xref: /webtrees/app/Http/RequestHandlers/CheckTree.php (revision 090a06287954f43677f06ea778b3f67c029de8fe) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2021 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\Gedcom;
23use Fisharebest\Webtrees\GedcomTag;
24use Fisharebest\Webtrees\Header;
25use Fisharebest\Webtrees\Http\ViewResponseTrait;
26use Fisharebest\Webtrees\I18N;
27use Fisharebest\Webtrees\Tree;
28use Illuminate\Database\Capsule\Manager as DB;
29use Illuminate\Database\Query\Expression;
30use Psr\Http\Message\ResponseInterface;
31use Psr\Http\Message\ServerRequestInterface;
32use Psr\Http\Server\RequestHandlerInterface;
33use stdClass;
34
35use function array_key_exists;
36use function assert;
37use function e;
38use function in_array;
39use function preg_match;
40use function preg_match_all;
41use function route;
42use function strtoupper;
43
44use const PREG_SET_ORDER;
45
46/**
47 * Check a tree for errors.
48 */
49class CheckTree implements RequestHandlerInterface
50{
51    use ViewResponseTrait;
52
53    /**
54     * @param ServerRequestInterface $request
55     *
56     * @return ResponseInterface
57     */
58    public function handle(ServerRequestInterface $request): ResponseInterface
59    {
60        $this->layout = 'layouts/administration';
61
62        $tree = $request->getAttribute('tree');
63        assert($tree instanceof Tree);
64
65        // We need to work with raw GEDCOM data, as we are looking for errors
66        // which may prevent the GedcomRecord objects from working.
67
68        $q1 = DB::table('individuals')
69            ->where('i_file', '=', $tree->id())
70            ->select(['i_id AS xref', 'i_gedcom AS gedcom', new Expression("'INDI' AS type")]);
71        $q2 = DB::table('families')
72            ->where('f_file', '=', $tree->id())
73            ->select(['f_id AS xref', 'f_gedcom AS gedcom', new Expression("'FAM' AS type")]);
74        $q3 = DB::table('media')
75            ->where('m_file', '=', $tree->id())
76            ->select(['m_id AS xref', 'm_gedcom AS gedcom', new Expression("'OBJE' AS type")]);
77        $q4 = DB::table('sources')
78            ->where('s_file', '=', $tree->id())
79            ->select(['s_id AS xref', 's_gedcom AS gedcom', new Expression("'SOUR' AS type")]);
80        $q5 = DB::table('other')
81            ->where('o_file', '=', $tree->id())
82            ->whereNotIn('o_type', [Header::RECORD_TYPE, 'TRLR'])
83            ->select(['o_id AS xref', 'o_gedcom AS gedcom', 'o_type']);
84        $q6 = DB::table('change')
85            ->where('gedcom_id', '=', $tree->id())
86            ->where('status', '=', 'pending')
87            ->orderBy('change_id')
88            ->select(['xref', 'new_gedcom AS gedcom', new Expression("'' AS type")]);
89
90        $rows = $q1
91            ->unionAll($q2)
92            ->unionAll($q3)
93            ->unionAll($q4)
94            ->unionAll($q5)
95            ->unionAll($q6)
96            ->get()
97            ->map(static function (stdClass $row): stdClass {
98                // Extract type for pending record
99                if ($row->type === '' && preg_match('/^0 @[^@]*@ ([_A-Z0-9]+)/', $row->gedcom, $match)) {
100                    $row->type = $match[1];
101                }
102
103                return $row;
104            });
105
106        $records = [];
107
108        foreach ($rows as $row) {
109            if ($row->gedcom !== '') {
110                // existing or updated record
111                $records[$row->xref] = $row;
112            } else {
113                // deleted record
114                unset($records[$row->xref]);
115            }
116        }
117
118        // LOOK FOR BROKEN LINKS
119        $XREF_LINKS = [
120            'NOTE'          => 'NOTE',
121            'SOUR'          => 'SOUR',
122            'REPO'          => 'REPO',
123            'OBJE'          => 'OBJE',
124            'SUBM'          => 'SUBM',
125            'FAMC'          => 'FAM',
126            'FAMS'          => 'FAM',
127            //'ADOP'=>'FAM', // Need to handle this case specially. We may have both ADOP and FAMC links to the same FAM, but only store one.
