xref: /webtrees/app/Http/RequestHandlers/CheckTree.php (revision 40106a7319c9d89830bdc5f3fafd09abcca8cc6f) 
1<?php
2
3/**
4 * webtrees: online genealogy
5 * Copyright (C) 2020 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 <http://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\Services\AdminService;
28use Fisharebest\Webtrees\Services\TreeService;
29use Fisharebest\Webtrees\Tree;
30use Illuminate\Database\Capsule\Manager as DB;
31use Illuminate\Database\Query\Expression;
32use Psr\Http\Message\ResponseInterface;
33use Psr\Http\Message\ServerRequestInterface;
34use Psr\Http\Server\RequestHandlerInterface;
35use stdClass;
36
37use function array_key_exists;
38use function assert;
39use function e;
40use function in_array;
41use function preg_match;
42use function preg_match_all;
43use function route;
44use function strtoupper;
45
46use const PREG_SET_ORDER;
47
48/**
49 * Check a tree for errors.
50 */
51class CheckTree implements RequestHandlerInterface
52{
53    use ViewResponseTrait;
54
55    /**
56     * @param ServerRequestInterface $request
57     *
58     * @return ResponseInterface
59     */
60    public function handle(ServerRequestInterface $request): ResponseInterface
61    {
62        $this->layout = 'layouts/administration';
63
64        $tree = $request->getAttribute('tree');
65        assert($tree instanceof Tree);
66
67        // We need to work with raw GEDCOM data, as we are looking for errors
68        // which may prevent the GedcomRecord objects from working.
69
70        $q1 = DB::table('individuals')
71            ->where('i_file', '=', $tree->id())
72            ->select(['i_id AS xref', 'i_gedcom AS gedcom', new Expression("'INDI' AS type")]);
73        $q2 = DB::table('families')
74            ->where('f_file', '=', $tree->id())
75            ->select(['f_id AS xref', 'f_gedcom AS gedcom', new Expression("'FAM' AS type")]);
76        $q3 = DB::table('media')
77            ->where('m_file', '=', $tree->id())
78            ->select(['m_id AS xref', 'm_gedcom AS gedcom', new Expression("'OBJE' AS type")]);
79        $q4 = DB::table('sources')
80            ->where('s_file', '=', $tree->id())
81            ->select(['s_id AS xref', 's_gedcom AS gedcom', new Expression("'SOUR' AS type")]);
82        $q5 = DB::table('other')
83            ->where('o_file', '=', $tree->id())
84            ->whereNotIn('o_type', [Header::RECORD_TYPE, 'TRLR'])
85            ->select(['o_id AS xref', 'o_gedcom AS gedcom', 'o_type']);
86        $q6 = DB::table('change')
87            ->where('gedcom_id', '=', $tree->id())
88            ->where('status', '=', 'pending')
89            ->orderBy('change_id')
90            ->select(['xref', 'new_gedcom AS gedcom', new Expression("'' AS type")]);
91
92        $rows = $q1
93            ->unionAll($q2)
94            ->unionAll($q3)
95            ->unionAll($q4)
96            ->unionAll($q5)
97            ->unionAll($q6)
98            ->get()
99            ->map(static function (stdClass $row): stdClass {
100                // Extract type for pending record
101                if ($row->type === '' && preg_match('/^0 @[^@]*@ ([_A-Z0-9]+)/', $row->gedcom, $match)) {
102                    $row->type = $match[1];
103                }
104
105                return $row;
106            });
107
108        $records = [];
109
110        foreach ($rows as $row) {
111            if ($row->gedcom !== '') {
112                // existing or updated record
113                $records[$row->xref] = $row;
114            } else {
115                // deleted record
116                unset($records[$row->xref]);
117            }
118        }
119
120        // LOOK FOR BROKEN LINKS
121        $XREF_LINKS = [
122            'NOTE'          => 'NOTE',
123            'SOUR'          => 'SOUR',
124            'REPO'          => 'REPO',
125            'OBJE'          => 'OBJE',
126            'SUBM'          => 'SUBM',
127            'FAMC'          => 'FAM',
128            'FAMS'          => 'FAM',
129            //'ADOP'=>'FAM', // Need to handle this case specially. We may have both ADOP and FAMC links to the same FAM, but only store one.
