xref: /webtrees/app/Http/RequestHandlers/GedcomLoad.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 Exception;
23use Fisharebest\Webtrees\Exceptions\GedcomErrorException;
24use Fisharebest\Webtrees\Functions\FunctionsImport;
25use Fisharebest\Webtrees\Gedcom;
26use Fisharebest\Webtrees\Http\ViewResponseTrait;
27use Fisharebest\Webtrees\I18N;
28use Fisharebest\Webtrees\Services\TimeoutService;
29use Fisharebest\Webtrees\Services\TreeService;
30use Fisharebest\Webtrees\Tree;
31use Illuminate\Database\Capsule\Manager as DB;
32use Illuminate\Database\DetectsConcurrencyErrors;
33use Illuminate\Database\Query\Expression;
34use Psr\Http\Message\ResponseInterface;
35use Psr\Http\Message\ServerRequestInterface;
36use Psr\Http\Server\RequestHandlerInterface;
37
38use function assert;
39use function preg_match;
40use function preg_split;
41use function response;
42use function str_replace;
43use function str_starts_with;
44use function strlen;
45use function strtoupper;
46use function substr;
47use function trim;
48use function view;
49
50/**
51 * Load a chunk of GEDCOM data.
52 */
53class GedcomLoad implements RequestHandlerInterface
54{
55    use ViewResponseTrait;
56    use DetectsConcurrencyErrors;
57
58    /** @var TimeoutService */
59    private $timeout_service;
60
61    /** @var TreeService */
62    private $tree_service;
63
64    /**
65     * GedcomLoad constructor.
66     *
67     * @param TimeoutService $timeout_service
68     * @param TreeService    $tree_service
69     */
70    public function __construct(TimeoutService $timeout_service, TreeService $tree_service)
71    {
72        $this->timeout_service = $timeout_service;
73        $this->tree_service    = $tree_service;
74    }
75
76    /**
77     * @param ServerRequestInterface $request
78     *
79     * @return ResponseInterface
80     */
81    public function handle(ServerRequestInterface $request): ResponseInterface
82    {
83        $this->layout = 'layouts/ajax';
84
85        $tree = $request->getAttribute('tree');
86        assert($tree instanceof Tree);
87
88        try {
89            // What is the current import status?
90            $import_offset = DB::table('gedcom_chunk')
91                ->where('gedcom_id', '=', $tree->id())
92                ->where('imported', '=', '1')
93                ->count();
94
95            $import_total = DB::table('gedcom_chunk')
96                ->where('gedcom_id', '=', $tree->id())
97                ->count();
98
99            // Finished?
100            if ($import_offset === $import_total) {
101                $tree->setPreference('imported', '1');
102
103                $html = view('admin/import-complete', ['tree' => $tree]);
104
105                return response($html);
106            }
107
108            // Calculate progress so far
109            $progress = $import_offset / $import_total;
110
111            $first_time = ($import_offset === 0);
112
113            // Collect up any errors, and show them later.
114            $errors = '';
115
116            // Run for a short period of time. This keeps the resource requirements low.
117            do {
118                $data = DB::table('gedcom_chunk')
119                    ->where('gedcom_id', '=', $tree->id())
120                    ->where('imported', '=', '0')
121                    ->orderBy('gedcom_chunk_id')
122                    ->select(['gedcom_chunk_id', 'chunk_data'])
123                    ->first();
124
125                if ($data === null) {
126                    break;
127                }
128
129                // Mark the chunk as imported.  This will create a row-lock, to prevent other
130                // processes from reading it until we have finished.
131                $n = DB::table('gedcom_chunk')
132                    ->where('gedcom_chunk_id', '=', $data->gedcom_chunk_id)
133                    ->where('imported', '=', '0')
134                    ->update(['imported' => 1]);
135
136                // Another process has already imported this data?
137                if ($n === 0) {
138                    break;
139                }
140
141                // If we are loading the first (header) record, then delete old data and convert to UTF-8.
142                if ($first_time) {
143                    $this->tree_service->deleteGenealogyData($tree, (bool) $tree->getPreference('keep_media'));
144
145                    // Remove any byte-order-mark
146                    if (str_starts_with($data->chunk_data, Gedcom::UTF8_BOM)) {
147                        $data->chunk_data = substr($data->chunk_data, strlen(Gedcom::UTF8_BOM));
148                        // Put it back in the database, so we can do character conversion
149                        DB::table('gedcom_chunk')
150                            ->where('gedcom_chunk_id', '=', $data->gedcom_chunk_id)
151                            ->update(['chunk_data' => $data->chunk_data]);
152                    }
153
154                    if (!str_starts_with($data->chunk_data, '0 HEAD')) {
155                        return $this->viewResponse('admin/import-fail', [
156                            'error' => I18N::translate('Invalid GEDCOM file - no header record found.'),
157                            'tree'  => $tree,
158                        ]);
159                    }
160
161                    // What character set is this? Need to convert it to UTF8
162                    if (preg_match('/[\r\n][ \t]*1 CHAR(?:ACTER)? ([^\r\n]+)/', $data->chunk_data, $match)) {
163                        $charset = strtoupper(trim($match[1]));
164                    } else {
165                        $charset = 'ASCII';
166                    }
167
168                    // MySQL supports a wide range of collation conversions. These are ones that
169                    // have been encountered "in the wild".
