]>
Commit | Line | Data |
---|---|---|
e4f7f8df AD |
1 | <?php |
2 | ||
3 | // | |
4 | // $Id: sphinxapi.php 2055 2009-11-06 23:09:58Z shodan $ | |
5 | // | |
6 | ||
7 | // | |
8 | // Copyright (c) 2001-2008, Andrew Aksyonoff. All rights reserved. | |
9 | // | |
10 | // This program is free software; you can redistribute it and/or modify | |
11 | // it under the terms of the GNU General Public License. You should have | |
12 | // received a copy of the GPL license along with this program; if you | |
13 | // did not, you can find it at http://www.gnu.org/ | |
14 | // | |
15 | ||
16 | ///////////////////////////////////////////////////////////////////////////// | |
17 | // PHP version of Sphinx searchd client (PHP API) | |
18 | ///////////////////////////////////////////////////////////////////////////// | |
19 | ||
20 | /// known searchd commands | |
21 | define ( "SEARCHD_COMMAND_SEARCH", 0 ); | |
22 | define ( "SEARCHD_COMMAND_EXCERPT", 1 ); | |
23 | define ( "SEARCHD_COMMAND_UPDATE", 2 ); | |
24 | define ( "SEARCHD_COMMAND_KEYWORDS",3 ); | |
25 | define ( "SEARCHD_COMMAND_PERSIST", 4 ); | |
26 | define ( "SEARCHD_COMMAND_STATUS", 5 ); | |
27 | define ( "SEARCHD_COMMAND_QUERY", 6 ); | |
28 | ||
29 | /// current client-side command implementation versions | |
30 | define ( "VER_COMMAND_SEARCH", 0x116 ); | |
31 | define ( "VER_COMMAND_EXCERPT", 0x100 ); | |
32 | define ( "VER_COMMAND_UPDATE", 0x102 ); | |
33 | define ( "VER_COMMAND_KEYWORDS", 0x100 ); | |
34 | define ( "VER_COMMAND_STATUS", 0x100 ); | |
35 | define ( "VER_COMMAND_QUERY", 0x100 ); | |
36 | ||
37 | /// known searchd status codes | |
38 | define ( "SEARCHD_OK", 0 ); | |
39 | define ( "SEARCHD_ERROR", 1 ); | |
40 | define ( "SEARCHD_RETRY", 2 ); | |
41 | define ( "SEARCHD_WARNING", 3 ); | |
42 | ||
43 | /// known match modes | |
44 | define ( "SPH_MATCH_ALL", 0 ); | |
45 | define ( "SPH_MATCH_ANY", 1 ); | |
46 | define ( "SPH_MATCH_PHRASE", 2 ); | |
47 | define ( "SPH_MATCH_BOOLEAN", 3 ); | |
48 | define ( "SPH_MATCH_EXTENDED", 4 ); | |
49 | define ( "SPH_MATCH_FULLSCAN", 5 ); | |
50 | define ( "SPH_MATCH_EXTENDED2", 6 ); // extended engine V2 (TEMPORARY, WILL BE REMOVED) | |
51 | ||
52 | /// known ranking modes (ext2 only) | |
53 | define ( "SPH_RANK_PROXIMITY_BM25", 0 ); ///< default mode, phrase proximity major factor and BM25 minor one | |
54 | define ( "SPH_RANK_BM25", 1 ); ///< statistical mode, BM25 ranking only (faster but worse quality) | |
55 | define ( "SPH_RANK_NONE", 2 ); ///< no ranking, all matches get a weight of 1 | |
56 | define ( "SPH_RANK_WORDCOUNT", 3 ); ///< simple word-count weighting, rank is a weighted sum of per-field keyword occurence counts | |
57 | define ( "SPH_RANK_PROXIMITY", 4 ); | |
58 | define ( "SPH_RANK_MATCHANY", 5 ); | |
59 | define ( "SPH_RANK_FIELDMASK", 6 ); | |
60 | ||
61 | /// known sort modes | |
62 | define ( "SPH_SORT_RELEVANCE", 0 ); | |
63 | define ( "SPH_SORT_ATTR_DESC", 1 ); | |
64 | define ( "SPH_SORT_ATTR_ASC", 2 ); | |
65 | define ( "SPH_SORT_TIME_SEGMENTS", 3 ); | |
66 | define ( "SPH_SORT_EXTENDED", 4 ); | |
67 | define ( "SPH_SORT_EXPR", 5 ); | |
68 | ||
69 | /// known filter types | |
70 | define ( "SPH_FILTER_VALUES", 0 ); | |
71 | define ( "SPH_FILTER_RANGE", 1 ); | |