128            'HUSB'          => 'INDI',
129            'WIFE'          => 'INDI',
130            'CHIL'          => 'INDI',
131            'ASSO'          => 'INDI',
132            '_ASSO'         => 'INDI',
133            // A webtrees extension
134            'ALIA'          => 'INDI',
135            'AUTH'          => 'INDI',
136            // A webtrees extension
137            'ANCI'          => 'SUBM',
138            'DESI'          => 'SUBM',
139            '_WT_OBJE_SORT' => 'OBJE',
140            '_LOC'          => '_LOC',
141        ];
142
143        $RECORD_LINKS = [
144            'INDI' => [
145                'NOTE',
146                'OBJE',
147                'SOUR',
148                'SUBM',
149                'ASSO',
150                '_ASSO',
151                'FAMC',
152                'FAMS',
153                'ALIA',
154                '_WT_OBJE_SORT',
155                '_LOC',
156            ],
157            'FAM'  => [
158                'NOTE',
159                'OBJE',
160                'SOUR',
161                'SUBM',
162                'ASSO',
163                '_ASSO',
164                'HUSB',
165                'WIFE',
166                'CHIL',
167                '_LOC',
168            ],
169            'SOUR' => [
170                'NOTE',
171                'OBJE',
172                'REPO',
173                'AUTH',
174            ],
175            'REPO' => ['NOTE'],
176            'OBJE' => ['NOTE'],
177            // The spec also allows SOUR, but we treat this as a warning
178            'NOTE' => [],
179            // The spec also allows SOUR, but we treat this as a warning
180            'SUBM' => [
181                'NOTE',
182                'OBJE',
183            ],
184            'SUBN' => ['SUBM'],
185            '_LOC' => [
186                'SOUR',
187                'OBJE',
188                '_LOC',
189                'NOTE',
190            ],
191        ];
192
193        $errors   = [];
194        $warnings = [];
195
196        // Generate lists of all links
197        $all_links   = [];
198        $upper_links = [];
199        foreach ($records as $record) {
200            $all_links[$record->xref]               = [];
201            $upper_links[strtoupper($record->xref)] = $record->xref;
202            preg_match_all('/\n\d (' . Gedcom::REGEX_TAG . ') @([^#@\n][^\n@]*)@/', $record->gedcom, $matches, PREG_SET_ORDER);
203            foreach ($matches as $match) {
204                $all_links[$record->xref][$match[2]] = $match[1];
205            }
206        }
207
208        foreach ($all_links as $xref1 => $links) {
209            // PHP converts array keys to integers.
210            $xref1 = (string) $xref1;
211
212            $type1 = $records[$xref1]->type;
213            foreach ($links as $xref2 => $type2) {
214                // PHP converts array keys to integers.
215                $xref2 = (string) $xref2;
216
217                $type3 = isset($records[$xref2]) ? $records[$xref2]->type : '';
218                if (!array_key_exists($xref2, $all_links)) {
219                    if (array_key_exists(strtoupper($xref2), $upper_links)) {
220                        $warnings[] =
221                            $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' .
222                            /* I18N: placeholders are GEDCOM XREFs, such as R123 */
223                            I18N::translate('%1$s does not exist. Did you mean %2$s?', $this->checkLink($tree, $xref2), $this->checkLink($tree, $upper_links[strtoupper($xref2)]));
224                    } else {
225                        /* I18N: placeholders are GEDCOM XREFs, such as R123 */
226                        $errors[] = $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' . I18N::translate('%1$s does not exist.', $this->checkLink($tree, $xref2));
227                    }
228                } elseif ($type2 === 'SOUR' && $type1 === 'NOTE') {
229                    // Notes are intended to add explanations and comments to other records. They should not have their own sources.
