Passed
Push — develop ( dc4166...c2d6d8 )
by Greg
12:30
created

CheckTree::__construct()   A

Complexity

Conditions 1
Paths 1

Size

Total Lines 4
Code Lines 2

Duplication

Lines 0
Ratio 0 %

Importance

Changes 0
Metric Value
cc 1
eloc 2
nc 1
nop 2
dl 0
loc 4
rs 10
c 0
b 0
f 0
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
18
declare(strict_types=1);
19
20
namespace Fisharebest\Webtrees\Http\RequestHandlers;
21
22
use Fisharebest\Webtrees\Elements\AbstractXrefElement;
23
use Fisharebest\Webtrees\Elements\MultimediaFormat;
24
use Fisharebest\Webtrees\Elements\SubmitterText;
25
use Fisharebest\Webtrees\Elements\UnknownElement;
26
use Fisharebest\Webtrees\Elements\XrefFamily;
27
use Fisharebest\Webtrees\Elements\XrefIndividual;
28
use Fisharebest\Webtrees\Elements\XrefLocation;
29
use Fisharebest\Webtrees\Elements\XrefMedia;
30
use Fisharebest\Webtrees\Elements\XrefNote;
31
use Fisharebest\Webtrees\Elements\XrefRepository;
32
use Fisharebest\Webtrees\Elements\XrefSource;
33
use Fisharebest\Webtrees\Elements\XrefSubmission;
34
use Fisharebest\Webtrees\Elements\XrefSubmitter;
35
use Fisharebest\Webtrees\Factories\ElementFactory;
36
use Fisharebest\Webtrees\Factories\ImageFactory;
37
use Fisharebest\Webtrees\Family;
38
use Fisharebest\Webtrees\Gedcom;
39
use Fisharebest\Webtrees\Header;
40
use Fisharebest\Webtrees\Http\ViewResponseTrait;
41
use Fisharebest\Webtrees\I18N;
42
use Fisharebest\Webtrees\Individual;
43
use Fisharebest\Webtrees\Location;
44
use Fisharebest\Webtrees\Media;
45
use Fisharebest\Webtrees\Mime;
46
use Fisharebest\Webtrees\Note;
47
use Fisharebest\Webtrees\Registry;
48
use Fisharebest\Webtrees\Repository;
49
use Fisharebest\Webtrees\Services\TimeoutService;
50
use Fisharebest\Webtrees\Source;
51
use Fisharebest\Webtrees\Submission;
52
use Fisharebest\Webtrees\Submitter;
53
use Fisharebest\Webtrees\Tree;
54
use Fisharebest\Webtrees\Validator;
55
use Illuminate\Database\Capsule\Manager as DB;
56
use Illuminate\Database\Query\Expression;
57
use Psr\Http\Message\ResponseInterface;
58
use Psr\Http\Message\ServerRequestInterface;
59
use Psr\Http\Server\RequestHandlerInterface;
60
61
use function array_key_exists;
62
use function array_slice;
63
use function e;
64
use function implode;
65
use function in_array;
66
use function preg_match;
67
use function route;
68
use function str_starts_with;
69
use function strtoupper;
70
use function substr_count;
71
72
/**
73
 * Check a tree for errors.
74
 */
75
class CheckTree implements RequestHandlerInterface
76
{
77
    use ViewResponseTrait;
78
79
    private Gedcom $gedcom;
80
81
    private TimeoutService $timeout_service;
82
83
    /**
84
     * @param Gedcom         $gedcom
85
     * @param TimeoutService $timeout_service
86
     */
87
    public function __construct(Gedcom $gedcom, TimeoutService $timeout_service)
88
    {
89
        $this->gedcom          = $gedcom;
90
        $this->timeout_service = $timeout_service;
91
    }
92
93
    /**
94
     * @param ServerRequestInterface $request
95
     *
96
     * @return ResponseInterface
97
     */
98
    public function handle(ServerRequestInterface $request): ResponseInterface
99
    {
100
        $this->layout = 'layouts/administration';
101
102
        $tree    = Validator::attributes($request)->tree();
103
        $skip_to = Validator::queryParams($request)->string('skip_to', '');
104
105
        // We need to work with raw GEDCOM data, as we are looking for errors
106
        // which may prevent the GedcomRecord objects from working.
107
108
        $q1 = DB::table('individuals')
109
            ->where('i_file', '=', $tree->id())
110
            ->select(['i_id AS xref', 'i_gedcom AS gedcom', new Expression("'INDI' AS type")]);
111
        $q2 = DB::table('families')
112
            ->where('f_file', '=', $tree->id())
113
            ->select(['f_id AS xref', 'f_gedcom AS gedcom', new Expression("'FAM' AS type")]);
114
        $q3 = DB::table('media')
115
            ->where('m_file', '=', $tree->id())
116
            ->select(['m_id AS xref', 'm_gedcom AS gedcom', new Expression("'OBJE' AS type")]);
117
        $q4 = DB::table('sources')
118
            ->where('s_file', '=', $tree->id())
119
            ->select(['s_id AS xref', 's_gedcom AS gedcom', new Expression("'SOUR' AS type")]);
120
        $q5 = DB::table('other')
121
            ->where('o_file', '=', $tree->id())
122
            ->select(['o_id AS xref', 'o_gedcom AS gedcom', 'o_type']);
123
        $q6 = DB::table('change')
124
            ->where('gedcom_id', '=', $tree->id())
125
            ->where('status', '=', 'pending')
126
            ->orderBy('change_id')
127
            ->select(['xref', 'new_gedcom AS gedcom', new Expression("'' AS type")]);
128
129
        $rows = $q1
130
            ->unionAll($q2)
131
            ->unionAll($q3)
132
            ->unionAll($q4)
133
            ->unionAll($q5)
134
            ->unionAll($q6)
135
            ->get()
136
