post_content, as a table of contents or a reading-time |
| 1635 |
// shortcode does — otherwise read the last saved body while the unsaved |
| 1636 |
// markup rendered around it, and lagged a save behind (#864). |
| 1637 |
// |
| 1638 |
// A clone, not the caller's object: the post is current only for the |
| 1639 |
// duration of the render and the caller's $post must come back |
| 1640 |
// unchanged. `thinkrank_analyzable_content` receives the clone too, |
| 1641 |
// which is what makes $post->post_content there agree with the markup |
| 1642 |
// being analyzed on the live path. |
| 1643 |
$bound = clone $post; |
| 1644 |
$bound->post_content = $live_content; |
| 1645 |
|
| 1646 |
return \ThinkRank\SEO\Builder_Content::resolve_markup($live_content, $bound); |
| 1647 |
} |
| 1648 |
|
| 1649 |
public function analyze_post_content(int $post_id): array { |
| 1650 |
$post = get_post($post_id); |
| 1651 |
if (!$post) { |
| 1652 |
return []; |
| 1653 |
} |
| 1654 |
|
| 1655 |
$content = self::resolve_analyzable_content($post); |
| 1656 |
$title = $post->post_title; |
| 1657 |
|
| 1658 |
// Extract headings from content |
| 1659 |
$headings = $this->extract_headings($content); |
| 1660 |
|
| 1661 |
// Count the markup, not wp_strip_all_tags() output: stripping deletes |
| 1662 |
// tags without a space, so "five</p><p>six" would already be one word |
| 1663 |
// before the counter saw it. |
| 1664 |
$word_count = $this->calculate_word_count_js_style($content); |
| 1665 |
|
| 1666 |
// Calculate readability |
| 1667 |
$readability_score = $this->calculate_readability_score($content); |
| 1668 |
|
| 1669 |
// Count links |
| 1670 |
$internal_links = $this->count_internal_links($content); |
| 1671 |
$external_links = $this->count_external_links($content); |
| 1672 |
$external_dofollow_links = $this->count_external_dofollow_links($content); |
| 1673 |
|
| 1674 |
// Analyze images |
| 1675 |
$images = $this->analyze_images($content); |
| 1676 |
|
| 1677 |
// Get URL |
| 1678 |
$url = get_permalink($post_id); |
| 1679 |
|
| 1680 |
return [ |
| 1681 |
'content' => $content, |
| 1682 |
'title' => $title, |
| 1683 |
'headings' => $headings, |
| 1684 |
'word_count' => $word_count, |
| 1685 |
'readability_score' => $readability_score, |
| 1686 |
'internal_links' => $internal_links, |
| 1687 |
'external_links' => $external_links, |
| 1688 |
'external_dofollow_links' => $external_dofollow_links, |
| 1689 |
'images' => $images, |
| 1690 |
'url' => $url, |
| 1691 |
'slug' => $post->post_name, |
| 1692 |
'post_modified' => $post->post_modified, |
| 1693 |
'schema_present' => $this->detect_schema_present($content) |
| 1694 |
|| $this->thinkrank_global_schema_active($post->post_type) |
| 1695 |
|| $this->thinkrank_deployed_schema_active($post), |
| 1696 |
]; |
| 1697 |
} |
| 1698 |
|
| 1699 |
/** |
| 1700 |
* Build the human-readable readability label (mirrors the editor's |
| 1701 |
* calculateReadabilityScore: "<level> (<flesch>)") from analyzed content. |
| 1702 |
* |
| 1703 |
* @param array $content_data Output of analyze_post_content()/analyze_live_content() |
| 1704 |
* @return string Readability label, e.g. "Standard (62)" |
| 1705 |
*/ |
| 1706 |
private function format_readability_label(array $content_data): string { |
| 1707 |
if ((int) ($content_data['word_count'] ?? 0) === 0) { |
| 1708 |
return 'No content'; |
| 1709 |
} |
| 1710 |
|
| 1711 |
$rounded = (int) round((float) ($content_data['readability_score'] ?? 0)); |
| 1712 |
|
| 1713 |
if ($rounded >= 90) { |
| 1714 |
$level = 'Very Easy'; |
| 1715 |
} elseif ($rounded >= 80) { |
| 1716 |
$level = 'Easy'; |
| 1717 |
} elseif ($rounded >= 70) { |
| 1718 |
$level = 'Fairly Easy'; |
| 1719 |
} elseif ($rounded >= 60) { |
| 1720 |
$level = 'Standard'; |
| 1721 |
} elseif ($rounded >= 50) { |
| 1722 |
$level = 'Fairly Difficult'; |
| 1723 |
} elseif ($rounded >= 30) { |
| 1724 |
$level = 'Difficult'; |
| 1725 |
} else { |
| 1726 |
$level = 'Very Difficult'; |
| 1727 |
} |
| 1728 |
|
| 1729 |
return "{$level} ({$rounded})"; |
| 1730 |
} |
| 1731 |
|
| 1732 |
/** |
| 1733 |
* Derive the content-quality label (mirrors the editor's |
| 1734 |
* calculateContentQuality: word count + long-paragraph + subheading scoring) |
| 1735 |
* so persisted scores carry a non-null quality value. |
| 1736 |
* |
| 1737 |
* @param array $content_data Output of analyze_post_content()/analyze_live_content() |
| 1738 |
* @return string One of: No content, Good, OK, Needs improvement |
| 1739 |
*/ |
| 1740 |
private function derive_content_quality(array $content_data): string { |
| 1741 |
$word_count = (int) ($content_data['word_count'] ?? 0); |
| 1742 |
if ($word_count === 0) { |
| 1743 |
return 'No content'; |
| 1744 |
} |
| 1745 |
|
| 1746 |
$final = $this->derive_content_quality_score($content_data); |
| 1747 |
|
| 1748 |
if ($final >= 80) { |
| 1749 |
return 'Good'; |
| 1750 |
} |
| 1751 |
if ($final >= 50) { |
| 1752 |
return 'OK'; |
| 1753 |
} |
| 1754 |
|
| 1755 |
return 'Needs improvement'; |
| 1756 |
} |
| 1757 |
|
| 1758 |
/** |
| 1759 |