130            'HUSB'          => 'INDI',
131            'WIFE'          => 'INDI',
132            'CHIL'          => 'INDI',
133            'ASSO'          => 'INDI',
134            '_ASSO'         => 'INDI',
135            // A webtrees extension
136            'ALIA'          => 'INDI',
137            'AUTH'          => 'INDI',
138            // A webtrees extension
139            'ANCI'          => 'SUBM',
140            'DESI'          => 'SUBM',
141            '_WT_OBJE_SORT' => 'OBJE',
142            '_LOC'          => '_LOC',
143        ];
144
145        $RECORD_LINKS = [
146            'INDI' => [
147                'NOTE',
148                'OBJE',
149                'SOUR',
150                'SUBM',
151                'ASSO',
152                '_ASSO',
153                'FAMC',
154                'FAMS',
155                'ALIA',
156                '_WT_OBJE_SORT',
157                '_LOC',
158            ],
159            'FAM'  => [
160                'NOTE',
161                'OBJE',
162                'SOUR',
163                'SUBM',
164                'ASSO',
165                '_ASSO',
166                'HUSB',
167                'WIFE',
168                'CHIL',
169                '_LOC',
170            ],
171            'SOUR' => [
172                'NOTE',
173                'OBJE',
174                'REPO',
175                'AUTH',
176            ],
177            'REPO' => ['NOTE'],
178            'OBJE' => ['NOTE'],
179            // The spec also allows SOUR, but we treat this as a warning
180            'NOTE' => [],
181            // The spec also allows SOUR, but we treat this as a warning
182            'SUBM' => [
183                'NOTE',
184                'OBJE',
185            ],
186            'SUBN' => ['SUBM'],
187            '_LOC' => [
188                'SOUR',
189                'OBJE',
190                '_LOC',
191                'NOTE',
192            ],
193        ];
194
195        $errors   = [];
196        $warnings = [];
197
198        // Generate lists of all links
199        $all_links   = [];
200        $upper_links = [];
201        foreach ($records as $record) {
202            $all_links[$record->xref]               = [];
203            $upper_links[strtoupper($record->xref)] = $record->xref;
204            preg_match_all('/\n\d (' . Gedcom::REGEX_TAG . ') @([^#@\n][^\n@]*)@/', $record->gedcom, $matches, PREG_SET_ORDER);
205            foreach ($matches as $match) {
206                $all_links[$record->xref][$match[2]] = $match[1];
207            }
208        }
209
210        foreach ($all_links as $xref1 => $links) {
211            // PHP converts array keys to integers.
212            $xref1 = (string) $xref1;
213
214            $type1 = $records[$xref1]->type;
215            foreach ($links as $xref2 => $type2) {
216                // PHP converts array keys to integers.
217                $xref2 = (string) $xref2;
218
219                $type3 = isset($records[$xref2]) ? $records[$xref2]->type : '';
220                if (!array_key_exists($xref2, $all_links)) {
221                    if (array_key_exists(strtoupper($xref2), $upper_links)) {
222                        $warnings[] =
223                            $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' .
224                            /* I18N: placeholders are GEDCOM XREFs, such as R123 */
225                            I18N::translate('%1$s does not exist. Did you mean %2$s?', $this->checkLink($tree, $xref2), $this->checkLink($tree, $upper_links[strtoupper($xref2)]));
226                    } else {
227                        /* I18N: placeholders are GEDCOM XREFs, such as R123 */
228                        $errors[] = $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' . I18N::translate('%1$s does not exist.', $this->checkLink($tree, $xref2));
229                    }
230                } elseif ($type2 === 'SOUR' && $type1 === 'NOTE') {
231                    // Notes are intended to add explanations and comments to other records. They should not have their own sources.