            ->map(static function (object $row): object {
137
                // Extract type for pending record
138
                if ($row->type === '' && preg_match('/^0 @[^@]*@ ([_A-Z0-9]+)/', $row->gedcom, $match) === 1) {
139
                    $row->type = $match[1];
140
                }
141
142
                return $row;
143
            });
144
145
        $records = [];
146
        $xrefs   = [];
147
148
        foreach ($rows as $row) {
149
            if ($row->gedcom !== '') {
150
                // existing or updated record
151
                $records[$row->xref] = $row;
152
            } else {
153
                // deleted record
154
                unset($records[$row->xref]);
155
            }
156
157
            $xrefs[strtoupper($row->xref)] = $row->xref;
158
        }
159
160
        unset($rows);
161
162
        $errors   = [];
163
        $warnings = [];
164
        $infos    = [];
165
166
        $element_factory = new ElementFactory();
167
        $this->gedcom->registerTags($element_factory, false);
168
169
        foreach ($records as $record) {
170
            // If we are nearly out of time, then stop processing here
171
            if ($skip_to === $record->xref) {
172
                $skip_to = '';
173
            } elseif ($skip_to !== '') {
174
                continue;
175
            } elseif ($this->timeout_service->isTimeNearlyUp()) {
176
                $skip_to = $record->xref;
177
                break;
178
            }
179
180
            $lines = explode("\n", $record->gedcom);
181
            array_shift($lines);
182
183
            $last_level = 0;
184
            $hierarchy  = [$record->type];
185
186
            foreach ($lines as $line_number => $line) {
187
                if (preg_match('/^(\d+) (\w+) ?(.*)/', $line, $match) !== 1) {
188
                    $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, I18N::translate('Invalid GEDCOM record'));
189
                    break;
190
                }
191
192
                $level = (int) $match[1];
193
                if ($level > $last_level + 1) {
194
                    $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, I18N::translate('Invalid GEDCOM line number'));
195
                    break;
196
                }
197
198
                $tag               = $match[2];
199
                $value             = $match[3];
200
                $hierarchy[$level] = $tag;
201
                $full_tag          = implode(':', array_slice($hierarchy, 0, 1 + $level));
202
                $element           = Registry::elementFactory()->make($full_tag);
203
                $last_level        = $level;
204
205
                if ($tag === 'CONT') {
206
                    $element = new SubmitterText('CONT');
207
                }
208
209
                if ($element instanceof UnknownElement) {
210
                    if (str_starts_with($tag, '_')) {
211
                        $message    = I18N::translate('Custom GEDCOM tags are discouraged. Try to use only standard GEDCOM tags.');
212
                        $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
213
                    } else {
214
                        $message  = I18N::translate('Invalid GEDCOM tag.') . ' ' . $full_tag;
215
                        $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
216
                    }
217
                } elseif ($element instanceof AbstractXrefElement) {
218
                    if (preg_match('/@(' . Gedcom::REGEX_XREF . ')@/', $value, $match) === 1) {
219
                        $xref1  = $match[1];
220
                        $xref2  = $xrefs[strtoupper($xref1)] ?? null;
221
                        $linked = $records[$xref2] ?? null;
222
223
                        if ($linked === null) {
224
                            $message  = I18N::translate('%s does not exist.', e($xref1));
225
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
226
                        } elseif ($element instanceof XrefFamily && $linked->type !== Family::RECORD_TYPE) {
227
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Family::RECORD_TYPE);
228
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
229
                        } elseif ($element instanceof XrefIndividual && $linked->type !== Individual::RECORD_TYPE) {
230
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Individual::RECORD_TYPE);
231
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
232
                        } elseif ($element instanceof XrefMedia && $linked->type !== Media::RECORD_TYPE) {
233
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Media::RECORD_TYPE);
234
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
235
                        } elseif ($element instanceof XrefNote && $linked->type !== Note::RECORD_TYPE) {
236
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Note::RECORD_TYPE);
237
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
238