* Numeric content-quality score (0-100): word count + long-paragraph + |
| 1760 |
* subheading distribution. Shared by the quality label and the |
| 1761 |
* satisfying-content factor so both stay in sync. |
| 1762 |
* |
| 1763 |
* @param array $content_data Output of analyze_post_content()/analyze_live_content() |
| 1764 |
* @return int Quality score 0-100. |
| 1765 |
*/ |
| 1766 |
private function derive_content_quality_score(array $content_data): int { |
| 1767 |
$word_count = (int) ($content_data['word_count'] ?? 0); |
| 1768 |
if ($word_count === 0) { |
| 1769 |
return 0; |
| 1770 |
} |
| 1771 |
|
| 1772 |
$content = (string) ($content_data['content'] ?? ''); |
| 1773 |
$score = 0; |
| 1774 |
|
| 1775 |
// 1. Word count (industry standard: 300+ words). |
| 1776 |
if ($word_count >= 600) { |
| 1777 |
$score += 100; |
| 1778 |
} elseif ($word_count >= 300) { |
| 1779 |
$score += 50; |
| 1780 |
} |
| 1781 |
|
| 1782 |
// 2. Long paragraphs (flag paragraphs over 150 words). |
| 1783 |
$long_paragraphs = 0; |
| 1784 |
if (preg_match_all('/<p[^>]*>(.*?)<\/p>/is', $content, $matches)) { |
| 1785 |
foreach ($matches[1] as $paragraph) { |
| 1786 |
if ($this->calculate_word_count_js_style($paragraph) > 150) { |
| 1787 |
$long_paragraphs++; |
| 1788 |
} |
| 1789 |
} |
| 1790 |
} |
| 1791 |
if ($long_paragraphs === 0) { |
| 1792 |
$score += 100; |
| 1793 |
} elseif ($long_paragraphs <= 2) { |
| 1794 |
$score += 50; |
| 1795 |
} |
| 1796 |
|
| 1797 |
// 3. Subheading distribution (H2–H6, ~one per 300 words). |
| 1798 |
$subheadings = 0; |
| 1799 |
foreach ((array) ($content_data['headings'] ?? []) as $heading) { |
| 1800 |
if ((int) ($heading['level'] ?? 0) >= 2) { |
| 1801 |
$subheadings++; |
| 1802 |
} |
| 1803 |
} |
| 1804 |
$expected = (int) floor($word_count / 300); |
| 1805 |
if ($subheadings > 0 && $subheadings >= $expected) { |
| 1806 |
$score += 100; |
| 1807 |
} elseif ($subheadings > 0) { |
| 1808 |
$score += 50; |
| 1809 |
} |
| 1810 |
|
| 1811 |
return (int) round($score / 3); |
| 1812 |
} |
| 1813 |
|
| 1814 |
/** |
| 1815 |
* Extract headings from content |
| 1816 |
* |
| 1817 |
* @param string $content Content HTML |
| 1818 |
* @return array Array of headings with levels |
| 1819 |
*/ |
| 1820 |
private function extract_headings(string $content): array { |
| 1821 |
$headings = []; |
| 1822 |
|
| 1823 |
// Match H1-H6 tags |
| 1824 |
if (preg_match_all('/<h([1-6])[^>]*>(.*?)<\/h[1-6]>/i', $content, $matches, PREG_SET_ORDER)) { |
| 1825 |
foreach ($matches as $match) { |
| 1826 |
$headings[] = [ |
| 1827 |
'level' => (int)$match[1], |
| 1828 |
'text' => wp_strip_all_tags($match[2]), |
| 1829 |
]; |
| 1830 |
} |
| 1831 |
} |
| 1832 |
|
| 1833 |
return $headings; |
| 1834 |
} |
| 1835 |
|
| 1836 |
/** |
| 1837 |
* Calculate readability score using Flesch Reading Ease |
| 1838 |
* |
| 1839 |
* @param string $content Content text |
| 1840 |
* @return float Readability score |
| 1841 |
*/ |
| 1842 |
private function calculate_readability_score(string $content): float { |
| 1843 |
// Sentences, words and syllables all come from one reading text. The |
| 1844 |
// word count drops shortcode syntax and punctuation-only tokens; when |
| 1845 |
// sentences and syllables were still read off wp_strip_all_tags() |
| 1846 |
// output, "[vc_column width="1/2"]" added syllables (and, with a "." in |
| 1847 |
// an attribute, sentences) to a word count that did not include it, |
| 1848 |
// and a WPBakery page's Flesch fell from 65 to 46. |
| 1849 |
$text = self::reading_text_of($content); |
| 1850 |
|
| 1851 |
if ('' === $text) { |
| 1852 |
return 0; |
| 1853 |
} |
| 1854 |
|
| 1855 |
$sentence_count = self::count_sentences($text); |
| 1856 |
$word_count = \ThinkRank\SEO\Word_Count_Index::count_words($text); |
| 1857 |
$syllable_count = $this->count_syllables($text); |
| 1858 |
|
| 1859 |
if ($sentence_count === 0 || $word_count === 0) { |
| 1860 |
return 0; |
| 1861 |
} |
| 1862 |
|
| 1863 |
// Flesch Reading Ease formula |
| 1864 |
$score = 206.835 - (1.015 * ($word_count / $sentence_count)) - (84.6 * ($syllable_count / $word_count)); |
| 1865 |
|
| 1866 |
return max(0, min(100, $score)); |
| 1867 |
} |
| 1868 |
|
| 1869 |
/** |
| 1870 |
* Count the sentences in reading text. |
| 1871 |
* |
| 1872 |
* A sentence is a run of text between terminal punctuation that holds at |
| 1873 |
* least one letter or digit. A fragment with no word in it (the space |
| 1874 |
* after the final full stop, a stray "!" between two "?") is not a |
| 1875 |
* sentence. contentAnalysis.js countSentences() applies the same rule. |
| 1876 |
* |
| 1877 |
* @since 2.14.2 |
| 1878 |
* |
| 1879 |
* @param string $text Text as returned by {@see self::reading_text_of()}. |
| 1880 |
* @return int |
| 1881 |
*/ |
| 1882 |
private static function count_sentences(string $text): int { |
| 1883 |
$fragments = preg_split('/[.!?]+/', $text, -1, PREG_SPLIT_NO_EMPTY) ?: []; |
| 1884 |
|
| 1885 |
return count(preg_grep('/[\p{L}\p{N}]/u', $fragments) ?: []); |
| 1886 |
} |
| 1887 |
|
| 1888 |
/** |
| 1889 |
