Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
Total | |
0.00% |
0 / 200 |
|
0.00% |
0 / 11 |
CRAP | |
0.00% |
0 / 2 |
GenerateCollationData | |
0.00% |
0 / 140 |
|
0.00% |
0 / 5 |
2162 | |
0.00% |
0 / 1 |
__construct | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
2 | |||
execute | |
0.00% |
0 / 31 |
|
0.00% |
0 / 1 |
90 | |||
loadUcd | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
charCallback | |
0.00% |
0 / 22 |
|
0.00% |
0 / 1 |
110 | |||
generateFirstChars | |
0.00% |
0 / 79 |
|
0.00% |
0 / 1 |
650 | |||
UcdXmlReader | |
0.00% |
0 / 60 |
|
0.00% |
0 / 6 |
992 | |
0.00% |
0 / 1 |
__construct | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
readChars | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
90 | |||
open | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
20 | |||
readAttributes | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
handleChar | |
0.00% |
0 / 21 |
|
0.00% |
0 / 1 |
72 | |||
getBlocks | |
0.00% |
0 / 13 |
|
0.00% |
0 / 1 |
56 |
1 | <?php |
2 | /** |
3 | * Maintenance script to generate first letter data files for Collation.php. |
4 | * |
5 | * This program is free software; you can redistribute it and/or modify |
6 | * it under the terms of the GNU General Public License as published by |
7 | * the Free Software Foundation; either version 2 of the License, or |
8 | * (at your option) any later version. |
9 | * |
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 | * |
15 | * You should have received a copy of the GNU General Public License along |
16 | * with this program; if not, write to the Free Software Foundation, Inc., |
17 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
18 | * http://www.gnu.org/copyleft/gpl.html |
19 | * |
20 | * @file |
21 | * @ingroup MaintenanceLanguage |
22 | */ |
23 | |
24 | // @codeCoverageIgnoreStart |
25 | require_once __DIR__ . '/../Maintenance.php'; |
26 | // @codeCoverageIgnoreEnd |
27 | |
28 | use Wikimedia\StaticArrayWriter; |
29 | |
30 | /** |
31 | * Generate first letter data files for Collation.php |
32 | * |
33 | * @ingroup MaintenanceLanguage |
34 | */ |
35 | class GenerateCollationData extends Maintenance { |
36 | /** @var string The directory with source data files in it */ |
37 | public $dataDir; |
38 | |
39 | /** @var int The primary weights, indexed by codepoint */ |
40 | public $weights; |
41 | |
42 | /** |
43 | * A hashtable keyed by codepoint, where presence indicates that a character |
44 | * has a decomposition mapping. This makes it non-preferred for group header |
45 | * selection. |
46 | * @var string[] |
47 | */ |
48 | public $mappedChars; |
49 | |
50 | /** @var string */ |
51 | public $debugOutFile; |
52 | |
53 | /** @var string[] */ |
54 | private $groups; |
55 | |
56 | public function __construct() { |
57 | parent::__construct(); |
58 | $this->addOption( 'data-dir', 'A directory on the local filesystem ' . |
59 | 'containing allkeys.txt and ucd.all.grouped.xml from unicode.org', |
60 | false, true ); |
61 | $this->addOption( 'debug-output', 'Filename for sending debug output to', |
62 | false, true ); |
63 | } |
64 | |
65 | public function execute() { |
66 | $this->dataDir = $this->getOption( 'data-dir', '.' ); |
67 | |
68 | $allkeysPresent = file_exists( "{$this->dataDir}/allkeys.txt" ); |
69 | $ucdallPresent = file_exists( "{$this->dataDir}/ucd.all.grouped.xml" ); |
70 | |
71 | if ( !$allkeysPresent || !$ucdallPresent ) { |
72 | $icuVersion = INTL_ICU_VERSION; |
73 | $unicodeVersion = implode( '.', array_slice( IntlChar::getUnicodeVersion(), 0, 3 ) ); |
74 | |
75 | $error = ""; |
76 | |
77 | if ( !$allkeysPresent ) { |
78 | $error .= "Unable to find allkeys.txt. " |
79 | . "Download it and specify its location with --data-dir=<DIR>. " |
80 | . "\n\n"; |
81 | } |
82 | if ( !$ucdallPresent ) { |