                        } elseif ($element instanceof XrefSource && $linked->type !== Source::RECORD_TYPE) {
239
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Source::RECORD_TYPE);
240
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
241
                        } elseif ($element instanceof XrefRepository && $linked->type !== Repository::RECORD_TYPE) {
242
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Repository::RECORD_TYPE);
243
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
244
                        } elseif ($element instanceof XrefSubmitter && $linked->type !== Submitter::RECORD_TYPE) {
245
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Submitter::RECORD_TYPE);
246
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
247
                        } elseif ($element instanceof XrefSubmission && $linked->type !== Submission::RECORD_TYPE) {
248
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Submission::RECORD_TYPE);
249
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
250
                        } elseif ($element instanceof XrefLocation && $linked->type !== Location::RECORD_TYPE) {
251
                            $message  = $this->linkErrorMessage($tree, $xref1, $linked->type, Location::RECORD_TYPE);
252
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
253
                        } elseif (($full_tag === 'FAM:HUSB' || $full_tag === 'FAM:WIFE') && !str_contains($linked->gedcom, "\n1 FAMS @" . $record->xref . '@')) {
254
                            $link1    = $this->recordLink($tree, $linked->xref);
255
                            $link2    = $this->recordLink($tree, $record->xref);
256
                            $message  = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2);
257
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
258
                        } elseif ($full_tag === 'FAM:CHIL' && !str_contains($linked->gedcom, "\n1 FAMC @" . $record->xref . '@')) {
259
                            $link1    = $this->recordLink($tree, $linked->xref);
260
                            $link2    = $this->recordLink($tree, $record->xref);
261
                            $message  = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2);
262
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
263
                        } elseif ($full_tag === 'INDI:FAMC' && !str_contains($linked->gedcom, "\n1 CHIL @" . $record->xref . '@')) {
264
                            $link1    = $this->recordLink($tree, $linked->xref);
265
                            $link2    = $this->recordLink($tree, $record->xref);
266
                            $message  = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2);
267
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
268
                        } elseif ($full_tag === 'INDI:FAMS' && !str_contains($linked->gedcom, "\n1 HUSB @" . $record->xref . '@') && !str_contains($linked->gedcom, "\n1 WIFE @" . $record->xref . '@')) {
269
                            $link1    = $this->recordLink($tree, $linked->xref);
270
                            $link2    = $this->recordLink($tree, $record->xref);
271
                            $message  = I18N::translate('%1$s does not have a link back to %2$s.', $link1, $link2);
272
                            $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
273
                        } elseif ($xref1 !== $xref2) {
274
                            $message    = I18N::translate('%1$s does not exist. Did you mean %2$s?', e($xref1), e($xref2));
275
                            $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
276
                        }
277
                    } elseif ($tag === 'SOUR') {
278
                        $message    = I18N::translate('Inline-source records are discouraged.');
279
                        $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
280
                    } else {
281
                        $message  = I18N::translate('Invalid GEDCOM value');
282
                        $errors[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
283
                    }
284
                } elseif ($element->canonical($value) !== $value) {
285
                    $expected = e($element->canonical($value));
286
                    $actual   = strtr(e($value), ["\t" => '&rarr;']);
287
                    $message  = I18N::translate('“%1$s” should be “%2$s”.', $actual, $expected);
288
                    $infos[]  = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
289
                } elseif ($element instanceof MultimediaFormat) {
290
                    $mime = Mime::TYPES[$value] ?? Mime::DEFAULT_TYPE;
291
292
                    if ($mime === Mime::DEFAULT_TYPE) {
293
                        $message    = I18N::translate('webtrees does not recognise this file format.');
294
                        $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
295