* Count syllables in text (approximate) |
| 1890 |
* |
| 1891 |
* Expects reading text ({@see self::reading_text_of()}); tags are not |
| 1892 |
* stripped here, so a decoded "<" in prose is not read as a tag. |
| 1893 |
* |
| 1894 |
* @param string $text Text to analyze |
| 1895 |
* @return int Syllable count |
| 1896 |
*/ |
| 1897 |
private function count_syllables(string $text): int { |
| 1898 |
$words = preg_split('/\s+/', trim(strtolower($text)), -1, PREG_SPLIT_NO_EMPTY) ?: []; |
| 1899 |
$syllables = 0; |
| 1900 |
|
| 1901 |
foreach ($words as $word) { |
| 1902 |
$word = preg_replace('/[^a-z]/', '', $word); |
| 1903 |
if ($word === '') { |
| 1904 |
continue; |
| 1905 |
} |
| 1906 |
|
| 1907 |
$groups = preg_match_all('/[aeiouy]+/', $word); |
| 1908 |
|
| 1909 |
// Standard Flesch heuristic: a trailing silent e does not form a |
| 1910 |
// syllable ("make", "time", "these") — but only when a consonant |
| 1911 |
// precedes it (a vowel+e ending like "movie" already shares its |
| 1912 |
// group) and never for consonant-le ("table"), which does count. |
| 1913 |
// Without this the counter inflated syllables/word by ~0.2-0.3 on |
| 1914 |
// ordinary prose, driving raw Flesch negative and the UI to a |
| 1915 |
// clamped "Very Difficult (0)" (#407). |
| 1916 |
if ($groups > 1 && preg_match('/[^aeiouy]e$/', $word) && !str_ends_with($word, 'le')) { |
| 1917 |
$groups--; |
| 1918 |
} |
| 1919 |
|
| 1920 |
$syllables += max(1, $groups); |
| 1921 |
} |
| 1922 |
|
| 1923 |
return $syllables; |
| 1924 |
} |
| 1925 |
|
| 1926 |
/** |
| 1927 |
* Count internal links in content |
| 1928 |
* |
| 1929 |
* @param string $content Content HTML |
| 1930 |
* @return int Internal link count |
| 1931 |
*/ |
| 1932 |
private function count_internal_links(string $content): int { |
| 1933 |
$site_url = get_site_url(); |
| 1934 |
$count = 0; |
| 1935 |
|
| 1936 |
if (preg_match_all('/<a[^>]+href=["\']([^"\']+)["\'][^>]*>/i', $content, $matches)) { |
| 1937 |
foreach ($matches[1] as $url) { |
| 1938 |
if (strpos($url, $site_url) !== false || strpos($url, '/') === 0) { |
| 1939 |
$count++; |
| 1940 |
} |
| 1941 |
} |
| 1942 |
} |
| 1943 |
|
| 1944 |
return $count; |
| 1945 |
} |
| 1946 |
|
| 1947 |
/** |
| 1948 |
* Count external links in content |
| 1949 |
* |
| 1950 |
* @param string $content Content HTML |
| 1951 |
* @return int External link count |
| 1952 |
*/ |
| 1953 |
private function count_external_links(string $content): int { |
| 1954 |
$site_url = get_site_url(); |
| 1955 |
$count = 0; |
| 1956 |
|
| 1957 |
if (preg_match_all('/<a[^>]+href=["\']([^"\']+)["\'][^>]*>/i', $content, $matches)) { |
| 1958 |
foreach ($matches[1] as $url) { |
| 1959 |
if (strpos($url, 'http') === 0 && strpos($url, $site_url) === false) { |
| 1960 |
$count++; |
| 1961 |
} |
| 1962 |
} |
| 1963 |
} |
| 1964 |
|
| 1965 |
return $count; |
| 1966 |
} |
| 1967 |
|
| 1968 |
/** |
| 1969 |
* Count external links that pass link equity (not rel="nofollow"). |
| 1970 |
* Mirrors Rank Math's "external dofollow link" check. |
| 1971 |
* |
| 1972 |
* @param string $content Content HTML |
| 1973 |
* @return int External dofollow link count |
| 1974 |
*/ |
| 1975 |
private function count_external_dofollow_links(string $content): int { |
| 1976 |
$site_url = get_site_url(); |
| 1977 |
$count = 0; |
| 1978 |
|
| 1979 |
if (preg_match_all('/<a\b[^>]*>/i', $content, $matches)) { |
| 1980 |
foreach ($matches[0] as $tag) { |
| 1981 |
if (!preg_match('/href=["\']([^"\']+)["\']/i', $tag, $href)) { |
| 1982 |
continue; |
| 1983 |
} |
| 1984 |
$url = $href[1]; |
| 1985 |
$is_external = strpos($url, 'http') === 0 && strpos($url, $site_url) === false; |
| 1986 |
if (!$is_external) { |
| 1987 |
continue; |
| 1988 |
} |
| 1989 |
if (preg_match('/rel=["\'][^"\']*\bnofollow\b[^"\']*["\']/i', $tag)) { |
| 1990 |
continue; |
| 1991 |
} |
| 1992 |
$count++; |
| 1993 |
} |
| 1994 |
} |
| 1995 |
|
| 1996 |
return $count; |
| 1997 |
} |
| 1998 |
|
| 1999 |
/** |
| 2000 |
* Analyze images in content |
| 2001 |
* |
| 2002 |
* @param string $content Content HTML |
| 2003 |
* @return array Image analysis data |
| 2004 |
*/ |
| 2005 |
/** |
| 2006 |
* Detect structured data embedded directly in the content (JSON-LD script |
| 2007 |
* blocks or microdata attributes). Site-wide schema injected at render time |
| 2008 |
* is a separate feature and intentionally out of scope here. |
| 2009 |
* |
| 2010 |
* @param string $content Raw/HTML content. |
| 2011 |
* @return bool |
| 2012 |
*/ |
| 2013 |
private function detect_schema_present(string $content): bool { |
| 2014 |
if ($content === '') { |
| 2015 |
return false; |
| 2016 |
} |
| 2017 |
return stripos($content, 'application/ld+json') !== false |
| 2018 |
|| stripos($content, 'itemscope') !== false |
| 2019 |
|| stripos($content, 'itemtype') !== false; |
| 2020 |
} |
| 2021 |
|
| 2022 |
/** |
| 2023 |
* Whether ThinkRank's Global SEO schema output is active for a post type. |
| 2024 |
* |
| 2025 |
* ThinkRank injects JSON-LD at render time (wp_head) when a schema type is |
| 2026 |
