Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
100.00% |
243 / 243 |
|
100.00% |
15 / 15 |
CRAP | |
100.00% |
1 / 1 |
| NameParser | |
100.00% |
243 / 243 |
|
100.00% |
15 / 15 |
92 | |
100.00% |
1 / 1 |
| parse | |
100.00% |
39 / 39 |
|
100.00% |
1 / 1 |
8 | |||
| clean | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| tokenize | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| normalizeCase | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
3 | |||
| normalizeWords | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| parseCommaMode | |
100.00% |
25 / 25 |
|
100.00% |
1 / 1 |
6 | |||
| absorbLeftovers | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
3 | |||
| finalizeInitials | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
4 | |||
| extractNickname | |
100.00% |
24 / 24 |
|
100.00% |
1 / 1 |
8 | |||
| extractSalutation | |
100.00% |
29 / 29 |
|
100.00% |
1 / 1 |
11 | |||
| extractSuffix | |
100.00% |
32 / 32 |
|
100.00% |
1 / 1 |
12 | |||
| extractInitials | |
100.00% |
24 / 24 |
|
100.00% |
1 / 1 |
14 | |||
| extractLastname | |
100.00% |
33 / 33 |
|
100.00% |
1 / 1 |
16 | |||
| extractFirstname | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
2 | |||
| extractMiddlename | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| 1 | <?php |
| 2 | declare(strict_types=1); |
| 3 | /** |
| 4 | * Pop PHP Framework (https://www.popphp.org/) |
| 5 | * |
| 6 | * @link https://github.com/popphp/popphp-framework |
| 7 | * @author Nick Sagona, III <nick@popphp.org> |
| 8 | * @copyright Copyright (c) 2009-2026 Nick Sagona, III |
| 9 | * @license https://www.popphp.org/license New BSD License |
| 10 | */ |
| 11 | |
| 12 | /** |
| 13 | * @namespace |
| 14 | */ |
| 15 | namespace Pop\Parser\Name; |
| 16 | |
| 17 | use Pop\Parser\AbstractParser; |
| 18 | use Pop\Parser\Exception; |
| 19 | |
| 20 | /** |
| 21 | * Name parser class |
| 22 | * |
| 23 | * @category Pop |
| 24 | * @package Pop\Parser |
| 25 | * @author Nick Sagona, III <nick@popphp.org> |
| 26 | * @copyright Copyright (c) 2009-2026 Nick Sagona, III |
| 27 | * @license https://www.popphp.org/license New BSD License |
| 28 | * @version 1.0.0 |
| 29 | */ |
| 30 | class NameParser extends AbstractParser |
| 31 | { |
| 32 | |
| 33 | /** |
| 34 | * Parse method |
| 35 | * |
| 36 | * Builds up the parsed fields in a local $fields array (and a local $initialsQueue), |
| 37 | * threaded by reference through each extraction step, rather than on $this - the |
| 38 | * parser holds no parsed-field state of its own. The very last step wraps that array |
| 39 | * into an immutable NameResult, which is what's actually returned. |
| 40 | * |
| 41 | * @param ?string $name |
| 42 | * @throws Exception |
| 43 | * @return NameResult |
| 44 | */ |
| 45 | public function parse(?string $name = null): NameResult |
| 46 | { |
| 47 | if (empty($this->data) && empty($name)) { |
| 48 | throw new Exception('Error: You must pass a name string to the parser object.'); |
| 49 | } |
| 50 | |
| 51 | if ((null === $name) && !empty($this->data)) { |
| 52 | $name = $this->data; |
| 53 | } else if (null !== $name) { |
| 54 | $this->data = $name; |
| 55 | } |
| 56 | |
| 57 | $name = $this->clean($name); |
| 58 | |
| 59 | if ($name === '') { |
| 60 | throw new Exception('Error: You must pass a name string to the parser object.'); |
| 61 | } |
| 62 | |
| 63 | $fields = [ |
| 64 | 'salutation' => null, |
| 65 | 'firstname' => null, |
| 66 | 'middlename' => null, |
| 67 | 'nickname' => null, |
| 68 | 'initials' => null, |
| 69 | 'lastnamePrefix' => null, |
| 70 | 'lastname' => null, |