72 | define ( "SPH_FILTER_FLOATRANGE", 2 ); | |
73 | ||
74 | /// known attribute types | |
75 | define ( "SPH_ATTR_INTEGER", 1 ); | |
76 | define ( "SPH_ATTR_TIMESTAMP", 2 ); | |
77 | define ( "SPH_ATTR_ORDINAL", 3 ); | |
78 | define ( "SPH_ATTR_BOOL", 4 ); | |
79 | define ( "SPH_ATTR_FLOAT", 5 ); | |
80 | define ( "SPH_ATTR_BIGINT", 6 ); | |
81 | define ( "SPH_ATTR_MULTI", 0x40000000 ); | |
82 | ||
83 | /// known grouping functions | |
84 | define ( "SPH_GROUPBY_DAY", 0 ); | |
85 | define ( "SPH_GROUPBY_WEEK", 1 ); | |
86 | define ( "SPH_GROUPBY_MONTH", 2 ); | |
87 | define ( "SPH_GROUPBY_YEAR", 3 ); | |
88 | define ( "SPH_GROUPBY_ATTR", 4 ); | |
89 | define ( "SPH_GROUPBY_ATTRPAIR", 5 ); | |
90 | ||
91 | // important properties of PHP's integers: | |
92 | // - always signed (one bit short of PHP_INT_SIZE) | |
93 | // - conversion from string to int is saturated | |
94 | // - float is double | |
95 | // - div converts arguments to floats | |
96 | // - mod converts arguments to ints | |
97 | ||
98 | // the packing code below works as follows: | |
99 | // - when we got an int, just pack it | |
100 | // if performance is a problem, this is the branch users should aim for | |
101 | // | |
102 | // - otherwise, we got a number in string form | |
103 | // this might be due to different reasons, but we assume that this is | |
104 | // because it didn't fit into PHP int | |
105 | // | |
106 | // - factor the string into high and low ints for packing | |
107 | // - if we have bcmath, then it is used | |
108 | // - if we don't, we have to do it manually (this is the fun part) | |
109 | // | |
110 | // - x64 branch does factoring using ints | |
111 | // - x32 (ab)uses floats, since we can't fit unsigned 32-bit number into an int | |
112 | // | |
113 | // unpacking routines are pretty much the same. | |
114 | // - return ints if we can | |
115 | // - otherwise format number into a string | |
116 | ||
117 | /// pack 64-bit signed | |
118 | function sphPackI64 ( $v ) | |
119 | { | |
120 | assert ( is_numeric($v) ); | |
121 | ||
122 | // x64 | |
123 | if ( PHP_INT_SIZE>=8 ) | |
124 | { | |
125 | $v = (int)$v; | |
126 | return pack ( "NN", $v>>32, $v&0xFFFFFFFF ); | |
127 | } | |
128 | ||
129 | // x32, int | |
130 | if ( is_int($v) ) | |
131 | return pack ( "NN", $v < 0 ? -1 : 0, $v ); | |
132 | ||
133 | // x32, bcmath | |
134 | if ( function_exists("bcmul") ) | |
135 | { | |
136 | if ( bccomp ( $v, 0 ) == -1 ) | |
137 | $v = bcadd ( "18446744073709551616", $v ); | |
138 | $h = bcdiv ( $v, "4294967296", 0 ); | |
139 | $l = bcmod ( $v, "4294967296" ); | |
140 | return pack ( "NN", (float)$h, (float)$l ); // conversion to float is intentional; int would lose 31st bit | |
141 | } | |
142 | ||
143 | // x32, no-bcmath | |
144 | $p = max(0, strlen($v) - 13); | |
145 | $lo = abs((float)substr($v, $p)); | |
146 | $hi = abs((float)substr($v, 0, $p)); | |
147 | ||
148 | $m = $lo + $hi*1316134912.0; // (10 ^ 13) % (1 << 32) = 1316134912 | |
149 | $q = floor($m/4294967296.0); | |
150 | $l = $m - ($q*4294967296.0); | |
151 | $h = $hi*2328.0 + $q; // (10 ^ 13) / (1 << 32) = 2328 | |
152 | ||
153 | if ( $v<0 ) | |
154 | { | |
155 | if ( $l==0 ) | |
156 | $h = 4294967296.0 - $h; | |
157 | else | |