230                } elseif ($type2 === 'SOUR' && $type1 === 'OBJE') {
231                    // Media objects are intended to illustrate other records, facts, and source/citations. They should not have their own sources.
232                } elseif ($type2 === 'OBJE' && $type1 === 'REPO') {
233                    $warnings[] =
234                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) .
235                        ' ' .
236                        I18N::translate('This type of link is not allowed here.');
237                } elseif (!array_key_exists($type1, $RECORD_LINKS) || !in_array($type2, $RECORD_LINKS[$type1], true) || !array_key_exists($type2, $XREF_LINKS)) {
238                    $errors[] =
239                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) .
240                        ' ' .
241                        I18N::translate('This type of link is not allowed here.');
242                } elseif ($XREF_LINKS[$type2] !== $type3) {
243                    // Target XREF does exist - but is invalid
244                    $errors[] =
245                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' .
246                        /* I18N: %1$s is an internal ID number such as R123. %2$s and %3$s are record types, such as INDI or SOUR */
247                        I18N::translate('%1$s is a %2$s but a %3$s is expected.', $this->checkLink($tree, $xref2), $this->formatType($type3), $this->formatType($type2));
248                } elseif (
249                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'FAMC', ['CHIL']) ||
250                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'FAMS', ['HUSB', 'WIFE']) ||
251                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'CHIL', ['FAMC']) ||
252                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'HUSB', ['FAMS']) ||
253                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'WIFE', ['FAMS'])
254                ) {
255                    /* I18N: %1$s and %2$s are internal ID numbers such as R123 */
256                    $errors[] = $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' . I18N::translate('%1$s does not have a link back to %2$s.', $this->checkLink($tree, $xref2), $this->checkLink($tree, $xref1));
257                }
258            }
259        }
260
261        $title = I18N::translate('Check for errors') . ' — ' . e($tree->title());
262
263        return $this->viewResponse('admin/trees-check', [
264            'errors'   => $errors,
265            'title'    => $title,
266            'tree'     => $tree,
267            'warnings' => $warnings,
268        ]);
269    }
270
271    /**
272     * @param string     $type
273     * @param string[][] $links
274     * @param string     $xref1
275     * @param string     $xref2
276     * @param string     $link
277     * @param string[]   $reciprocal
278     *
279     * @return bool
280     */
281    private function checkReverseLink(string $type, array $links, string $xref1, string $xref2, string $link, array $reciprocal): bool
282    {
283        return $type === $link && (!array_key_exists($xref1, $links[$xref2]) || !in_array($links[$xref2][$xref1], $reciprocal, true));
284    }
285
286    /**
287     * Create a message linking one record to another.
288     *
289     * @param Tree   $tree
290     * @param string $type1
291     * @param string $xref1
292     * @param string $type2
293     * @param string $xref2
294     *
295     * @return string
296     */
297    private function checkLinkMessage(Tree $tree, string $type1, string $xref1, string $type2, string $xref2): string
298    {
299        /* I18N: The placeholders are GEDCOM XREFs and tags. e.g. “INDI I123 contains a FAMC link to F234.” */
300        return I18N::translate(
301            '%1$s %2$s has a %3$s link to %4$s.',
302            $this->formatType($type1),
303            $this->checkLink($tree, $xref1),
304            $this->formatType($type2),
305            $this->checkLink($tree, $xref2)
306        );
307    }
308
309    /**
310     * Format a link to a record.
311     *
312     * @param Tree   $tree
313     * @param string $xref
314     *
315     * @return string
316     */
317    private function checkLink(Tree $tree, string $xref): string
318    {
319        return '<b><a href="' . e(route(GedcomRecordPage::class, [
320                'xref' => $xref,
321                'tree' => $tree->name(),
322            ])) . '">' . $xref . '</a></b>';
323    }
324
325    /**
326     * Format a record type.
327     *
328     * @param string $type
329     *
330     * @return string
331     */
332    private function formatType(string $type): string
333    {
334        return '<b title="' . GedcomTag::getLabel($type) . '">' . $type . '</b>';
335    }
336}
337