| 71 | 'suffix' => null, |
| 72 | 'credentials' => null, |
| 73 | ]; |
| 74 | $initialsQueue = []; |
| 75 | |
| 76 | $nameValues = new NameValues(); |
| 77 | |
| 78 | if (str_contains($name, ',')) { |
| 79 | $this->parseCommaMode($name, $nameValues, $fields, $initialsQueue); |
| 80 | } else { |
| 81 | $tokens = $this->tokenize($name); |
| 82 | $originalCount = count($tokens); |
| 83 | $tokens = $this->extractNickname($tokens, $nameValues, $fields); |
| 84 | $tokens = $this->extractSalutation($tokens, $nameValues, $fields); |
| 85 | $tokens = $this->extractSuffix($tokens, $nameValues, $fields, 2); |
| 86 | $tokens = $this->extractInitials($tokens, false, $nameValues, $initialsQueue); |
| 87 | $tokens = $this->extractLastname($tokens, $nameValues, $fields, false, $originalCount); |
| 88 | $tokens = $this->extractFirstname($tokens, $fields); |
| 89 | $tokens = $this->extractMiddlename($tokens, $fields); |
| 90 | $this->absorbLeftovers($tokens, $fields); |
| 91 | } |
| 92 | |
| 93 | $this->finalizeInitials($fields, $initialsQueue); |
| 94 | |
| 95 | $fields['confidence'] = $this->calculateConfidence($fields['confidenceSignalCount'] ?? 0); |
| 96 | unset($fields['confidenceSignalCount']); |
| 97 | |
| 98 | $this->result = new NameResult($fields); |
| 99 | |
| 100 | return $this->result; |
| 101 | } |
| 102 | |
| 103 | /** |
| 104 | * Clean method |
| 105 | * |
| 106 | * @param string $name |
| 107 | * @return string |
| 108 | */ |
| 109 | public function clean(string $name): string |
| 110 | { |
| 111 | return trim(preg_replace('/\s+/', ' ', $name)); |
| 112 | } |
| 113 | |
| 114 | /** |
| 115 | * Tokenize method |
| 116 | * |
| 117 | * Splits on whitespace and drops any resulting empty tokens - not just the ordinary "no |
| 118 | * internal double spaces" case, but the edge case where $name is itself empty (e.g. an |
| 119 | * empty comma-mode segment from a leading/trailing/doubled comma, like ",John Smith,"), |
| 120 | * where preg_split would otherwise yield a single spurious "" token instead of no tokens |
| 121 | * at all, which downstream steps would then treat as a real (blank) word to claim. |
| 122 | * |
| 123 | * @param string $name |
| 124 | * @return array |
| 125 | */ |
| 126 | protected function tokenize(string $name): array |
| 127 | { |
| 128 | return array_values(array_filter(preg_split('/\s+/', trim($name)), fn($token) => $token !== '')); |
| 129 | } |
| 130 | |
| 131 | /** |
| 132 | * Normalize the case of a single word: an all-uppercase or all-lowercase word gets |
| 133 | * title-cased ("MACDONALD" / "macdonald" -> "Macdonald"); a word with any existing |
| 134 | * mixed case (e.g. "MacDonald", "McDonald", "O'Brien") is left exactly as typed, since |
| 135 | * that mixed case is almost always deliberate. "O'Brien"-style words are already handled |
| 136 | * correctly by the title-casing above (the apostrophe splits it into two letter-runs, |
| 137 | * each capitalized independently) with no extra logic needed. "Mc" gets one further, |
| 138 | * explicit fix-up: capitalize the letter right after it ("Mcdonald" -> "McDonald"), since |
| 139 | * it's a reliable surname-prefix marker in English with very few false positives. "Mac" is |
| 140 | * deliberately NOT special-cased the same way - it's also the start of many ordinary |
| 141 | * names/words ("Macy", "Mack", "Macon") where blindly capitalizing the next letter would |
| 142 | * be wrong more often than right. |
| 143 | * |
| 144 | * @param string $word |
| 145 | * @return string |
| 146 | */ |
| 147 | protected function normalizeCase(string $word): string |
| 148 | { |