83 | $error .= "Unable to find ucd.all.grouped.xml. " |
84 | . "Download it, unzip, and specify its location with --data-dir=<DIR>. " |
85 | . "\n\n"; |
86 | } |
87 | |
88 | $error .= "You are using ICU $icuVersion, intended for Unicode $unicodeVersion. " |
89 | . "Appropriate file(s) should be available at:\n"; |
90 | |
91 | $allkeysURL = "https://www.unicode.org/Public/UCA/$unicodeVersion/allkeys.txt"; |
92 | $ucdallURL = "https://www.unicode.org/Public/$unicodeVersion/ucdxml/ucd.all.grouped.zip"; |
93 | |
94 | if ( !$allkeysPresent ) { |
95 | $error .= "* $allkeysURL\n"; |
96 | } |
97 | if ( !$ucdallPresent ) { |
98 | $error .= "* $ucdallURL\n"; |
99 | } |
100 | |
101 | $this->fatalError( $error ); |
102 | } |
103 | |
104 | $debugOutFileName = $this->getOption( 'debug-output' ); |
105 | if ( $debugOutFileName ) { |
106 | $this->debugOutFile = fopen( $debugOutFileName, 'w' ); |
107 | if ( !$this->debugOutFile ) { |
108 | $this->fatalError( "Unable to open debug output file for writing" ); |
109 | } |
110 | } |
111 | $this->loadUcd(); |
112 | $this->generateFirstChars(); |
113 | } |
114 | |
115 | private function loadUcd() { |
116 | $uxr = new UcdXmlReader( "{$this->dataDir}/ucd.all.grouped.xml" ); |
117 | $uxr->readChars( [ $this, 'charCallback' ] ); |
118 | } |
119 | |
120 | private function charCallback( $data ) { |
121 | // Skip non-printable characters, |
122 | // but do not skip a normal space (U+0020) since |
123 | // people like to use that as a fake no header symbol. |
124 | $category = substr( $data['gc'], 0, 1 ); |
125 | if ( strpos( 'LNPS', $category ) === false |
126 | && $data['cp'] !== '0020' |
127 | ) { |
128 | return; |
129 | } |
130 | $cp = hexdec( $data['cp'] ); |
131 | |
132 | // Skip the CJK ideograph blocks, as an optimisation measure. |
133 | // UCA doesn't sort them properly anyway, without tailoring. |
134 | if ( IcuCollation::isCjk( $cp ) ) { |
135 | return; |
136 | } |
137 | |
138 | // Skip the composed Hangul syllables, we will use the bare Jamo |
139 | // as first letters |
140 | if ( $data['block'] == 'Hangul Syllables' ) { |
141 | return; |
142 | } |
143 | |
144 | // Calculate implicit weight per UTS #10 v6.0.0, sec 7.1.3 |
145 | if ( $data['UIdeo'] === 'Y' ) { |
146 | if ( $data['block'] == 'CJK Unified Ideographs' |
147 | || $data['block'] == 'CJK Compatibility Ideographs' |
148 | ) { |
149 | $base = 0xFB40; |
150 | } else { |
151 | $base = 0xFB80; |
152 | } |
153 | } else { |
154 | $base = 0xFBC0; |
155 | } |
156 | $a = $base + ( $cp >> 15 ); |
157 | $b = ( $cp & 0x7fff ) | 0x8000; |
158 | |
159 | $this->weights[$cp] = sprintf( ".%04X.%04X", $a, $b ); |
160 | |
161 | if ( $data['dm'] !== '#' ) { |
162 | $this->mappedChars[$cp] = true; |
163 | } |
164 | |
165 | if ( $cp % 4096 == 0 ) { |
166 | print "{$data['cp']}\n"; |
167 | } |
168 | } |
169 | |
170 | private function generateFirstChars() { |
171 | $file = fopen( "{$this->dataDir}/allkeys.txt", 'r' ); |
172 | if ( !$file ) { |
173 | $this->fatalError( "Unable to open allkeys.txt" ); |
174 | } |
175 | |
176 | $goodTertiaryChars = []; |
177 | |
178 | // For each character with an entry in allkeys.txt, overwrite the implicit |
179 | // entry in $this->weights that came from the UCD. |
180 | // Also gather a list of tertiary weights, for use in selecting the group header |
181 | // phpcs:ignore Generic.CodeAnalysis.AssignmentInCondition.FoundInWhileCondition |
182 | while ( ( $line = fgets( $file ) ) !== false ) { |
183 | // We're only interested in single-character weights, pick them out with a regex |
184 | $line = trim( $line ); |
185 | if ( !preg_match( '/^([0-9A-F]+)\s*;\s*([^#]*)/', $line, $m ) ) { |
186 | continue; |
187 | } |
188 | |
189 | $cp = hexdec( $m[1] ); |
190 | $allWeights = trim( $m[2] ); |
191 | $primary = ''; |
192 | $tertiary = ''; |
193 | |