* configured for the post type, so a post can have valid structured data |
| 2027 |
* even when none is embedded in the post body. The score credits this so the |
| 2028 |
* "add structured data" suggestion reflects ThinkRank's own schema engine. |
| 2029 |
* |
| 2030 |
* @param string $post_type Post type slug. |
| 2031 |
* @return bool |
| 2032 |
*/ |
| 2033 |
private function thinkrank_global_schema_active(string $post_type): bool { |
| 2034 |
if ($post_type === '') { |
| 2035 |
return false; |
| 2036 |
} |
| 2037 |
$all_settings = get_option('thinkrank_global_seo_settings', []); |
| 2038 |
return !empty($all_settings[$post_type]['schema_type']); |
| 2039 |
} |
| 2040 |
|
| 2041 |
/** |
| 2042 |
* Whether the Schema Manager has an active deployed schema for this post. |
| 2043 |
* |
| 2044 |
* Per-post schema deployed from the editor's Schema tab is stored in the |
| 2045 |
* Schema Manager's own table and emitted at wp_head by |
| 2046 |
* Frontend\SEO_Manager::output_site_schema_markup(). Neither |
| 2047 |
* detect_schema_present() (body scan) nor thinkrank_global_schema_active() |
| 2048 |
* (post-type option) sees it, so without this the score reported "no |
| 2049 |
* structured data" for posts that do emit it. |
| 2050 |
* |
| 2051 |
* Mirrors the context_type whitelist the emitter and the metabox both use, so |
| 2052 |
* the lookup targets the same row the front end reads. |
| 2053 |
* |
| 2054 |
* @param \WP_Post $post Post being scored. |
| 2055 |
* @return bool |
| 2056 |
*/ |
| 2057 |
private function thinkrank_deployed_schema_active(\WP_Post $post): bool { |
| 2058 |
if (!class_exists('ThinkRank\\SEO\\Schema_Management_System')) { |
| 2059 |
$manager_file = THINKRANK_PLUGIN_DIR . 'includes/seo/class-schema-management-system.php'; |
| 2060 |
if (!file_exists($manager_file)) { |
| 2061 |
return false; |
| 2062 |
} |
| 2063 |
require_once $manager_file; |
| 2064 |
} |
| 2065 |
|
| 2066 |
$context_type = in_array($post->post_type, ['site', 'post', 'page', 'product'], true) |
| 2067 |
? $post->post_type |
| 2068 |
: 'post'; |
| 2069 |
|
| 2070 |
try { |
| 2071 |
$manager = new \ThinkRank\SEO\Schema_Management_System(); |
| 2072 |
return !empty($manager->get_deployed_schemas($context_type, (int) $post->ID)); |
| 2073 |
} catch (\Throwable $e) { |
| 2074 |
return false; |
| 2075 |
} |
| 2076 |
} |
| 2077 |
|
| 2078 |
private function analyze_images(string $content): array { |
| 2079 |
$images = []; |
| 2080 |
|
| 2081 |
if (preg_match_all('/<img[^>]+>/i', $content, $matches)) { |
| 2082 |
foreach ($matches[0] as $img_tag) { |
| 2083 |
$alt = ''; |
| 2084 |
if (preg_match('/alt=["\']([^"\']*)["\']/', $img_tag, $alt_match)) { |
| 2085 |
$alt = $alt_match[1]; |
| 2086 |
} |
| 2087 |
|
| 2088 |
$src = ''; |
| 2089 |
if (preg_match('/src=["\']([^"\']*)["\']/', $img_tag, $src_match)) { |
| 2090 |
$src = $src_match[1]; |
| 2091 |
} |
| 2092 |
|
| 2093 |
$images[] = [ |
| 2094 |
'src' => $src, |
| 2095 |
'alt' => $alt, |
| 2096 |
]; |
| 2097 |
} |
| 2098 |
} |
| 2099 |
|
| 2100 |
return $images; |
| 2101 |
} |
| 2102 |
|
| 2103 |
/** |
| 2104 |
* Save SEO score to database |
| 2105 |
* |
| 2106 |
* @param int $post_id Post ID |
| 2107 |
* @param int $user_id User ID |
| 2108 |
* @param array $score_data Score data |
| 2109 |
* @return int|false Score ID or false on failure |
| 2110 |
*/ |
| 2111 |
public function save_score(int $post_id, int $user_id, array $score_data) { |
| 2112 |
global $wpdb; |
| 2113 |
|
| 2114 |
$table_name = $wpdb->prefix . 'thinkrank_seo_scores'; |
| 2115 |
|
| 2116 |
// Prepare data for insertion |
| 2117 |
$insert_data = [ |
| 2118 |
'post_id' => $post_id, |
| 2119 |
'user_id' => $user_id, |
| 2120 |
'overall_score' => $score_data['overall_score'], |
| 2121 |
'score_breakdown' => wp_json_encode($score_data['score_breakdown']), |
| 2122 |
'suggestions' => wp_json_encode($score_data['suggestions']), |
| 2123 |
'grade' => $score_data['grade'], |
| 2124 |
'algorithm_version' => $score_data['algorithm_version'] ?? '2024.1', |
| 2125 |
'calculated_at' => $score_data['calculated_at'], |
| 2126 |
'created_at' => current_time('mysql'), |
| 2127 |
]; |
| 2128 |
|
| 2129 |
$format = [ |
| 2130 |
'%d', '%d', '%d', '%s', '%s', '%s', '%s', '%s', '%s' |
| 2131 |
]; |
| 2132 |
|
| 2133 |
// Add readability_score if provided |
| 2134 |
if (isset($score_data['readability_score'])) { |
| 2135 |
$insert_data['readability_score'] = $score_data['readability_score']; |
| 2136 |
$format[] = '%s'; |
| 2137 |
} |
| 2138 |
|
| 2139 |
// Add content_quality if provided |
| 2140 |
if (isset($score_data['content_quality'])) { |
| 2141 |
$insert_data['content_quality'] = $score_data['content_quality']; |
| 2142 |
$format[] = '%s'; |
| 2143 |
} |
| 2144 |
|
| 2145 |
// phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery,WordPress.DB.DirectDatabaseQuery.NoCaching, PluginCheck.Security.DirectDB.UnescapedDBParameter -- SEO score storage requires direct database access |
| 2146 |
$result = $wpdb->insert( |
| 2147 |
$table_name, |
| 2148 |
$insert_data, |
| 2149 |
$format |
| 2150 |