158 | { | |
159 | $h = 4294967295.0 - $h; | |
160 | $l = 4294967296.0 - $l; | |
161 | } | |
162 | } | |
163 | return pack ( "NN", $h, $l ); | |
164 | } | |
165 | ||
166 | /// pack 64-bit unsigned | |
167 | function sphPackU64 ( $v ) | |
168 | { | |
169 | assert ( is_numeric($v) ); | |
170 | ||
171 | // x64 | |
172 | if ( PHP_INT_SIZE>=8 ) | |
173 | { | |
174 | assert ( $v>=0 ); | |
175 | ||
176 | // x64, int | |
177 | if ( is_int($v) ) | |
178 | return pack ( "NN", $v>>32, $v&0xFFFFFFFF ); | |
179 | ||
180 | // x64, bcmath | |
181 | if ( function_exists("bcmul") ) | |
182 | { | |
183 | $h = bcdiv ( $v, 4294967296, 0 ); | |
184 | $l = bcmod ( $v, 4294967296 ); | |
185 | return pack ( "NN", $h, $l ); | |
186 | } | |
187 | ||
188 | // x64, no-bcmath | |
189 | $p = max ( 0, strlen($v) - 13 ); | |
190 | $lo = (int)substr ( $v, $p ); | |
191 | $hi = (int)substr ( $v, 0, $p ); | |
192 | ||
193 | $m = $lo + $hi*1316134912; | |
194 | $l = $m % 4294967296; | |
195 | $h = $hi*2328 + (int)($m/4294967296); | |
196 | ||
197 | return pack ( "NN", $h, $l ); | |
198 | } | |
199 | ||
200 | // x32, int | |
201 | if ( is_int($v) ) | |
202 | return pack ( "NN", 0, $v ); | |
203 | ||
204 | // x32, bcmath | |
205 | if ( function_exists("bcmul") ) | |
206 | { | |
207 | $h = bcdiv ( $v, "4294967296", 0 ); | |
208 | $l = bcmod ( $v, "4294967296" ); | |
209 | return pack ( "NN", (float)$h, (float)$l ); // conversion to float is intentional; int would lose 31st bit | |
210 | } | |
211 | ||
212 | // x32, no-bcmath | |
213 | $p = max(0, strlen($v) - 13); | |
214 | $lo = (float)substr($v, $p); | |
215 | $hi = (float)substr($v, 0, $p); | |
216 | ||
217 | $m = $lo + $hi*1316134912.0; | |
218 | $q = floor($m / 4294967296.0); | |
219 | $l = $m - ($q * 4294967296.0); | |
220 | $h = $hi*2328.0 + $q; | |
221 | ||
222 | return pack ( "NN", $h, $l ); | |
223 | } | |
224 | ||
225 | // unpack 64-bit unsigned | |
226 | function sphUnpackU64 ( $v ) | |
227 | { | |
228 | list ( $hi, $lo ) = array_values ( unpack ( "N*N*", $v ) ); | |
229 | ||
230 | if ( PHP_INT_SIZE>=8 ) | |
231 | { | |
232 | if ( $hi<0 ) $hi += (1<<32); // because php 5.2.2 to 5.2.5 is totally fucked up again | |
233 | if ( $lo<0 ) $lo += (1<<32); | |
234 | ||
235 | // x64, int | |
236 | if ( $hi<=2147483647 ) | |
237 | return ($hi<<32) + $lo; | |
238 | ||
239 | // x64, bcmath | |
240 | if ( function_exists("bcmul") ) | |
241 | return bcadd ( $lo, bcmul ( $hi, "4294967296" ) ); | |
242 | ||
243 | // x64, no-bcmath | |
244 | $C = 100000; | |
245 | $h = ((int)($hi / $C) << 32) + (int)($lo / $C); | |
246 | $l = (($hi % $C) << 32) + ($lo % $C); | |
247 | if ( $l>$C ) | |
248 | { | |
249 | $h += (int)($l / $C); | |
250 | $l = $l % $C; | |
251 | } | |
252 | ||
253 | if ( $h==0 ) | |
254 | return $l; | |
255 | return sprintf ( "%d%05d", $h, $l ); | |
256 | } | |
257 | ||
258 | // x32, int | |
259 | if ( $hi==0 ) | |
260 | { | |
261 | if ( $lo>0 ) | |
262 | return $lo; | |
263 | return sprintf ( "%u", $lo ); | |
264 | } | |
265 | ||
266 | $hi = sprintf ( "%u", $hi ); | |
267 | $lo = sprintf ( "%u", $lo ); | |
268 | ||
269 | // x32, bcmath | |
270 | if ( function_exists("bcmul") ) | |
271 | return bcadd ( $lo, bcmul ( $hi, "4294967296" ) ); | |
272 | ||