194 | if ( !isset( $this->weights[$cp] ) ) { |
195 | // Non-printable, ignore |
196 | continue; |
197 | } |
198 | foreach ( StringUtils::explode( '[', $allWeights ) as $weightStr ) { |
199 | if ( preg_match_all( '/[*.]([0-9A-F]+)/', $weightStr, $m ) ) { |
200 | if ( $m[1][0] !== '0000' ) { |
201 | $primary .= '.' . $m[1][0]; |
202 | } |
203 | if ( $m[1][2] !== '0000' ) { |
204 | $tertiary .= '.' . $m[1][2]; |
205 | } |
206 | } |
207 | } |
208 | $this->weights[$cp] = $primary; |
209 | if ( $tertiary === '.0008' |
210 | || $tertiary === '.000E' |
211 | ) { |
212 | $goodTertiaryChars[$cp] = true; |
213 | } |
214 | } |
215 | fclose( $file ); |
216 | |
217 | // Identify groups of characters with the same primary weight |
218 | $this->groups = []; |
219 | asort( $this->weights, SORT_STRING ); |
220 | $prevWeight = reset( $this->weights ); |
221 | $group = []; |
222 | foreach ( $this->weights as $cp => $weight ) { |
223 | if ( $weight !== $prevWeight ) { |
224 | $this->groups[$prevWeight] = $group; |
225 | $prevWeight = $weight; |
226 | $group = $this->groups[$weight] ?? []; |
227 | } |
228 | $group[] = $cp; |
229 | } |
230 | if ( $group ) { |
231 | $this->groups[$prevWeight] = $group; |
232 | } |
233 | |
234 | // If one character has a given primary weight sequence, and a second |
235 | // character has a longer primary weight sequence with an initial |
236 | // portion equal to the first character, then remove the second |
237 | // character. This avoids having characters like U+A732 (double A) |
238 | // polluting the basic Latin sort area. |
239 | |
240 | foreach ( $this->groups as $weight => $group ) { |
241 | if ( preg_match( '/(\.[0-9A-F]*)\./', $weight, $m ) ) { |
242 | if ( isset( $this->groups[$m[1]] ) ) { |
243 | unset( $this->groups[$weight] ); |
244 | } |
245 | } |
246 | } |
247 | |
248 | ksort( $this->groups, SORT_STRING ); |
249 | |
250 | // Identify the header character in each group |
251 | $headerChars = []; |
252 | $prevChar = "\000"; |
253 | $tertiaryCollator = new Collator( 'root' ); |
254 | $primaryCollator = new Collator( 'root' ); |
255 | $primaryCollator->setStrength( Collator::PRIMARY ); |
256 | $numOutOfOrder = 0; |
257 | foreach ( $this->groups as $weight => $group ) { |
258 | $uncomposedChars = []; |
259 | $goodChars = []; |
260 | foreach ( $group as $cp ) { |
261 | if ( isset( $goodTertiaryChars[$cp] ) ) { |
262 | $goodChars[] = $cp; |
263 | } |
264 | if ( !isset( $this->mappedChars[$cp] ) ) { |
265 | $uncomposedChars[] = $cp; |
266 | } |
267 | } |
268 | $x = array_intersect( $goodChars, $uncomposedChars ); |
269 | if ( !$x ) { |
270 | $x = $uncomposedChars; |
271 | if ( !$x ) { |
272 | $x = $group; |
273 | } |
274 | } |
275 | |
276 | // Use ICU to pick the lowest sorting character in the selection |
277 | $tertiaryCollator->sort( $x ); |
278 | $cp = $x[0]; |
279 | |
280 | $char = UtfNormal\Utils::codepointToUtf8( $cp ); |
281 | $headerChars[] = $char; |
282 | if ( $primaryCollator->compare( $char, $prevChar ) <= 0 ) { |
283 | $numOutOfOrder++; |
284 | } |
285 | $prevChar = $char; |
286 | |
287 | if ( $this->debugOutFile ) { |
288 | fwrite( $this->debugOutFile, sprintf( "%05X %s %s (%s)\n", $cp, $weight, $char, |
289 | implode( ' ', array_map( [ UtfNormal\Utils::class, 'codepointToUtf8' ], $group ) ) ) ); |
290 | } |
291 | } |
292 | |
293 | print "Out of order: $numOutOfOrder / " . count( $headerChars ) . "\n"; |
294 | |
295 | global $IP; |
296 | $writer = new StaticArrayWriter(); |
297 | file_put_contents( |
298 | "$IP/includes/collation/data/first-letters-root.php", |
299 | $writer->create( $headerChars, 'File created by generateCollationData.php' ) |
300 | ); |
301 | echo "first-letters-root: file written.\n"; |
302 | } |
303 | } |
304 | |
305 | class UcdXmlReader { |
306 | /** @var string */ |
307 | public $fileName; |