                    } elseif (str_starts_with($mime, 'image/') && !array_key_exists($mime, ImageFactory::SUPPORTED_FORMATS)) {
296
                        $message    = I18N::translate('webtrees cannot create thumbnails for this file format.');
297
                        $warnings[] = $this->lineError($tree, $record->type, $record->xref, $line_number, $line, $message);
298
                    }
299
                }
300
            }
301
302
            if ($record->type === Family::RECORD_TYPE) {
303
                if (substr_count($record->gedcom, "\n1 HUSB @") > 1) {
304
                    $message  = I18N::translate('%s occurs too many times.', 'FAM:HUSB');
305
                    $errors[] = $this->recordError($tree, $record->type, $record->xref, $message);
306
                }
307
                if (substr_count($record->gedcom, "\n1 WIFE @") > 1) {
308
                    $message  = I18N::translate('%s occurs too many times.', 'FAM:WIFE');
309
                    $errors[] = $this->recordError($tree, $record->type, $record->xref, $message);
310
                }
311
            }
312
        }
313
314
        $title = I18N::translate('Check for errors') . ' — ' . e($tree->title());
315
316
        if ($skip_to === '') {
317
            $more_url = '';
318
        } else {
319
            $more_url = route(self::class, ['tree' => $tree->name(), 'skip_to' => $skip_to]);
320
        }
321
322
        return $this->viewResponse('admin/trees-check', [
323
            'errors'   => $errors,
324
            'infos'    => $infos,
325
            'more_url' => $more_url,
326
            'title'    => $title,
327
            'tree'     => $tree,
328
            'warnings' => $warnings,
329
        ]);
330
    }
331
332
    /**
333
     * @param string $type
334
     *
335
     * @return string
336
     */
337
    private function recordType(string $type): string
338
    {
339
        $types = [
340
            Family::RECORD_TYPE     => I18N::translate('Family'),
341
            Header::RECORD_TYPE     => I18N::translate('Header'),
342
            Individual::RECORD_TYPE => I18N::translate('Individual'),
343
            Location::RECORD_TYPE   => I18N::translate('Location'),
344
            Media::RECORD_TYPE      => I18N::translate('Media object'),
345
            Note::RECORD_TYPE       => I18N::translate('Note'),
346
            Repository::RECORD_TYPE => I18N::translate('Repository'),
347
            Source::RECORD_TYPE     => I18N::translate('Source'),
348
            Submission::RECORD_TYPE => I18N::translate('Submission'),
349
            Submitter::RECORD_TYPE  => I18N::translate('Submitter'),
350
        ];
351
352
        return $types[$type] ?? e($type);
353
    }
354
355
    /**
356
     * @param Tree   $tree
357
     * @param string $xref
358
     *
359
     * @return string
360
     */
361
    private function recordLink(Tree $tree, string $xref): string
362
    {
363
        $url = route(GedcomRecordPage::class, ['xref' => $xref, 'tree' => $tree->name()]);
364
365
        return '<a href="' . e($url) . '">' . e($xref) . '</a>';
366
    }
367
368
    /**
369
     * Format a link to a record.
370
     *
371
     * @param Tree   $tree
372
     * @param string $type
373
     * @param string $xref
374
     * @param int    $line_number
375
     * @param string $line
376
     * @param string $message
377
     *
378
     * @return string
379
     */
380
    private function lineError(Tree $tree, string $type, string $xref, int $line_number, string $line, string $message): string
381
    {
382
        return
383
            I18N::translate('%1$s: %2$s', $this->recordType($type), $this->recordLink($tree, $xref)) .
384
            ' — ' .
385
            I18N::translate('%1$s: %2$s', I18N::translate('Line number'), I18N::number($line_number)) .
386
            ' — ' .
387
            '<code>' . e($line) . '</code>' .
388
            '<br>' . $message;
389
    }
390
391
    /**
392
     * Format a link to a record.
393
     *
394
     * @param Tree   $tree
395
     * @param string $type
396
     * @param string $xref
397
     * @param string $message
398
     *
399
     * @return string
400
     */
401
    private function recordError(Tree $tree, string $type, string $xref, string $message): string
402
    {
403
        return I18N::translate('%1$s: %2$s', $this->recordType($type), $this->recordLink($tree, $xref)) . ' — ' . $message;
404
    }
405
406
    /**
407
     * @param Tree   $tree
408
     * @param string $xref
409
     * @param string $type1
410
     * @param string $type2
411
     *
412
     * @return string
413
     */
414
    private function linkErrorMessage(Tree $tree, string $xref, string $type1, string $type2): string
415
    {
416
        $link  = $this->recordLink($tree, $xref);
417
        $type1 = $this->recordType($type1);
418
        $type2 = $this->recordType($type2);
419
420
        return I18N::translate('%1$s is a %2$s but a %3$s is expected.', $link, $type1, $type2);
421
    }
422
}
423