); |
| 2151 |
|
| 2152 |
return $result ? $wpdb->insert_id : false; |
| 2153 |
} |
| 2154 |
|
| 2155 |
/** |
| 2156 |
* Get score history for a post |
| 2157 |
* |
| 2158 |
* @param int $post_id Post ID |
| 2159 |
* @param int $limit Number of scores to retrieve |
| 2160 |
* @return array Score history |
| 2161 |
*/ |
| 2162 |
public function get_score_history(int $post_id, int $limit = 10): array { |
| 2163 |
global $wpdb; |
| 2164 |
|
| 2165 |
// Get table name and escape it properly (table names cannot be parameterized) |
| 2166 |
$table_name = esc_sql($wpdb->prefix . 'thinkrank_seo_scores'); |
| 2167 |
|
| 2168 |
// phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery,WordPress.DB.DirectDatabaseQuery.NoCaching, PluginCheck.Security.DirectDB.UnescapedDBParameter -- SEO score history requires direct database access |
| 2169 |
$results = $wpdb->get_results( |
| 2170 |
$wpdb->prepare( |
| 2171 |
// phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared, PluginCheck.Security.DirectDB.UnescapedDBParameter -- Table name is properly escaped using esc_sql() |
| 2172 |
"SELECT * FROM `{$table_name}` WHERE post_id = %d ORDER BY created_at DESC LIMIT %d", |
| 2173 |
$post_id, |
| 2174 |
$limit |
| 2175 |
), |
| 2176 |
ARRAY_A |
| 2177 |
); |
| 2178 |
|
| 2179 |
// Decode JSON fields |
| 2180 |
foreach ($results as &$result) { |
| 2181 |
$result['score_breakdown'] = json_decode($result['score_breakdown'], true); |
| 2182 |
$result['suggestions'] = json_decode($result['suggestions'], true); |
| 2183 |
} |
| 2184 |
|
| 2185 |
return $results ?: []; |
| 2186 |
} |
| 2187 |
|
| 2188 |
/** |
| 2189 |
* Get latest score for a post |
| 2190 |
* |
| 2191 |
* @param int $post_id Post ID |
| 2192 |
* @return array|null Latest score data |
| 2193 |
*/ |
| 2194 |
public function get_latest_score(int $post_id): ?array { |
| 2195 |
$history = $this->get_score_history($post_id, 1); |
| 2196 |
return !empty($history) ? $history[0] : null; |
| 2197 |
} |
| 2198 |
|
| 2199 |
/** |
| 2200 |
* Score mobile experience - NEW 2025 factor (5 points) |
| 2201 |
* |
| 2202 |
* @param array $content_data Content analysis data |
| 2203 |
* @return array Scoring result |
| 2204 |
*/ |
| 2205 |
private function score_mobile_experience(array $content_data): array { |
| 2206 |
$max = $this->scoring_factors['mobile_experience']; |
| 2207 |
|
| 2208 |
// Mobile experience is theme/site-level, not controlled by post content |
| 2209 |
// — but the plugin already measures it. When a mobile Lighthouse score |
| 2210 |
// has been collected, score against it; the "benefit of the doubt" below |
| 2211 |
// is for sites nobody has measured, not for sites measured as slow. |
| 2212 |
$performance_score = $this->measured_performance_score(); |
| 2213 |
|
| 2214 |
if ($performance_score === null) { |
| 2215 |
return [ |
| 2216 |
'score' => $max, |
| 2217 |
'max_score' => $max, |
| 2218 |
'suggestions' => ['Ensure mobile-first design and fast loading on mobile devices'], |
| 2219 |
'details' => ['mobile_score' => 'Assumed adequate', 'measured' => false], |
| 2220 |
]; |
| 2221 |
} |
| 2222 |
|
| 2223 |
$score = (int) round($max * $performance_score / 100); |
| 2224 |
|
| 2225 |
return [ |
| 2226 |
'score' => $score, |
| 2227 |
'max_score' => $max, |
| 2228 |
'suggestions' => $score < $max |
| 2229 |
? ['Improve mobile page speed: the last PageSpeed run scored ' . $performance_score . '/100 on mobile'] |
| 2230 |
: [], |
| 2231 |
'details' => [ |
| 2232 |
'mobile_score' => $performance_score, |
| 2233 |
'measured' => true, |
| 2234 |
'source' => 'pagespeed_mobile', |
| 2235 |
], |
| 2236 |
]; |
| 2237 |
} |
| 2238 |
|
| 2239 |
/** |
| 2240 |
* The last collected mobile Lighthouse score, or null when unmeasured. |
| 2241 |
* |
| 2242 |
* Memoised per instance: compute_score() asks twice, and a post-list screen |
| 2243 |
* scores a page of posts at a time. |
| 2244 |
* |
| 2245 |
* Every failure — no performance module, no collected row, an unreadable |
| 2246 |
* table — resolves to null, which the callers read as "not measured" and |
| 2247 |
* answer with the full-credit fallback. A site is never penalised for |
| 2248 |
* ThinkRank being unable to look. |
| 2249 |
* |
| 2250 |
* @since 2.3.1 |
| 2251 |
* @return int|null Score 0-100, or null when nothing has been collected. |
| 2252 |
*/ |
| 2253 |
private function measured_performance_score(): ?int { |
| 2254 |
$measurement = $this->measured_performance(); |
| 2255 |
|
| 2256 |
if ($measurement === null || !isset($measurement['performance_score'])) { |
| 2257 |
return null; |
| 2258 |
} |
| 2259 |
|
| 2260 |
$score = $measurement['performance_score']; |
| 2261 |
|
| 2262 |
if (!is_numeric($score)) { |
| 2263 |
return null; |
| 2264 |
} |
| 2265 |
|
| 2266 |
return (int) round(max(0, min(100, (float) $score))); |
| 2267 |
} |
| 2268 |
|
| 2269 |
/** |
| 2270 |
* The last collected mobile measurement, or null when there is none. |
| 2271 |
* |
| 2272 |
* @since 2.3.1 |
| 2273 |
* @return array|null { core_web_vitals: array, performance_score: float|null } |