| 149 | $stripped = str_replace('.', '', $word); |
| 150 | |
| 151 | if (($stripped === mb_strtoupper($stripped)) || ($stripped === mb_strtolower($stripped))) { |
| 152 | $titled = preg_replace_callback('/\p{L}+/u', function ($matches) { |
| 153 | return mb_convert_case($matches[0], MB_CASE_TITLE); |
| 154 | }, $word); |
| 155 | |
| 156 | return preg_replace_callback('/\bMc(\p{L})/u', function ($matches) { |
| 157 | return 'Mc' . mb_strtoupper($matches[1]); |
| 158 | }, $titled); |
| 159 | } |
| 160 | |
| 161 | return $word; |
| 162 | } |
| 163 | |
| 164 | /** |
| 165 | * Normalize the case of each word in an array and join them with a space |
| 166 | * |
| 167 | * @param array $words |
| 168 | * @return string |
| 169 | */ |
| 170 | protected function normalizeWords(array $words): string |
| 171 | { |
| 172 | return implode(' ', array_map([$this, 'normalizeCase'], $words)); |
| 173 | } |
| 174 | |
| 175 | /** |
| 176 | * Parse comma-separated "Last, First Middle[, Suffix]" format |
| 177 | * |
| 178 | * @param string $name |
| 179 | * @param NameValues $nameValues |
| 180 | * @param array $fields |
| 181 | * @param array $initialsQueue |
| 182 | * @return void |
| 183 | */ |
| 184 | protected function parseCommaMode(string $name, NameValues $nameValues, array &$fields, array &$initialsQueue): void |
| 185 | { |
| 186 | $segments = array_map('trim', explode(',', $name)); |
| 187 | |
| 188 | // Segment 1 (before the first comma): the lastname segment. Salutation, suffix and |
| 189 | // lastname(-with-prefix) extraction run here. Whatever's left over (e.g. "Garcia" in |
| 190 | // "Garcia Marquez, Gabriel", or a lastname-prefix word extractLastname's ordinary |
| 191 | // guards wouldn't fold at position 0) is deliberately NOT assigned to firstname here |
| 192 | // - it's carried forward and absorbed only after segment 2 runs, so it can never |
| 193 | // silently overwrite or lose to segment 2's own firstname; absorbLeftovers() merges |
| 194 | // it into middlename once firstname is already set instead. |
| 195 | $segment1 = $this->tokenize($segments[0]); |
| 196 | $originalCount1 = count($segment1); |
| 197 | $segment1 = $this->extractSalutation($segment1, $nameValues, $fields); |
| 198 | $segment1 = $this->extractSuffix($segment1, $nameValues, $fields, 0, true); |
| 199 | $segment1 = $this->extractLastname($segment1, $nameValues, $fields, true, $originalCount1); |
| 200 | |
| 201 | // Segment 2 (between commas, or everything after the first comma): the given-name segment |
| 202 | if (isset($segments[1]) && ($segments[1] !== '')) { |
| 203 | $segment2 = $this->tokenize($segments[1]); |
| 204 | $segment2 = $this->extractSalutation($segment2, $nameValues, $fields); |
| 205 | $segment2 = $this->extractSuffix($segment2, $nameValues, $fields, 0, true, true); |
| 206 | $segment2 = $this->extractNickname($segment2, $nameValues, $fields); |
| 207 | $segment2 = $this->extractInitials($segment2, true, $nameValues, $initialsQueue); |
| 208 | $segment2 = $this->extractFirstname($segment2, $fields); |
| 209 | $segment2 = $this->extractMiddlename($segment2, $fields); |
| 210 | $this->absorbLeftovers($segment2, $fields); |
| 211 | } |
| 212 | |
| 213 | // Now that segment 2 has had its chance to claim the firstname slot: if it didn't |
| 214 | // (firstname is still null), segment 1's leftover IS the given name, so split it the |
| 215 | // normal way (first word -> firstname, the rest -> middlename) rather than joining it |
| 216 | // into one string. If segment 2 already provided a firstname, segment 1's leftover is |
| 217 | // secondary content - absorbLeftovers() merges the whole thing into middlename. |
