1<?php 2 3/** 4 * webtrees: online genealogy 5 * Copyright (C) 2022 webtrees development team 6 * This program is free software: you can redistribute it and/or modify 7 * it under the terms of the GNU General Public License as published by 8 * the Free Software Foundation, either version 3 of the License, or 9 * (at your option) any later version. 10 * This program is distributed in the hope that it will be useful, 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 13 * GNU General Public License for more details. 14 * You should have received a copy of the GNU General Public License 15 * along with this program. If not, see <https://www.gnu.org/licenses/>. 16 */ 17 18declare(strict_types=1); 19 20namespace Fisharebest\Webtrees\Http\RequestHandlers; 21 22use Fisharebest\Webtrees\Elements\AbstractXrefElement; 23use Fisharebest\Webtrees\Elements\SubmitterText; 24use Fisharebest\Webtrees\Elements\UnknownElement; 25use Fisharebest\Webtrees\Elements\XrefFamily; 26use Fisharebest\Webtrees\Elements\XrefIndividual; 27use Fisharebest\Webtrees\Elements\XrefLocation; 28use Fisharebest\Webtrees\Elements\XrefMedia; 29use Fisharebest\Webtrees\Elements\XrefNote; 30use Fisharebest\Webtrees\Elements\XrefRepository; 31use Fisharebest\Webtrees\Elements\XrefSource; 32use Fisharebest\Webtrees\Elements\XrefSubmission; 33use Fisharebest\Webtrees\Elements\XrefSubmitter; 34use Fisharebest\Webtrees\Factories\ElementFactory; 35use Fisharebest\Webtrees\Family; 36use Fisharebest\Webtrees\Gedcom; 37use Fisharebest\Webtrees\Header; 38use Fisharebest\Webtrees\Http\ViewResponseTrait; 39use Fisharebest\Webtrees\I18N; 40use Fisharebest\Webtrees\Individual; 41use Fisharebest\Webtrees\Location; 42use Fisharebest\Webtrees\Media; 43use Fisharebest\Webtrees\Note; 44use Fisharebest\Webtrees\Registry; 45use Fisharebest\Webtrees\Repository; 46use Fisharebest\Webtrees\Services\TimeoutService; 47use Fisharebest\Webtrees\Source; 48use Fisharebest\Webtrees\Submission; 49use Fisharebest\Webtrees\Submitter; 50use Fisharebest\Webtrees\Tree; 51use Fisharebest\Webtrees\Validator; 52use Illuminate\Database\Capsule\Manager as DB; 53use Illuminate\Database\Query\Expression; 54use Psr\Http\Message\ResponseInterface; 55use Psr\Http\Message\ServerRequestInterface; 56use Psr\Http\Server\RequestHandlerInterface; 57 58use function array_slice; 59use function e; 60use function implode; 61use function preg_match; 62use function route; 63use function str_starts_with; 64use function strtoupper; 65use function substr_count; 66 67/** 68 * Check a tree for errors. 69 */ 70class CheckTree implements RequestHandlerInterface 71{ 72 use ViewResponseTrait; 73 74 private Gedcom $gedcom; 75 76 private TimeoutService $timeout_service; 77 78 /** 79 * @param Gedcom $gedcom 80 * @param TimeoutService $timeout_service 81 */ 82 public function __construct(Gedcom $gedcom, TimeoutService $timeout_service) 83 { 84 $this->gedcom = $gedcom; 85 $this->timeout_service = $timeout_service; 86 } 87 88 /** 89 * @param ServerRequestInterface $request 90 * 91 * @return ResponseInterface 92 */ 93 public function handle(ServerRequestInterface $request): ResponseInterface 94 { 95 $this->layout = 'layouts/administration'; 96 97 $tree = Validator::attributes($request)->tree(); 98 $skip_to = Validator::queryParams($request)->string('skip_to', ''); 99 100 // We need to work with raw GEDCOM data, as we are looking for errors 101 // which may prevent the GedcomRecord objects from working. 