308 | /** @var callable */ |
309 | public $callback; |
310 | /** @var array */ |
311 | public $groupAttrs; |
312 | /** @var XMLReader */ |
313 | public $xml; |
314 | /** @var array[] */ |
315 | public $blocks = []; |
316 | /** @var array */ |
317 | public $currentBlock; |
318 | |
319 | public function __construct( $fileName ) { |
320 | $this->fileName = $fileName; |
321 | } |
322 | |
323 | public function readChars( $callback ) { |
324 | $this->getBlocks(); |
325 | $this->currentBlock = reset( $this->blocks ); |
326 | $xml = $this->open(); |
327 | $this->callback = $callback; |
328 | |
329 | while ( $xml->name !== 'repertoire' && $xml->next() ); |
330 | |
331 | while ( $xml->read() ) { |
332 | if ( $xml->nodeType == XMLReader::ELEMENT ) { |
333 | if ( $xml->name === 'group' ) { |
334 | $this->groupAttrs = $this->readAttributes(); |
335 | } elseif ( $xml->name === 'char' ) { |
336 | $this->handleChar(); |
337 | } |
338 | } elseif ( $xml->nodeType === XMLReader::END_ELEMENT ) { |
339 | if ( $xml->name === 'group' ) { |
340 | $this->groupAttrs = []; |
341 | } |
342 | } |
343 | } |
344 | $xml->close(); |
345 | } |
346 | |
347 | protected function open() { |
348 | $this->xml = new XMLReader; |
349 | if ( !$this->xml->open( $this->fileName ) ) { |
350 | throw new RuntimeException( __METHOD__ . ": unable to open {$this->fileName}" ); |
351 | } |
352 | while ( $this->xml->name !== 'ucd' && $this->xml->read() ); |
353 | $this->xml->read(); |
354 | |
355 | return $this->xml; |
356 | } |
357 | |
358 | /** |
359 | * Read the attributes of the current element node and return them |
360 | * as an array |
361 | * @return array |
362 | */ |
363 | protected function readAttributes() { |
364 | $attrs = []; |
365 | while ( $this->xml->moveToNextAttribute() ) { |
366 | $attrs[$this->xml->name] = $this->xml->value; |
367 | } |
368 | |
369 | return $attrs; |
370 | } |
371 | |
372 | protected function handleChar() { |
373 | $attrs = $this->readAttributes() + $this->groupAttrs; |
374 | if ( isset( $attrs['cp'] ) ) { |
375 | $first = $last = hexdec( $attrs['cp'] ); |
376 | } else { |
377 | $first = hexdec( $attrs['first-cp'] ); |
378 | $last = hexdec( $attrs['last-cp'] ); |
379 | unset( $attrs['first-cp'] ); |
380 | unset( $attrs['last-cp'] ); |
381 | } |
382 | |
383 | for ( $cp = $first; $cp <= $last; $cp++ ) { |
384 | $hexCp = sprintf( "%04X", $cp ); |
385 | foreach ( [ 'na', 'na1' ] as $nameProp ) { |
386 | if ( isset( $attrs[$nameProp] ) ) { |
387 | $attrs[$nameProp] = str_replace( '#', $hexCp, $attrs[$nameProp] ); |
388 | } |
389 | } |
390 | |
391 | while ( $this->currentBlock ) { |
392 | if ( $cp < $this->currentBlock[0] ) { |
393 | break; |
394 | } elseif ( $cp <= $this->currentBlock[1] ) { |
395 | $attrs['block'] = key( $this->blocks ); |
396 | break; |
397 | } else { |
398 | $this->currentBlock = next( $this->blocks ); |
399 | } |
400 | } |
401 | |
402 | $attrs['cp'] = $hexCp; |
403 | call_user_func( $this->callback, $attrs ); |
404 | } |
405 | } |
406 | |
407 | public function getBlocks() { |
408 | if ( $this->blocks ) { |
409 | return $this->blocks; |
410 | } |
411 | |
412 | $xml = $this->open(); |
413 | while ( $xml->name !== 'blocks' && $xml->read() ); |
414 | |
415 | while ( $xml->read() ) { |
416 | if ( $xml->nodeType == XMLReader::ELEMENT ) { |
417 | if ( $xml->name === 'block' ) { |
418 | $attrs = $this->readAttributes(); |
419 | $first = hexdec( $attrs['first-cp'] ); |
420 | $last = hexdec( $attrs['last-cp'] ); |
421 | $this->blocks[$attrs['name']] = [ $first, $last ]; |
422 | } |
423 | } |
424 | } |
425 | $xml->close(); |
426 | |
427 | return $this->blocks; |
428 | } |
429 | } |
430 | |
431 | // @codeCoverageIgnoreStart |
432 | $maintClass = GenerateCollationData::class; |
433 | require_once RUN_MAINTENANCE_IF_MAIN; |
434 | // @codeCoverageIgnoreEnd |