| 218 | if ($fields['firstname'] === null) { |
| 219 | $segment1 = $this->extractFirstname($segment1, $fields); |
| 220 | $segment1 = $this->extractMiddlename($segment1, $fields); |
| 221 | } |
| 222 | $this->absorbLeftovers($segment1, $fields); |
| 223 | |
| 224 | // Segment 3 onward (after a second comma, if present): suffix only per segment, with |
| 225 | // any non-suffix leftover absorbed rather than discarded (e.g. "Smith, John, Michael" |
| 226 | // must not lose "Michael" just because it isn't a recognized suffix). Looping over |
| 227 | // every remaining segment - not just $segments[2] - means a name with more than 3 |
| 228 | // comma-separated parts (e.g. "Smith, John, PhD, Esq") doesn't silently drop |
| 229 | // anything past the third. |
| 230 | foreach (array_slice($segments, 2) as $segment) { |
| 231 | if ($segment === '') { |
| 232 | continue; |
| 233 | } |
| 234 | $extraSegment = $this->tokenize($segment); |
| 235 | $extraSegment = $this->extractSuffix($extraSegment, $nameValues, $fields, 0, true); |
| 236 | $this->absorbLeftovers($extraSegment, $fields); |
| 237 | } |
| 238 | } |
| 239 | |
| 240 | /** |
| 241 | * Absorb any tokens no extraction step claimed: the first time this is called with a |
| 242 | * non-empty leftover, it becomes firstname (since every name needs one); after that, |
| 243 | * leftovers are appended to middlename. This is a deliberate difference from |
| 244 | * theiconic/name-parser, which can silently drop an unrecognized leading word (e.g. "The" |
| 245 | * in "The Rev. Mark Williams") - here nothing is ever discarded. That second case (a |
| 246 | * leftover appended to middlename) is a confidence signal - it means content exists that |
| 247 | * didn't fit any recognized category and had to be defaulted into middlename, e.g. the |
| 248 | * "Garcia" in "Garcia Marquez, Gabriel" not being clearly lastname or given-name content. |
| 249 | * |
| 250 | * @param array $tokens |
| 251 | * @param array $fields |
| 252 | * @return void |
| 253 | */ |
| 254 | protected function absorbLeftovers(array $tokens, array &$fields): void |
| 255 | { |
| 256 | if (empty($tokens)) { |
| 257 | return; |
| 258 | } |
| 259 | |
| 260 | $text = $this->normalizeWords($tokens); |
| 261 | |
| 262 | if ($fields['firstname'] === null) { |
| 263 | $fields['firstname'] = $text; |
| 264 | } else { |
| 265 | $fields['middlename'] = trim(($fields['middlename'] ?? '') . ' ' . $text); |
| 266 | $fields['confidenceSignalCount'] = ($fields['confidenceSignalCount'] ?? 0) + 1; |
| 267 | } |
| 268 | } |
| 269 | |
| 270 | /** |
| 271 | * If no raw token ever became firstname but one or more initials were set aside, promote |
| 272 | * the FIRST claimed initial back to firstname - handles "J. B. Hunt" (firstname="J.", |
| 273 | * initials="B.", lastname="Hunt"), since a name consisting only of initials plus a |
| 274 | * lastname still needs a firstname. That promotion is itself a confidence signal - the |
| 275 | * name never actually supplied a real firstname token, just an initial standing in for one. |
| 276 | * |
| 277 | * @param array $fields |
| 278 | * @param array $initialsQueue |
| 279 | * @return void |
| 280 | */ |
| 281 | protected function finalizeInitials(array &$fields, array &$initialsQueue): void |
| 282 | { |
| 283 | if (empty($initialsQueue)) { |
| 284 | return; |
| 285 | } |
| 286 | |
| 287 | if ($fields['firstname'] === null) { |
| 288 | $promoted = array_shift($initialsQueue); |
| 289 | $fields['firstname'] = $this->normalizeCase($promoted); |