102 103 $q1 = DB::table('individuals') 104 ->where('i_file', '=', $tree->id()) 105 ->select(['i_id AS xref', 'i_gedcom AS gedcom', new Expression("'INDI' AS type")]); 106 $q2 = DB::table('families') 107 ->where('f_file', '=', $tree->id()) 108 ->select(['f_id AS xref', 'f_gedcom AS gedcom', new Expression("'FAM' AS type")]); 109 $q3 = DB::table('media') 110 ->where('m_file', '=', $tree->id()) 111 ->select(['m_id AS xref', 'm_gedcom AS gedcom', new Expression("'OBJE' AS type")]); 112 $q4 = DB::table('sources') 113 ->where('s_file', '=', $tree->id()) 114 ->select(['s_id AS xref', 's_gedcom AS gedcom', new Expression("'SOUR' AS type")]); 115 $q5 = DB::table('other') 116 ->where('o_file', '=', $tree->id()) 117 ->select(['o_id AS xref', 'o_gedcom AS gedcom', 'o_type']); 118 $q6 = DB::table('change') 119 ->where('gedcom_id', '=', $tree->id()) 120 ->where('status', '=', 'pending') 121 ->orderBy('change_id') 122 ->select(['xref', 'new_gedcom AS gedcom', new Expression("'' AS type")]); 123 124 $rows = $q1 125 ->unionAll($q2) 126 ->unionAll($q3) 127 ->unionAll($q4) 128 ->unionAll($q5) 129 ->unionAll($q6) 130 ->get() 131 ->map(static function (object $row): object { 132 // Extract type for pending record 133 if ($row->type === '' && preg_match('/^0 @[^@]*@ ([_A-Z0-9]+)/', $row->gedcom, $match) === 1) { 134 $row->type = $match[1]; 135 } 136 137 return $row; 138 }); 139 140 $records = []; 141 $xrefs = []; 142 143 foreach ($rows as $row) { 144 if ($row->gedcom !== '') { 145 // existing or updated record 146 $records[$row->xref] = $row; 147 } else { 148 // deleted record 149 unset($records[$row->xref]); 150 } 151 152 $xrefs[strtoupper($row->xref)] = $row->xref; 153 } 154 155 unset($rows); 156 157 $errors = []; 158 $warnings = []; 159 160 $element_factory = new ElementFactory(); 161 $this->gedcom->registerTags($element_factory, false); 162 163 foreach ($records as $record) { 164 // If we are nearly out of time, then stop processing here 165 if ($skip_to === $record->xref) { 166 $skip_to = ''; 167 } elseif ($skip_to !== '') { 168 continue; 169 } elseif ($this->timeout_service->isTimeNearlyUp()) { 170 $skip_to = $record->xref; 171 break; 172 } 173 174 $lines = explode("\n", $record->gedcom); 175 array_shift($lines); 176 177 $last_level = 0; 178 $hierarchy = [$record->type]; 179 180 foreach ($lines as $line_number => $line) { 181 if (preg_match('/^(\d+) (\w+) ?(.*)/', $line, $match) !== 1) { 182 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, I18N::translate('Invalid GEDCOM record')); 183 break; 184 } 185 186 $level = (int) $match[1]; 187 if ($level > $last_level + 1) { 188 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, I18N::translate('Invalid GEDCOM line number')); 189 break; 190 } 191 192 $tag = $match[2]; 193 $value = $match[3]; 194 $hierarchy[$level] = $tag; 195 $full_tag = implode(':', array_slice($hierarchy, 0, 1 + $level)); 196 $element = Registry::elementFactory()->make($full_tag); 197 $last_level = $level; 198 199 if ($tag === 'CONT') { 200 $element = new SubmitterText('CONT'); 201 } 202 203 if ($element instanceof UnknownElement) { 204 if (str_starts_with($tag, '_')) { 205 $message = I18N::translate('Custom GEDCOM tags are discouraged. Try to use only standard GEDCOM tags.'); 206 $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 207 } else { 208 $message = I18N::translate('Invalid GEDCOM tag.') . ' ' . $full_tag; 209 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 210 } 211 } elseif ($element instanceof AbstractXrefElement) { 212 if (preg_match('/@(' . Gedcom::REGEX_XREF . ')@/', $value, $match) === 1) { 213 $xref1 = $match[1]; 214 $xref2 = $xrefs[strtoupper($xref1)] ?? null; 215 $linked = $records[$xref2] ?? null; 216 217 if ($linked === null) { 218 $message = I18N::translate('%s does not exist.', e($xref1)); 219 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 220 } elseif ($element instanceof XrefFamily && $linked->type !== Family::RECORD_TYPE) { 221 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Family::RECORD_TYPE); 222 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 223 } elseif ($element instanceof XrefIndividual && $linked->type !