| 2274 |
*/ |
| 2275 |
private function measured_performance(): ?array { |
| 2276 |
if ($this->measured_performance_resolved) { |
| 2277 |
return $this->measured_performance; |
| 2278 |
} |
| 2279 |
|
| 2280 |
$this->measured_performance_resolved = true; |
| 2281 |
|
| 2282 |
if (!class_exists('ThinkRank\\SEO\\Performance_Monitoring_Manager')) { |
| 2283 |
return null; |
| 2284 |
} |
| 2285 |
|
| 2286 |
try { |
| 2287 |
$manager = new \ThinkRank\SEO\Performance_Monitoring_Manager(); |
| 2288 |
// Mobile deliberately: Google indexes mobile-first, and it is the |
| 2289 |
// device the mobile_experience factor is named after. |
| 2290 |
$this->measured_performance = $manager->get_stored_performance_measurement('mobile'); |
| 2291 |
} catch (\Throwable $e) { |
| 2292 |
$this->measured_performance = null; |
| 2293 |
} |
| 2294 |
|
| 2295 |
return $this->measured_performance; |
| 2296 |
} |
| 2297 |
|
| 2298 |
/** |
| 2299 |
* Score core web vitals - 2025 version (3 points) |
| 2300 |
* |
| 2301 |
* @param array $content_data Content analysis data |
| 2302 |
* @return array Scoring result |
| 2303 |
*/ |
| 2304 |
private function score_core_web_vitals(array $content_data): array { |
| 2305 |
$max = $this->scoring_factors['core_web_vitals']; |
| 2306 |
|
| 2307 |
// Not derivable from post content — but it is measured, and the audit |
| 2308 |
// stores LCP, INP and CLS with a rating each. Score against those when |
| 2309 |
// they exist; fall back to the benefit of the doubt when they do not. |
| 2310 |
$rated = $this->measured_vitals_score(); |
| 2311 |
|
| 2312 |
if ($rated === null) { |
| 2313 |
return [ |
| 2314 |
'score' => $max, |
| 2315 |
'max_score' => $max, |
| 2316 |
'suggestions' => ['Optimize Core Web Vitals: LCP, INP, and CLS for better user experience'], |
| 2317 |
'details' => ['vitals_status' => 'Assumed adequate', 'measured' => false], |
| 2318 |
]; |
| 2319 |
} |
| 2320 |
|
| 2321 |
$score = (int) round($max * $rated['average'] / 100); |
| 2322 |
|
| 2323 |
return [ |
| 2324 |
'score' => $score, |
| 2325 |
'max_score' => $max, |
| 2326 |
// Gated on the measurement, not the rounded score: two good metrics |
| 2327 |
// and one needing improvement averages 88.33, which rounds to the |
| 2328 |
// full 3 of 3 and used to swallow the suggestion naming the metric |
| 2329 |
// that is actually failing. |
| 2330 |
'suggestions' => !empty($rated['failing']) |
| 2331 |
? ['Optimize Core Web Vitals: ' . implode(', ', $rated['failing']) . ' below target on mobile'] |
| 2332 |
: [], |
| 2333 |
'details' => [ |
| 2334 |
'vitals_status' => $rated['statuses'], |
| 2335 |
'measured' => true, |
| 2336 |
'source' => 'pagespeed_mobile', |
| 2337 |
], |
| 2338 |
]; |
| 2339 |
} |
| 2340 |
|
| 2341 |
/** |
| 2342 |
* Rate the collected Core Web Vitals, or null when none were measured. |
| 2343 |
* |
| 2344 |
* Reuses the per-metric score the performance module already assigns |
| 2345 |
* (good 100, needs improvement 65, poor 30) rather than inventing a second |
| 2346 |
* scale, so the SEO score and the performance card cannot disagree about |
| 2347 |
* whether a metric is healthy. |
| 2348 |
* |
| 2349 |
* Metrics with no stored value — fcp is not always collected — are skipped |
| 2350 |
* rather than counted as failures. |
| 2351 |
* |
| 2352 |
* @since 2.3.1 |
| 2353 |
* @return array|null { average: float, statuses: array, failing: string[] } |
| 2354 |
*/ |
| 2355 |
private function measured_vitals_score(): ?array { |
| 2356 |
$measurement = $this->measured_performance(); |
| 2357 |
$vitals = $measurement['core_web_vitals'] ?? null; |
| 2358 |
|
| 2359 |
if (!is_array($vitals)) { |
| 2360 |
return null; |
| 2361 |
} |
| 2362 |
|
| 2363 |
$scores = []; |
| 2364 |
$statuses = []; |
| 2365 |
$failing = []; |
| 2366 |
|
| 2367 |
// The three Google ranks on. fcp is diagnostic and not a Core Web Vital. |
| 2368 |
foreach (['lcp', 'inp', 'cls'] as $metric) { |
| 2369 |
$data = $vitals[$metric] ?? null; |
| 2370 |
|
| 2371 |
if (!is_array($data) || !isset($data['value'], $data['score']) || $data['value'] === null) { |
| 2372 |
continue; |
| 2373 |
} |
| 2374 |
|
| 2375 |
$scores[] = (float) $data['score']; |
| 2376 |
$statuses[$metric] = $data['status'] ?? 'unknown'; |
| 2377 |
|
| 2378 |
if (($data['status'] ?? '') !== 'good') { |
| 2379 |
$failing[] = strtoupper($metric); |
| 2380 |
} |
| 2381 |
} |
| 2382 |
|
| 2383 |
if (empty($scores)) { |
| 2384 |
return null; |
| 2385 |
} |
| 2386 |
|
| 2387 |
return [ |
| 2388 |
'average' => array_sum($scores) / count($scores), |
| 2389 |
'statuses' => $statuses, |
| 2390 |
'failing' => $failing, |
| 2391 |
]; |
| 2392 |
} |
| 2393 |
|
| 2394 |
/** |
| 2395 |
* Score internal linking - declining importance (1 point) |
| 2396 |
* |
| 2397 |
* @param array $content_data Content analysis data |
| 2398 |
* @return array Scoring result |
| 2399 |
*/ |
| 2400 |
private function score_internal_linking(array $content_data): array { |
| 2401 |
$internal_links = $content_data['internal_links'] ?? 0; |
| 2402 |
$score = $internal_links > 0 ? 1 : 0; |