| 290 | $fields['confidenceSignalCount'] = ($fields['confidenceSignalCount'] ?? 0) + 1; |
| 291 | } |
| 292 | |
| 293 | if (!empty($initialsQueue)) { |
| 294 | $fields['initials'] = trim(($fields['initials'] ?? '') . ' ' . $this->normalizeWords($initialsQueue)); |
| 295 | } |
| 296 | |
| 297 | $initialsQueue = []; |
| 298 | } |
| 299 | |
| 300 | /** |
| 301 | * Extract nickname |
| 302 | * |
| 303 | * Scans for a token starting with an opening delimiter and collects tokens until one |
| 304 | * ends with the matching closing delimiter (supports multi-word nicknames). |
| 305 | * |
| 306 | * @param array $tokens |
| 307 | * @param NameValues $nameValues |
| 308 | * @param array $fields |
| 309 | * @return array |
| 310 | */ |
| 311 | protected function extractNickname(array $tokens, NameValues $nameValues, array &$fields): array |
| 312 | { |
| 313 | $delimiters = $nameValues->getNicknameDelimiters(); |
| 314 | $openChars = array_keys($delimiters); |
| 315 | |
| 316 | foreach ($tokens as $start => $token) { |
| 317 | $firstChar = mb_substr($token, 0, 1); |
| 318 | if (!in_array($firstChar, $openChars, true)) { |
| 319 | continue; |
| 320 | } |
| 321 | |
| 322 | $closeChar = $delimiters[$firstChar]; |
| 323 | $end = null; |
| 324 | |
| 325 | for ($i = $start; $i < count($tokens); $i++) { |
| 326 | if (str_ends_with($tokens[$i], $closeChar) && (($i > $start) || (strlen($tokens[$i]) > 1))) { |
| 327 | $end = $i; |
| 328 | break; |
| 329 | } |
| 330 | } |
| 331 | |
| 332 | if ($end === null) { |
| 333 | continue; |
| 334 | } |
| 335 | |
| 336 | $span = array_slice($tokens, $start, $end - $start + 1); |
| 337 | $span[0] = ltrim($span[0], $firstChar); |
| 338 | $lastIndex = count($span) - 1; |
| 339 | $span[$lastIndex] = rtrim($span[$lastIndex], $closeChar); |
| 340 | $span = array_map(fn($word) => trim($word, '\'"'), $span); |
| 341 | $span = array_filter($span, fn($word) => $word !== ''); |
| 342 | |
| 343 | $fields['nickname'] = $this->normalizeWords($span); |
| 344 | array_splice($tokens, $start, $end - $start + 1); |
| 345 | break; |
| 346 | } |
| 347 | |
| 348 | return array_values($tokens); |
| 349 | } |
| 350 | |
| 351 | /** |
| 352 | * Extract salutation |
| 353 | * |
| 354 | * Scans from the start of the tokens (bounded to roughly the first half) for matches |
| 355 | * against the salutation list, checked as both single tokens and multi-word phrases. |
| 356 | * Multiple consecutive salutations (e.g. "Rev. Dr John Doe") are all claimed. |
| 357 | * |
| 358 | * @param array $tokens |
| 359 | * @param NameValues $nameValues |
| 360 | * @param array $fields |
| 361 | * @return array |
| 362 | */ |
| 363 | protected function extractSalutation(array $tokens, NameValues $nameValues, array &$fields): array |
| 364 | { |
| 365 | $salutations = []; |
| 366 | foreach ($nameValues->getSalutations() as $key => $display) { |
| 367 | $keyWords = explode(' ', $key); |
| 368 | $salutations[] = ['keyWords' => $keyWords, 'length' => count($keyWords), 'display' => $display]; |
| 369 | } |
| 370 | |
| 371 | $claimed = []; |
| 372 | $index = 0; |
| 373 | |
| 374 | while (true) { |
| 375 | $max = !empty($tokens) ? max(1, (int) floor(count($tokens) / 2)) : 0; |
| 376 | if (($index >= $max) || ($index >= count($tokens))) { |
| 377 | break; |
| 378 | } |
| 379 | |
| 380 | $matchedLength = null; |
| 381 | $matchedValue = null; |
| 382 | |
| 383 | foreach ($salutations as ['keyWords' => $keyWords, 'length' => $length, 'display' => $display]) { |
| 384 | if (($index + $length) > count($tokens)) { |
| 385 | continue; |
| 386 | } |
| 387 | $subset = array_slice($tokens, $index, $length); |