== Individual::RECORD_TYPE) { 224 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Individual::RECORD_TYPE); 225 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 226 } elseif ($element instanceof XrefMedia && $linked->type !== Media::RECORD_TYPE) { 227 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Media::RECORD_TYPE); 228 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 229 } elseif ($element instanceof XrefNote && $linked->type !== Note::RECORD_TYPE) { 230 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Note::RECORD_TYPE); 231 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 232 } elseif ($element instanceof XrefSource && $linked->type !== Source::RECORD_TYPE) { 233 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Source::RECORD_TYPE); 234 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 235 } elseif ($element instanceof XrefRepository && $linked->type !== Repository::RECORD_TYPE) { 236 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Repository::RECORD_TYPE); 237 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 238 } elseif ($element instanceof XrefSubmitter && $linked->type !== Submitter::RECORD_TYPE) { 239 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Submitter::RECORD_TYPE); 240 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 241 } elseif ($element instanceof XrefSubmission && $linked->type !== Submission::RECORD_TYPE) { 242 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Submission::RECORD_TYPE); 243 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 244 } elseif ($element instanceof XrefLocation && $linked->type !== Location::RECORD_TYPE) { 245 $message = $this->linkErrorMessage($tree, $xref1, $linked->type, Location::RECORD_TYPE); 246 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 247 } elseif (($full_tag === 'FAM:HUSB' || $full_tag === 'FAM:WIFE') && !str_contains($linked->gedcom, "\n1 FAMS @" . $record->xref . '@')) { 248 $link1 = $this->recordLink($tree, $linked->xref); 249 $link2 = $this->recordLink($tree, $record->xref); 250 $message = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2); 251 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 252 } elseif ($full_tag === 'FAM:CHIL' && !str_contains($linked->gedcom, "\n1 FAMC @" . $record->xref . '@')) { 253 $link1 = $this->recordLink($tree, $linked->xref); 254 $link2 = $this->recordLink($tree, $record->xref); 255 $message = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2); 256 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 257 } elseif ($full_tag === 'INDI:FAMC' && !str_contains($linked->gedcom, "\n1 CHIL @" . $record->xref . '@')) { 258 $link1 = $this->recordLink($tree, $linked->xref); 259 $link2 = $this->recordLink($tree, $record->xref); 260 $message = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2); 261 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 262 } elseif ($full_tag === 'INDI:FAMS' && !str_contains($linked->gedcom, "\n1 HUSB @" . $record->xref . '@') && !str_contains($linked->gedcom, "\n1 WIFE @" . $record->xref . '@')) { 263 $link1 = $this->recordLink($tree, $linked->xref); 264 $link2 = $this->recordLink($tree, $record->xref); 265 $message = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2); 266 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 267 } elseif ($xref1 !== $xref2) { 268 $message = I18N::translate('%1$s does not exist. Did you mean %2$s?', e($xref1), e($xref2)); 269 $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 270 } 271 } elseif ($tag === 'SOUR') { 272 $message = I18N::translate('Inline-source records are discouraged.'); 273 $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 274 } else { 275 $message = I18N::translate('Invalid GEDCOM value'); 276 $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 277 } 278 } elseif ($element->canonical($value) !