| 2403 |
|
| 2404 |
return [ |
| 2405 |
'score' => $score, |
| 2406 |
'max_score' => $this->scoring_factors['internal_linking'], |
| 2407 |
'suggestions' => $score === 0 ? ['Add relevant internal links to other pages on your site'] : [], |
| 2408 |
'details' => ['internal_links_count' => $internal_links] |
| 2409 |
]; |
| 2410 |
} |
| 2411 |
|
| 2412 |
/** |
| 2413 |
* Score technical factors (1 point) |
| 2414 |
* |
| 2415 |
* @param array $content_data Content analysis data |
| 2416 |
* @param array $metadata Post metadata |
| 2417 |
* @param array $options Additional options |
| 2418 |
* @return array Scoring result |
| 2419 |
*/ |
| 2420 |
private function score_technical_factors(array $content_data, array $metadata, array $options): array { |
| 2421 |
$score = 0; |
| 2422 |
$suggestions = []; |
| 2423 |
|
| 2424 |
// Meta description check |
| 2425 |
$meta_desc = $metadata['description'] ?? ''; |
| 2426 |
if (!empty($meta_desc) && mb_strlen($meta_desc) >= self::DESCRIPTION_OPTIMAL_MIN && mb_strlen($meta_desc) <= self::DESCRIPTION_OPTIMAL_MAX) { |
| 2427 |
$score += 0.5; |
| 2428 |
} else { |
| 2429 |
$suggestions[] = 'Add a compelling meta description (120-160 characters)'; |
| 2430 |
} |
| 2431 |
|
| 2432 |
// Schema markup check (simplified) |
| 2433 |
if (!empty($content_data['schema_present'])) { |
| 2434 |
$score += 0.5; |
| 2435 |
} else { |
| 2436 |
$suggestions[] = 'Consider adding structured data (schema markup)'; |
| 2437 |
} |
| 2438 |
|
| 2439 |
return [ |
| 2440 |
'score' => $score, |
| 2441 |
'max_score' => $this->scoring_factors['technical_factors'], |
| 2442 |
'suggestions' => $suggestions, |
| 2443 |
'details' => [ |
| 2444 |
'meta_description_length' => mb_strlen($meta_desc), |
| 2445 |
'schema_present' => !empty($content_data['schema_present']) |
| 2446 |
] |
| 2447 |
]; |
| 2448 |
} |
| 2449 |
|
| 2450 |
/** |
| 2451 |
* Get grade from score |
| 2452 |
* |
| 2453 |
* @param mixed $score Numeric score |
| 2454 |
* @return string Letter grade |
| 2455 |
*/ |
| 2456 |
private function get_grade_from_score($score): string { |
| 2457 |
$score = (int) $score; // Ensure it's an integer |
| 2458 |
|
| 2459 |
if ($score >= 95) { return 'A+'; |
| 2460 |
} |
| 2461 |
if ($score >= 90) { return 'A'; |
| 2462 |
} |
| 2463 |
if ($score >= 85) { return 'A-'; |
| 2464 |
} |
| 2465 |
if ($score >= 80) { return 'B+'; |
| 2466 |
} |
| 2467 |
if ($score >= 75) { return 'B'; |
| 2468 |
} |
| 2469 |
if ($score >= 70) { return 'B-'; |
| 2470 |
} |
| 2471 |
if ($score >= 65) { return 'C+'; |
| 2472 |
} |
| 2473 |
if ($score >= 60) { return 'C'; |
| 2474 |
} |
| 2475 |
if ($score >= 55) { return 'C-'; |
| 2476 |
} |
| 2477 |
if ($score >= 45) { return 'D+'; |
| 2478 |
} |
| 2479 |
if ($score >= 35) { return 'D'; |
| 2480 |
} |
| 2481 |
return 'F'; |
| 2482 |
} |
| 2483 |
|
| 2484 |
/** |
| 2485 |
* Analyze live content from editor (not saved to database yet) |
| 2486 |
* Same as analyze_post_content but uses provided content instead of saved content |
| 2487 |
* |
| 2488 |
* @param string $live_content Live content from editor |
| 2489 |
* @param int $post_id Post ID for metadata |
| 2490 |
* @return array Content analysis data |
| 2491 |
*/ |
| 2492 |
public function analyze_live_content(string $live_content, int $post_id): array { |
| 2493 |
$post = get_post($post_id); |
| 2494 |
if (!$post) { |
| 2495 |
return []; |
| 2496 |
} |
| 2497 |
|
| 2498 |
// Resolve the live string the same way stored content is resolved. On a |
| 2499 |
// builder page the editor hands over raw builder markup (the block |
| 2500 |
// editor cannot render blocks it has no client-side registration for), |
| 2501 |
// which analyzed as-is reads as zero words — the reason a Divi page |
| 2502 |
// could show a correct saved score beside a live panel still claiming |
| 2503 |
// "No content". |
| 2504 |
$content = self::resolve_live_content($live_content, $post); |
| 2505 |
$title = $post->post_title; |
| 2506 |
|
| 2507 |
// Extract headings from content |
| 2508 |
$headings = $this->extract_headings($content); |
| 2509 |
|
| 2510 |
// Count the markup, not wp_strip_all_tags() output: stripping deletes |
| 2511 |
// tags without a space, so "five</p><p>six" would already be one word |
| 2512 |
// before the counter saw it. |
| 2513 |
$word_count = $this->calculate_word_count_js_style($content); |
| 2514 |
|
| 2515 |
// Calculate readability |
| 2516 |
$readability_score = $this->calculate_readability_score($content); |
| 2517 |
|
| 2518 |
// Count links |
| 2519 |
$internal_links = $this->count_internal_links($content); |
| 2520 |
$external_links = $this->count_external_links($content); |
| 2521 |
$external_dofollow_links = $this->count_external_dofollow_links($content); |
| 2522 |
|
| 2523 |
// Analyze images |
| 2524 |
$images = $this->analyze_images($content); |
| 2525 |
|
| 2526 |
// Get URL |
| 2527 |
$url = get_permalink($post_id); |
| 2528 |
|
| 2529 |
return [ |
| 2530 |
'content' => $content, |
| 2531 |
'title' => $title, |
| 2532 |
'headings' => $headings, |
| 2533 |