| 388 | $subsetKeys = array_map(fn($word) => strtolower(str_replace('.', '', $word)), $subset); |
| 389 | if ($subsetKeys === $keyWords) { |
| 390 | $matchedLength = $length; |
| 391 | $matchedValue = $display; |
| 392 | break; |
| 393 | } |
| 394 | } |
| 395 | |
| 396 | if ($matchedLength === null) { |
| 397 | $index++; |
| 398 | continue; |
| 399 | } |
| 400 | |
| 401 | $claimed[] = $matchedValue; |
| 402 | array_splice($tokens, $index, $matchedLength); |
| 403 | } |
| 404 | |
| 405 | if (!empty($claimed)) { |
| 406 | $fields['salutation'] = trim(($fields['salutation'] ?? '') . ' ' . implode(' ', $claimed)); |
| 407 | } |
| 408 | |
| 409 | return array_values($tokens); |
| 410 | } |
| 411 | |
| 412 | /** |
| 413 | * Extract suffix |
| 414 | * |
| 415 | * Scans from the end backward while trailing tokens keep matching either the generational |
| 416 | * suffix list or the professional credentials list, stopping before it would eat into the |
| 417 | * reserved leading tokens (or, in single-part mode, matches only when exactly one token |
| 418 | * remains). The two are kept in separate fields ("Von Fange III, PhD" -> suffix "III", |
| 419 | * credentials "PhD"), each preserving its own relative order even when the two are |
| 420 | * interleaved among the trailing tokens. |
| 421 | * |
| 422 | * @param array $tokens |
| 423 | * @param NameValues $nameValues |
| 424 | * @param array $fields |
| 425 | * @param int $reservedParts |
| 426 | * @param bool $matchSinglePart |
| 427 | * @param bool $reserveLastToken |
| 428 | * @return array |
| 429 | */ |
| 430 | protected function extractSuffix( |
| 431 | array $tokens, |
| 432 | NameValues $nameValues, |
| 433 | array &$fields, |
| 434 | int $reservedParts = 2, |
| 435 | bool $matchSinglePart = false, |
| 436 | bool $reserveLastToken = false |
| 437 | ): array |
| 438 | { |
| 439 | $suffixes = $nameValues->getSuffixes(); |
| 440 | $credentials = $nameValues->getCredentials(); |
| 441 | |
| 442 | if ($matchSinglePart && (count($tokens) === 1)) { |
| 443 | $key = strtolower(str_replace('.', '', $tokens[0])); |
| 444 | if (isset($suffixes[$key])) { |
| 445 | $fields['suffix'] = trim(($fields['suffix'] ?? '') . ' ' . $suffixes[$key]); |
| 446 | return []; |
| 447 | } |
| 448 | if (isset($credentials[$key])) { |
| 449 | $fields['credentials'] = trim(($fields['credentials'] ?? '') . ' ' . $credentials[$key]); |
| 450 | return []; |
| 451 | } |
| 452 | return $tokens; |
| 453 | } |
| 454 | |
| 455 | $claimed = []; |
| 456 | $stop = $reserveLastToken ? 1 : $reservedParts; |
| 457 | $index = count($tokens) - 1; |
| 458 | |
| 459 | while ($index >= $stop) { |
| 460 | $key = strtolower(str_replace('.', '', $tokens[$index])); |
| 461 | if (isset($suffixes[$key])) { |
| 462 | array_unshift($claimed, ['type' => 'suffix', 'value' => $suffixes[$key]]); |
| 463 | } else if (isset($credentials[$key])) { |
| 464 | array_unshift($claimed, ['type' => 'credentials', 'value' => $credentials[$key]]); |
| 465 | } else { |
| 466 | break; |
| 467 | } |
| 468 | $index--; |
| 469 | } |
| 470 | |
| 471 | if (!empty($claimed)) { |
| 472 | $count = count($claimed); |
| 473 | array_splice($tokens, count($tokens) - $count, $count); |
| 474 | |
| 475 | $suffixParts = array_column(array_filter($claimed, fn($claim) => $claim['type'] === 'suffix'), 'value'); |
| 476 | if (!empty($suffixParts)) { |
| 477 | $fields['suffix'] = trim(($fields['suffix'] ?? '') . ' ' . implode(' ', $suffixParts)); |
| 478 | } |
| 479 | |
| 480 | $credentialParts = array_column(array_filter($claimed, fn($claim) => $claim['type'] === 'credentials'), 'value'); |