== $value) { 279 $expected = e($element->canonical($value)); 280 $actual = strtr(e($value), ["\t" => '→']); 281 $message = I18N::translate('“%1$s” should be “%2$s”.', $actual, $expected); 282 $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message); 283 } 284 } 285 286 if ($record->type === Family::RECORD_TYPE) { 287 if (substr_count($record->gedcom, "\n1 HUSB @") > 1) { 288 $message = I18N::translate('%s occurs too many times.', 'FAM:HUSB'); 289 $errors[] = $this->recordError($tree, $record->type, $record->xref, $message); 290 } 291 if (substr_count($record->gedcom, "\n1 WIFE @") > 1) { 292 $message = I18N::translate('%s occurs too many times.', 'FAM:WIFE'); 293 $errors[] = $this->recordError($tree, $record->type, $record->xref, $message); 294 } 295 } 296 } 297 298 $title = I18N::translate('Check for errors') . ' — ' . e($tree->title()); 299 300 if ($skip_to === '') { 301 $more_url = ''; 302 } else { 303 $more_url = route(self::class, ['tree' => $tree->name(), 'skip_to' => $skip_to]); 304 } 305 306 return $this->viewResponse('admin/trees-check', [ 307 'errors' => $errors, 308 'more_url' => $more_url, 309 'title' => $title, 310 'tree' => $tree, 311 'warnings' => $warnings, 312 ]); 313 } 314 315 /** 316 * @param string $type 317 * 318 * @return string 319 */ 320 private function recordType(string $type): string 321 { 322 $types = [ 323 Family::RECORD_TYPE => I18N::translate('Family'), 324 Header::RECORD_TYPE => I18N::translate('Header'), 325 Individual::RECORD_TYPE => I18N::translate('Individual'), 326 Location::RECORD_TYPE => I18N::translate('Location'), 327 Media::RECORD_TYPE => I18N::translate('Media object'), 328 Note::RECORD_TYPE => I18N::translate('Note'), 329 Repository::RECORD_TYPE => I18N::translate('Repository'), 330 Source::RECORD_TYPE => I18N::translate('Source'), 331 Submission::RECORD_TYPE => I18N::translate('Submission'), 332 Submitter::RECORD_TYPE => I18N::translate('Submitter'), 333 ]; 334 335 return $types[$type] ?? e($type); 336 } 337 338 /** 339 * @param Tree $tree 340 * @param string $xref 341 * 342 * @return string 343 */ 344 private function recordLink(Tree $tree, string $xref): string 345 { 346 $url = route(GedcomRecordPage::class, ['xref' => $xref, 'tree' => $tree->name()]); 347 348 return '<a href="' . e($url) . '">' . e($xref) . '</a>'; 349 } 350 351 /** 352 * Format a link to a record. 353 * 354 * @param Tree $tree 355 * @param string $type 356 * @param string $xref 357 * @param int $line_number 358 * @param string $line 359 * @param string $message 360 * 361 * @return string 362 */ 363 private function lineError(Tree $tree, string $type, string $xref, int $line_number, string $line, string $message): string 364 { 365 return 366 I18N::translate('%1$s: %2$s', $this->recordType($type), $this->recordLink($tree, $xref)) . 367 ' — ' . 368 I18N::translate('%1$s: %2$s', I18N::translate('Line number'), I18N::number($line_number)) . 369 ' — ' . 370 '<code>' . e($line) . '</code>' . 371 '<br>' . $message; 372 } 373 374 /** 375 * Format a link to a record. 376 * 377 * @param Tree $tree 378 * @param string $type 379 * @param string $xref 380 * @param string $message 381 * 382 * @return string 383 */ 384 private function recordError(Tree $tree, string $type, string $xref, string $message): string 385 { 386 return I18N::translate('%1$s: %2$s', $this->recordType($type), $this->recordLink($tree, $xref)) . ' — ' . $message; 387 } 388 389 /** 390 * @param Tree $tree 391 * @param string $xref 392 * @param string $type1 393 * @param string $type2 394 * 395 * @return string 396 */ 397 private function linkErrorMessage(Tree $tree, string $xref, string $type1, string $type2): string 398 { 399 $link = $this->recordLink($tree, $xref); 400 $type1 = $this->recordType($type1); 401 $type2 = $this->recordType($type2); 402 403 return I18N::translate('%1$s is a %2$s but a %3$s is expected.', $link, $type1, $type2); 404 } 405} 406