1 <?php
  2  /**
  3  * This file is part of phpMorphy library
  4  *
  5  * Copyright c 2007-2008 Kamaev Vladimir <heromantor@users.sourceforge.net>
  6  *
  7  * This library is free software; you can redistribute it and/or
  8  * modify it under the terms of the GNU Lesser General Public
  9  * License as published by the Free Software Foundation; either
 10  * version 2 of the License, or (at your option) any later version.
 11  *
 12  * This library is distributed in the hope that it will be useful,
 13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
 14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 15  * Lesser General Public License for more details.
 16  *
 17  * You should have received a copy of the GNU Lesser General Public
 18  * License along with this library; if not, write to the
 19  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 20  * Boston, MA 02111-1307, USA.
 21  */
 22 
 23 interface phpMorphy_GramInfo_Interace {
 24     /**
 25      * Returns langugage for graminfo file
 26      * @return string
 27      */
 28     function getLocale();
 29     
 30     /**
 31      * Return encoding for graminfo file
 32      * @return string
 33      */
 34     function getEncoding();
 35     
 36     /**
 37      * Return size of character (cp1251 - 1, utf8 - 1, utf16 - 2, utf32 - 4 etc)
 38      * @return int
 39      */
 40     function getCharSize();
 41     
 42     /**
 43      * Return end of string value (usually string with \0 value of char_size + 1 length)
 44      * @return string
 45      */
 46     function getEnds();
 47     
 48     /**
 49      * Reads graminfo header
 50      *
 51      * @param int $offset
 52      * @return array
 53      */
 54     function readGramInfoHeader($offset);
 55 
 56     /**
 57      * Returns size of header struct
 58      */
 59     function getGramInfoHeaderSize();
 60     
 61     /**
 62      * Read ancodes section for header retrieved with readGramInfoHeader
 63      *
 64      * @param array $info
 65      * @return array
 66      */
 67     function readAncodes($info);
 68     
 69     /**
 70      * Read flexias section for header retrieved with readGramInfoHeader
 71      *
 72      * @param array $info
 73      * @return array
 74      */
 75     function readFlexiaData($info);
 76     
 77     /**
 78      * Read all graminfo headers offsets, which can be used latter for readGramInfoHeader method
 79      * @return array
 80      */
 81     function readAllGramInfoOffsets();
 82     
 83     function getHeader();
 84     function readAllPartOfSpeech();
 85     function readAllGrammems();
 86     function readAllAncodes();
 87 }
 88  
 89 abstract class phpMorphy_GramInfo implements phpMorphy_GramInfo_Interace {
 90     const HEADER_SIZE = 128;
 91     
 92     protected
 93         $resource,
 94         $header,
 95         $ends,
 96         $ends_size;
 97     
 98     protected function phpMorphy_GramInfo($resource, $header) {
 99         $this->resource = $resource;
100         $this->header = $header;
101         
102         $this->ends = str_repeat("\0", $header['char_size'] + 1);
103         $this->ends_size = $GLOBALS['__phpmorphy_strlen']($this->ends);
104     }
105     
106     static function create(phpMorphy_Storage $storage, $lazy) {
107         if($lazy) {
108             return new phpMorphy_GramInfo_Proxy($storage);
109         }
110         
111         $header = phpMorphy_GramInfo::readHeader(
112             $storage->read(0, self::HEADER_SIZE)
113         );
114         
115         if(!phpMorphy_GramInfo::validateHeader($header)) {
116             throw new phpMorphy_Exception('Invalid graminfo format');
117         }
118         
119         $storage_type = $storage->getTypeAsString();
120         $file_path = dirname(__FILE__) . "/access/graminfo_{$storage_type}.php";
121         $clazz = 'phpMorphy_GramInfo_' . ucfirst($storage_type);
122         
123         require_once($file_path);
124         return new $clazz($storage->getResource(), $header);
125     }
126     
127     function getLocale() {
128         return $this->header['lang'];
129     }
130     
131     function getEncoding() {
132         return $this->header['encoding'];
133     }
134     
135     function getCharSize() {
136         return $this->header['char_size'];
137     }
138     
139     function getEnds() {
140         return $this->ends;
141     }
142     
143     function getHeader() {
144         return $this->header;
145     }
146     
147     static protected function readHeader($headerRaw) {
148         $header = unpack(
149             'Vver/Vis_be/Vflex_count_old/' .
150             'Vflex_offset/Vflex_size/Vflex_count/Vflex_index_offset/Vflex_index_size/' .
151             'Vposes_offset/Vposes_size/Vposes_count/Vposes_index_offset/Vposes_index_size/' .
152             'Vgrammems_offset/Vgrammems_size/Vgrammems_count/Vgrammems_index_offset/Vgrammems_index_size/' .
153             'Vancodes_offset/Vancodes_size/Vancodes_count/Vancodes_index_offset/Vancodes_index_size/' .