| 481 | if (!empty($credentialParts)) { |
| 482 | $fields['credentials'] = trim(($fields['credentials'] ?? '') . ' ' . implode(' ', $credentialParts)); |
| 483 | } |
| 484 | } |
| 485 | |
| 486 | return array_values($tokens); |
| 487 | } |
| 488 | |
| 489 | /** |
| 490 | * Extract initials |
| 491 | * |
| 492 | * A remaining single letter (optionally with a trailing period) is an initial. An |
| 493 | * all-caps 2-letter run (e.g. "JR") is split into two separate initials first - unless |
| 494 | * that run is also a recognized lastname prefix ("DE", "LA", "ST", ...), in which case |
| 495 | * it's left alone so extractLastname() can fold it as a prefix; without this guard, |
| 496 | * all-caps input like "JAMES DE LUCA" would have "DE" shredded into two fake initials |
| 497 | * before extractLastname() ever saw it. The very last remaining token is never treated |
| 498 | * as an initial unless $matchLastPart is true. |
| 499 | * |
| 500 | * @param array $tokens |
| 501 | * @param bool $matchLastPart |
| 502 | * @param NameValues $nameValues |
| 503 | * @param array $initialsQueue |
| 504 | * @return array |
| 505 | */ |
| 506 | protected function extractInitials(array $tokens, bool $matchLastPart, NameValues $nameValues, array &$initialsQueue): array |
| 507 | { |
| 508 | $prefixes = $nameValues->getLastnamePrefixes(); |
| 509 | |
| 510 | $last = count($tokens) - 1; |
| 511 | for ($i = 0; $i < count($tokens); $i++) { |
| 512 | if (!$matchLastPart && ($i === $last)) { |
| 513 | continue; |
| 514 | } |
| 515 | $stripped = str_replace('.', '', $tokens[$i]); |
| 516 | if ((strlen($stripped) === 2) && ($stripped === strtoupper($stripped)) && ctype_alpha($stripped) |
| 517 | && !isset($prefixes[strtolower($stripped)])) { |
| 518 | array_splice($tokens, $i, 1, [$stripped[0], $stripped[1]]); |
| 519 | $last = count($tokens) - 1; |
| 520 | $i++; |
| 521 | } |
| 522 | } |
| 523 | |
| 524 | $last = count($tokens) - 1; |
| 525 | $claimedIndexes = []; |
| 526 | |
| 527 | foreach ($tokens as $i => $token) { |
| 528 | if (!$matchLastPart && ($i === $last)) { |
| 529 | continue; |
| 530 | } |
| 531 | $stripped = str_replace('.', '', $token); |
| 532 | if (strlen($stripped) === 1) { |
| 533 | $claimedIndexes[] = $i; |
| 534 | } |
| 535 | } |
| 536 | |
| 537 | foreach ($claimedIndexes as $i) { |
| 538 | $initialsQueue[] = $tokens[$i]; |
| 539 | } |
| 540 | foreach (array_reverse($claimedIndexes) as $i) { |
| 541 | array_splice($tokens, $i, 1); |
| 542 | } |
| 543 | |
| 544 | return array_values($tokens); |
| 545 | } |
| 546 | |
| 547 | /** |
| 548 | * Extract lastname (with prefix folding) |
| 549 | * |
| 550 | * Scans remaining tokens from the end backward, claiming them as lastname. A claimed run |
| 551 | * immediately preceded by a recognized lastname-prefix word, with at least one unclaimed |
| 552 | * token still before it, folds the prefix into lastnamePrefix. Stops once it hits a word |
| 553 | * long enough to look like a complete lastname on its own with more still unclaimed |
| 554 | * before it - what keeps a middle name from being swallowed into the lastname. |
| 555 | * |
| 556 | * $originalCount is the token count BEFORE any earlier extraction step ran; it (not the |
| 557 | * current, shrunk token count) determines whether there was ever more than one word in |
| 558 | * this name to begin with, since a name reduced to a single remaining token by earlier |
| 559 | * steps (e.g. "J. B. Hunt" -> "Hunt" once both initials are claimed) should still have |
| 560 | * that token claimed as lastname. |
| 561 | * |
| 562 | * @param array $tokens |