'word_count' => $word_count, |
| 2534 |
'readability_score' => $readability_score, |
| 2535 |
'internal_links' => $internal_links, |
| 2536 |
'external_links' => $external_links, |
| 2537 |
'external_dofollow_links' => $external_dofollow_links, |
| 2538 |
'images' => $images, |
| 2539 |
'url' => $url, |
| 2540 |
'slug' => $post->post_name, |
| 2541 |
'post_modified' => $post->post_modified, |
| 2542 |
'schema_present' => $this->detect_schema_present($content) |
| 2543 |
|| $this->thinkrank_global_schema_active($post->post_type) |
| 2544 |
|| $this->thinkrank_deployed_schema_active($post), |
| 2545 |
]; |
| 2546 |
} |
| 2547 |
|
| 2548 |
/** |
| 2549 |
* Count the words a reader reads in HTML or text. |
| 2550 |
* |
| 2551 |
* The same extraction and counting as the thin content report |
| 2552 |
* ({@see \ThinkRank\SEO\Word_Count_Index::reading_text()} and |
| 2553 |
* {@see \ThinkRank\SEO\Word_Count_Index::count_words()}), so the editor |
| 2554 |
* score and the report give one number for one page. Splitting on |
| 2555 |
* whitespace counted shortcode syntax as words: a WPBakery page with 216 |
| 2556 |
* words of prose scored a word count of 325 while the report said 216 |
| 2557 |
* (#893 fixed the report only). It also counted tokens of punctuation |
| 2558 |
* alone, such as a full stop after a link. |
| 2559 |
* |
| 2560 |
* Always words, whatever the locale's unit, because every threshold that |
| 2561 |
* reads this value (content length, long paragraphs, headings per 300 |
| 2562 |
* words, readability, keyword density) is in words. |
| 2563 |
* |
| 2564 |
* contentAnalysis.js calculateWordCount() applies the same two rules in |
| 2565 |
* the editor. |
| 2566 |
* |
| 2567 |
* @param string $text HTML or text to count words in. |
| 2568 |
* @return int Word count |
| 2569 |
*/ |
| 2570 |
private function calculate_word_count_js_style(string $text): int { |
| 2571 |
if ('' === $text) { |
| 2572 |
return 0; |
| 2573 |
} |
| 2574 |
|
| 2575 |
return \ThinkRank\SEO\Word_Count_Index::count_words(self::reading_text_of($text)); |
| 2576 |
} |
| 2577 |
|
| 2578 |
/** |
| 2579 |
* The text a reader reads in HTML or text, as the word count sees it. |
| 2580 |
* |
| 2581 |
* Every check that divides by the word count (readability, keyword |
| 2582 |
* density, topic relevance) reads its numerator from this same text, so |
| 2583 |
* shortcode syntax is never on one side of a ratio and not the other. |
| 2584 |
* |
| 2585 |
* @since 2.14.2 |
| 2586 |
* |
| 2587 |
* @param string $content HTML or text. |
| 2588 |
* @return string Plain text, whitespace collapsed. |
| 2589 |
*/ |
| 2590 |
private static function reading_text_of(string $content): string { |
| 2591 |
if ('' === $content) { |
| 2592 |
return ''; |
| 2593 |
} |
| 2594 |
|
| 2595 |
if (!class_exists('\ThinkRank\SEO\Word_Count_Index')) { |
| 2596 |
require_once THINKRANK_PLUGIN_DIR . 'includes/seo/class-word-count-index.php'; |
| 2597 |
} |
| 2598 |
|
| 2599 |
return \ThinkRank\SEO\Word_Count_Index::reading_text($content); |
| 2600 |
} |
| 2601 |
|
| 2602 |
/** |
| 2603 |
* Get existing score data for a post from database |
| 2604 |
* |
| 2605 |
* @param int $post_id Post ID |
| 2606 |
* @return array|null Existing score data or null if not found |
| 2607 |
*/ |
| 2608 |
public function get_existing_score_data(int $post_id): ?array { |
| 2609 |
global $wpdb; |
| 2610 |
|
| 2611 |
// Get table name and escape it properly (table names cannot be parameterized) |
| 2612 |
$table_name = esc_sql($wpdb->prefix . 'thinkrank_seo_scores'); |
| 2613 |
|
| 2614 |
// Get the most recent score for this post |
| 2615 |
// phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery,WordPress.DB.DirectDatabaseQuery.NoCaching, PluginCheck.Security.DirectDB.UnescapedDBParameter -- SEO score retrieval requires direct database access |
| 2616 |
$result = $wpdb->get_row($wpdb->prepare( |
| 2617 |
// phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared, PluginCheck.Security.DirectDB.UnescapedDBParameter -- Table name is properly escaped using esc_sql() |
| 2618 |
"SELECT * FROM `{$table_name}` |
| 2619 |
WHERE post_id = %d |
| 2620 |
ORDER BY calculated_at DESC |
| 2621 |
LIMIT 1", |
| 2622 |
$post_id |
| 2623 |
), ARRAY_A); |
| 2624 |
|
| 2625 |
if (!$result) { |
| 2626 |
return null; |
| 2627 |
} |
| 2628 |
|
| 2629 |
// Decode JSON data (stored with json_encode) |
| 2630 |
$score_breakdown = json_decode($result['score_breakdown'], true); |
| 2631 |
$suggestions = json_decode($result['suggestions'], true); |
| 2632 |
|
| 2633 |
// Format the data to match the expected structure |
| 2634 |
return [ |
| 2635 |
'overall_score' => (int) $result['overall_score'], |
| 2636 |
'grade' => $result['grade'], |
| 2637 |
'score_breakdown' => $score_breakdown, |
| 2638 |
'suggestions' => $suggestions ?: [], |
| 2639 |
'target_keyword' => null, // Not stored in database, will be provided by frontend |
| 2640 |
'calculated_at' => $result['calculated_at'], |
| 2641 |
'score_id' => $result['id'] |
| 2642 |
]; |
| 2643 |
} |
| 2644 |
} |
| 2645 |
|