232                } elseif ($type2 === 'SOUR' && $type1 === 'OBJE') {
233                    // Media objects are intended to illustrate other records, facts, and source/citations. They should not have their own sources.
234                } elseif ($type2 === 'OBJE' && $type1 === 'REPO') {
235                    $warnings[] =
236                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) .
237                        ' ' .
238                        I18N::translate('This type of link is not allowed here.');
239                } elseif (!array_key_exists($type1, $RECORD_LINKS) || !in_array($type2, $RECORD_LINKS[$type1], true) || !array_key_exists($type2, $XREF_LINKS)) {
240                    $errors[] =
241                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) .
242                        ' ' .
243                        I18N::translate('This type of link is not allowed here.');
244                } elseif ($XREF_LINKS[$type2] !== $type3) {
245                    // Target XREF does exist - but is invalid
246                    $errors[] =
247                        $this->checkLinkMessage($tree, $type1, $xref1, $type2, $xref2) . ' ' .
248                        /* I18N: %1$s is an internal ID number such as R123. %2$s and %3$s are record types, such as INDI or SOUR */
249                        I18N::translate('%1$s is a %2$s but a %3$s is expected.', $this->checkLink($tree, $xref2), $this->formatType($type3), $this->formatType($type2));
250                } elseif (
251                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'FAMC', ['CHIL']) ||
252                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'FAMS', ['HUSB', 'WIFE']) ||
253                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'CHIL', ['FAMC']) ||
254                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'HUSB', ['FAMS']) ||
255                    $this->checkReverseLink($type2, $all_links, $xref1, $xref2, 'WIFE', ['FAMS'])
256                ) {
257                    /* I18N: %1$s and %2$s are internal ID numbers such as R123 */
258                    $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));
259                }
260            }
261        }
262
263        $title = I18N::translate('Check for errors') . ' — ' . e($tree->title());
264
265        return $this->viewResponse('admin/trees-check', [
266            'errors'   => $errors,
267            'title'    => $title,
268            'tree'     => $tree,
269            'warnings' => $warnings,
270        ]);
271    }
272
273    /**
274     * @param string     $type
275     * @param string[][] $links
276     * @param string     $xref1
277     * @param string     $xref2
278     * @param string     $link
279     * @param string[]   $reciprocal
280     *
281     * @return bool
282     */
283    private function checkReverseLink(string $type, array $links, string $xref1, string $xref2, string $link, array $reciprocal): bool
284    {
285        return $type === $link && (!array_key_exists($xref1, $links[$xref2]) || !in_array($links[$xref2][$xref1], $reciprocal, true));
286    }
287
288    /**
289     * Create a message linking one record to another.
290     *
291     * @param Tree   $tree
292     * @param string $type1
293     * @param string $xref1
294     * @param string $type2
295     * @param string $xref2
296     *
297     * @return string
298     */
299    private function checkLinkMessage(Tree $tree, string $type1, string $xref1, string $type2, $xref2): string
300    {
301        /* I18N: The placeholders are GEDCOM XREFs and tags. e.g. “INDI I123 contains a FAMC link to F234.” */
302        return I18N::translate(
303            '%1$s %2$s has a %3$s link to %4$s.',
304            $this->formatType($type1),
305            $this->checkLink($tree, $xref1),
306            $this->formatType($type2),
307            $this->checkLink($tree, $xref2)
308        );
309    }
310
311    /**
312     * Format a link to a record.
313     *
314     * @param Tree   $tree
315     * @param string $xref
316     *
317     * @return string
318     */
319    private function checkLink(Tree $tree, string $xref): string
320    {
321        return '<b><a href="' . e(route(GedcomRecordPage::class, [
322                'xref' => $xref,
323                'tree' => $tree->name(),
324            ])) . '">' . $xref . '</a></b>';
325    }
326
327    /**
328     * Format a record type.
329     *
330     * @param string $type
331     *
332     * @return string
333     */
334    private function formatType(string $type): string
335    {
336        return '<b title="' . GedcomTag::getLabel($type) . '">' . $type . '</b>';
337    }
338}
339