| 563 | * @param NameValues $nameValues |
| 564 | * @param array $fields |
| 565 | * @param bool $singlePartOk |
| 566 | * @param ?int $originalCount |
| 567 | * @return array |
| 568 | */ |
| 569 | protected function extractLastname( |
| 570 | array $tokens, |
| 571 | NameValues $nameValues, |
| 572 | array &$fields, |
| 573 | bool $singlePartOk = false, |
| 574 | ?int $originalCount = null |
| 575 | ): array |
| 576 | { |
| 577 | $originalCount ??= count($tokens); |
| 578 | |
| 579 | if ((!$singlePartOk && ($originalCount < 2)) || empty($tokens)) { |
| 580 | return $tokens; |
| 581 | } |
| 582 | |
| 583 | $prefixes = $nameValues->getLastnamePrefixes(); |
| 584 | $lastnameWords = []; |
| 585 | $prefixWords = []; |
| 586 | $index = count($tokens) - 1; |
| 587 | $claimedAny = false; |
| 588 | $lastClaimedWord = null; |
| 589 | |
| 590 | while ($index >= 0) { |
| 591 | $word = $tokens[$index]; |
| 592 | $key = strtolower(str_replace('.', '', $word)); |
| 593 | |
| 594 | // The "must be at index > 0" guard exists to always leave at least one token |
| 595 | // unclaimed for a firstname - but in singlePartOk mode (comma-mode segment 1), |
| 596 | // firstname comes from segment 2, not this segment, so there's nothing to |
| 597 | // reserve: allow folding and continued lastname-claiming all the way to index 0. |
| 598 | if ($claimedAny && isset($prefixes[$key]) && (($index > 0) || $singlePartOk)) { |
| 599 | array_unshift($prefixWords, $prefixes[$key]); |
| 600 | array_splice($tokens, $index, 1); |
| 601 | $index--; |
| 602 | continue; |
| 603 | } |
| 604 | |
| 605 | if ($claimedAny) { |
| 606 | if (!$singlePartOk && ($index < 1)) { |
| 607 | break; |
| 608 | } |
| 609 | if (strlen($lastClaimedWord) >= 3) { |
| 610 | break; |
| 611 | } |
| 612 | } |
| 613 | |
| 614 | array_unshift($lastnameWords, $word); |
| 615 | array_splice($tokens, $index, 1); |
| 616 | $claimedAny = true; |
| 617 | $lastClaimedWord = $word; |
| 618 | $index--; |
| 619 | |
| 620 | if (!$singlePartOk && ($index < 0)) { |
| 621 | break; |
| 622 | } |
| 623 | } |
| 624 | |
| 625 | $fields['lastname'] = $this->normalizeWords($lastnameWords); |
| 626 | if (!empty($prefixWords)) { |
| 627 | $fields['lastnamePrefix'] = implode(' ', $prefixWords); |
| 628 | } |
| 629 | |
| 630 | return array_values($tokens); |
| 631 | } |
| 632 | |
| 633 | /** |
| 634 | * Extract firstname |
| 635 | * |
| 636 | * If exactly one raw token remains, it's the firstname outright; otherwise the first |
| 637 | * remaining token becomes firstname. |
| 638 | * |
| 639 | * @param array $tokens |
| 640 | * @param array $fields |
| 641 | * @return array |
| 642 | */ |
| 643 | protected function extractFirstname(array $tokens, array &$fields): array |
| 644 | { |
| 645 | if (empty($tokens)) { |
| 646 | return $tokens; |
| 647 | } |
| 648 | |
| 649 | $fields['firstname'] = $this->normalizeCase($tokens[0]); |
| 650 | array_splice($tokens, 0, 1); |
| 651 | |
| 652 | return array_values($tokens); |
| 653 | } |
| 654 | |
| 655 | /** |
| 656 | * Extract middlename |
| 657 | * |
| 658 | * Whatever raw tokens remain after firstname extraction join as middlename. |
| 659 | * |
| 660 | * @param array $tokens |
| 661 | * @param array $fields |
| 662 | * @return array |
| 663 | */ |
| 664 | protected function extractMiddlename(array $tokens, array &$fields): array |
| 665 | { |
| 666 | if (empty($tokens)) { |
| 667 | return $tokens; |
| 668 | } |
| 669 | |
| 670 | $fields['middlename'] = trim(($fields['middlename'] ?? '') . ' ' . $this->normalizeWords($tokens)); |
| 671 | |
| 672 | return []; |
| 673 | } |
| 674 | |
| 675 | } |