154             'Vchar_size/',
155             $headerRaw
156         );
157         
158         $offset = 24 * 4;
159         $len = ord($GLOBALS['__phpmorphy_substr']($headerRaw, $offset++, 1));
160         $header['lang'] = rtrim($GLOBALS['__phpmorphy_substr']($headerRaw, $offset, $len));
161         
162         $offset += $len;
163         
164         $len = ord($GLOBALS['__phpmorphy_substr']($headerRaw, $offset++, 1));
165         $header['encoding'] = rtrim($GLOBALS['__phpmorphy_substr']($headerRaw, $offset, $len));
166         
167         return $header;
168     }
169     
170     static protected function validateHeader($header) {
171         if(
172             3 != $header['ver'] ||
173             1 == $header['is_be']
174         ) {
175             return false;
176         }
177         
178         return true;
179     }
180     
181     protected function cleanupCString($string) {
182         if(false !== ($pos = $GLOBALS['__phpmorphy_strpos']($string, $this->ends))) {
183             $string = $GLOBALS['__phpmorphy_substr']($string, 0, $pos);
184         }
185         
186         return $string;
187     }
188     
189     abstract protected function readSectionIndex($offset, $count);
190     
191     protected function readSectionIndexAsSize($offset, $count, $total_size) {
192         if(!$count) {
193             return array();
194         }
195         
196         $index = $this->readSectionIndex($offset, $count);
197         $index[$count] = $index[0] + $total_size;
198         
199         for($i = 0; $i < $count; $i++) {
200             $index[$i] = $index[$i + 1] - $index[$i];
201         }
202         
203         unset($index[$count]);
204         
205         return $index;
206     }
207 };
208 
209 class phpMorphy_GramInfo_Decorator implements phpMorphy_GramInfo_Interace {
210     protected $info;
211     
212     function phpMorphy_GramInfo_Decorator(phpMorphy_GramInfo_Interace $info) {
213         $this->info = $info;
214     }
215     
216     function readGramInfoHeader($offset) { return $this->info->readGramInfoHeader($offset); }
217     function getGramInfoHeaderSize() { return $this->info->getGramInfoHeaderSize($offset); }
218     function readAncodes($info) { return $this->info->readAncodes($info); }
219     function readFlexiaData($info) { return $this->info->readFlexiaData($info); }
220     function readAllGramInfoOffsets() { return $this->info->readAllGramInfoOffsets(); }
221     function readAllPartOfSpeech() { return $this->info->readAllPartOfSpeech(); }
222     function readAllGrammems() { return $this->info->readAllGrammems(); }
223     function readAllAncodes() { return $this->info->readAllAncodes(); }
224     
225     function getLocale()  { return $this->info->getLocale(); }
226     function getEncoding()  { return $this->info->getEncoding(); }
227     function getCharSize()  { return $this->info->getCharSize(); }
228     function getEnds() { return $this->info->getEnds(); }
229     function getHeader() { return $this->info->getHeader(); }
230 }
231 
232 class phpMorphy_GramInfo_Proxy extends phpMorphy_GramInfo_Decorator {
233     protected $storage;
234     
235     function __construct(phpMorphy_Storage $storage) {
236         $this->storage = $storage;
237         unset($this->info);
238     }
239     
240     function __get($propName) {
241         if($propName == 'info') {
242             $this->info = phpMorphy_GramInfo::create($this->storage, false);
243             unset($this->storage);
244             return $this->info;
245         }
246         
247         throw new phpMorphy_Exception("Unknown prop name '$propName'");
248     }
249 }
250 
251 class phpMorphy_GramInfo_Proxy_WithHeader extends phpMorphy_GramInfo_Proxy {
252     protected
253         $cache,
254         $ends;
255     
256     function __construct(phpMorphy_Storage $storage, $cacheFile) {
257         parent::__construct($storage);
258         
259         $this->cache = $this->readCache($cacheFile);
260         $this->ends = str_repeat("\0", $this->getCharSize() + 1);
261     }
262     
263     protected function readCache($fileName) {
264         if(!is_array($result = include($fileName))) {
265             throw new phpMorphy_Exception("Can`t get header cache from '$fileName' file'");
266         }
267         
268         return $result;
269     }
270     
271     function getLocale()  {
272         return $this->cache['lang'];
273     }
274     
275     function getEncoding()  {
276         return $this->cache['encoding'];
277     }
278     
279     function getCharSize()  {
280         return $this->cache['char_size'];
281     }
282     
283     function getEnds() {
284         return $this->ends;
285     }
286     
287     function getHeader() {
288         return $this->cache;
289     }
290 }
291 
292 class phpMorphy_GramInfo_RuntimeCaching extends phpMorphy_GramInfo_Decorator {
293     protected
294         $flexia = array(),
295         $ancodes = array();
296     
297     function readFlexiaData($info) {
298         $offset = $info['offset'];
299         
300         if(!isset($this->flexia_all[$offset])) {
301             $this->flexia_all[$offset] = $this->info->readFlexiaData($info);
302         }
303         
304         return $this->flexia_all[$offset];
305     }
306 }
307 
308 class phpMorphy_GramInfo_AncodeCache extends phpMorphy_GramInfo_Decorator {
309     public
310         $hits = 0,
311         $miss = 0;
312 
313     protected
314         $cache;
315 
316     function __construct(phpMorphy_GramInfo_Interace $inner, $resource) {
317         parent::__construct($inner);
318 
319         if(false === ($this->cache = unserialize($resource->read(0, $resource->getFileSize())))) {
320             throw new phpMorphy_Exception("Can`t read ancodes cache");
321         }
322     }
323 
324     function readAncodes($info) {
325         $offset = $info['offset'];
326 
327         if(isset($this->cache[$offset])) {
328             $this->hits++;
329 
330             return $this->cache[$offset];
331         } else {
332             // in theory misses never occur
333             $this->miss++;
334 
335             return parent::readAncodes($info);
336         }
337     }
338 }