170                    switch ($charset) {
171                        case 'ASCII':
172                            DB::table('gedcom_chunk')
173                                ->where('gedcom_id', '=', $tree->id())
174                                ->update(['chunk_data' => new Expression('CONVERT(CONVERT(chunk_data USING ascii) USING utf8)')]);
175                            break;
176                        case 'IBMPC':   // IBMPC, IBM WINDOWS and MS-DOS could be anything. Mostly it means CP850.
177                        case 'IBM WINDOWS':
178                        case 'MS-DOS':
179                        case 'CP437':
180                        case 'CP850':
181                            // CP850 has extra letters with diacritics to replace box-drawing chars in CP437.
182                            DB::table('gedcom_chunk')
183                                ->where('gedcom_id', '=', $tree->id())
184                                ->update(['chunk_data' => new Expression('CONVERT(CONVERT(chunk_data USING cp850) USING utf8)')]);
185                            break;
186                        case 'ANSI': // ANSI could be anything. Most applications seem to treat it as latin1.
187                        case 'WINDOWS':
188                        case 'CP1252':
189                        case 'ISO8859-1':
190                        case 'ISO-8859-1':
191                        case 'LATIN1':
192                        case 'LATIN-1':
193                            // Convert from ISO-8859-1 (western european) to UTF8.
194                            DB::table('gedcom_chunk')
195                                ->where('gedcom_id', '=', $tree->id())
196                                ->update(['chunk_data' => new Expression('CONVERT(CONVERT(chunk_data USING latin1) USING utf8)')]);
197                            break;
198                        case 'CP1250':
199                        case 'ISO8859-2':
200                        case 'ISO-8859-2':
201                        case 'LATIN2':
202                        case 'LATIN-2':
203                            // Convert from ISO-8859-2 (eastern european) to UTF8.
204                            DB::table('gedcom_chunk')
205                                ->where('gedcom_id', '=', $tree->id())
206                                ->update(['chunk_data' => new Expression('CONVERT(CONVERT(chunk_data USING latin2) USING utf8)')]);
207                            break;
208                        case 'MACINTOSH':
209                            // Convert from MAC Roman to UTF8.
210                            DB::table('gedcom_chunk')
211                                ->where('gedcom_id', '=', $tree->id())
212                                ->update(['chunk_data' => new Expression('CONVERT(CONVERT(chunk_data USING macroman) USING utf8)')]);
213                            break;
214                        case 'UTF8':
215                        case 'UTF-8':
216                            // Already UTF-8 so nothing to do!
217                            break;
218                        case 'ANSEL':
219                        default:
220                            return $this->viewResponse('admin/import-fail', [
221                                'error' => I18N::translate('Error: converting GEDCOM files from %s encoding to UTF-8 encoding not currently supported.', $charset),
222                                'tree'  => $tree,
223                            ]);
224                    }
225                    $first_time = false;
226
227                    // Re-fetch the data, now that we have performed character set conversion.
228                    $data = DB::table('gedcom_chunk')
229                        ->where('gedcom_chunk_id', '=', $data->gedcom_chunk_id)
230                        ->select(['gedcom_chunk_id', 'chunk_data'])
231                        ->first();
232                }
233
234                $data->chunk_data = str_replace("\r", "\n", $data->chunk_data);
235
236                // Import all the records in this chunk of data
237                foreach (preg_split('/\n+(?=0)/', $data->chunk_data) as $rec) {
238                    try {
239                        FunctionsImport::importRecord($rec, $tree, false);
240                    } catch (GedcomErrorException $exception) {
241                        $errors .= $exception->getMessage();
242                    }
243                }
244
245                // Do not need the data any more.
246                DB::table('gedcom_chunk')
247                    ->where('gedcom_chunk_id', '=', $data->gedcom_chunk_id)
248                    ->update(['chunk_data' => '']);
249            } while (!$this->timeout_service->isTimeLimitUp());
250
251            return $this->viewResponse('admin/import-progress', [
252                'errors'   => $errors,
253                'progress' => $progress,
254                'tree'     => $tree,
255            ]);
256        } catch (Exception $ex) {
257            DB::connection()->rollBack();
258
259            // Deadlock? Try again.
260            if ($this->causedByDeadlock($ex)) {
261                return $this->viewResponse('admin/import-progress', [
262                    'errors'   => '',
263                    'progress' => $progress ?? 0.0,
264                    'tree'     => $tree,
265                ]);
266            }
267
268            return $this->viewResponse('admin/import-fail', [
269                'error' => $ex->getMessage(),
270                'tree'  => $tree,
271            ]);
272        }
273    }
274}
275