273 | // x32, no-bcmath | |
274 | $hi = (float)$hi; | |
275 | $lo = (float)$lo; | |
276 | ||
277 | $q = floor($hi/10000000.0); | |
278 | $r = $hi - $q*10000000.0; | |
279 | $m = $lo + $r*4967296.0; | |
280 | $mq = floor($m/10000000.0); | |
281 | $l = $m - $mq*10000000.0; | |
282 | $h = $q*4294967296.0 + $r*429.0 + $mq; | |
283 | ||
284 | $h = sprintf ( "%.0f", $h ); | |
285 | $l = sprintf ( "%07.0f", $l ); | |
286 | if ( $h=="0" ) | |
287 | return sprintf( "%.0f", (float)$l ); | |
288 | return $h . $l; | |
289 | } | |
290 | ||
291 | // unpack 64-bit signed | |
292 | function sphUnpackI64 ( $v ) | |
293 | { | |
294 | list ( $hi, $lo ) = array_values ( unpack ( "N*N*", $v ) ); | |
295 | ||
296 | // x64 | |
297 | if ( PHP_INT_SIZE>=8 ) | |
298 | { | |
299 | if ( $hi<0 ) $hi += (1<<32); // because php 5.2.2 to 5.2.5 is totally fucked up again | |
300 | if ( $lo<0 ) $lo += (1<<32); | |
301 | ||
302 | return ($hi<<32) + $lo; | |
303 | } | |
304 | ||
305 | // x32, int | |
306 | if ( $hi==0 ) | |
307 | { | |
308 | if ( $lo>0 ) | |
309 | return $lo; | |
310 | return sprintf ( "%u", $lo ); | |
311 | } | |
312 | // x32, int | |
313 | elseif ( $hi==-1 ) | |
314 | { | |
315 | if ( $lo<0 ) | |
316 | return $lo; | |
317 | return sprintf ( "%.0f", $lo - 4294967296.0 ); | |
318 | } | |
319 | ||
320 | $neg = ""; | |
321 | $c = 0; | |
322 | if ( $hi<0 ) | |
323 | { | |
324 | $hi = ~$hi; | |
325 | $lo = ~$lo; | |
326 | $c = 1; | |
327 | $neg = "-"; | |
328 | } | |
329 | ||
330 | $hi = sprintf ( "%u", $hi ); | |
331 | $lo = sprintf ( "%u", $lo ); | |
332 | ||
333 | // x32, bcmath | |
334 | if ( function_exists("bcmul") ) | |
335 | return $neg . bcadd ( bcadd ( $lo, bcmul ( $hi, "4294967296" ) ), $c ); | |
336 | ||
337 | // x32, no-bcmath | |
338 | $hi = (float)$hi; | |
339 | $lo = (float)$lo; | |
340 | ||
341 | $q = floor($hi/10000000.0); | |
342 | $r = $hi - $q*10000000.0; | |
343 | $m = $lo + $r*4967296.0; | |
344 | $mq = floor($m/10000000.0); | |
345 | $l = $m - $mq*10000000.0 + $c; | |
346 | $h = $q*4294967296.0 + $r*429.0 + $mq; | |
347 | if ( $l==10000000 ) | |
348 | { | |
349 | $l = 0; | |
350 | $h += 1; | |
351 | } | |
352 | ||
353 | $h = sprintf ( "%.0f", $h ); | |
354 | $l = sprintf ( "%07.0f", $l ); | |
355 | if ( $h=="0" ) | |
356 | return $neg . sprintf( "%.0f", (float)$l ); | |
357 | return $neg . $h . $l; | |
358 | } | |
359 | ||
360 | ||
361 | function sphFixUint ( $value ) | |
362 | { | |
363 | if ( PHP_INT_SIZE>=8 ) | |
364 | { | |
365 | // x64 route, workaround broken unpack() in 5.2.2+ | |
366 | if ( $value<0 ) $value += (1<<32); | |
367 | return $value; | |
368 | } | |
369 | else | |
370 | { | |
371 | // x32 route, workaround php signed/unsigned braindamage | |
372 | return sprintf ( "%u", $value ); | |
373 | } | |
374 | } | |
375 | ||
376 | ||
377 | /// sphinx searchd client class | |
378 | class SphinxClient | |
379 | { | |
380 | var $_host; ///< searchd host (default is "localhost") | |
381 | var $_port; ///< searchd port (default is 9312) | |
382 | var $_offset; ///< how many records to seek from result-set start (default is 0) | |
383 | var $_limit; ///< how many records to return from result-set starting at offset (default is 20) | |
384 | var $_mode; ///< query matching mode (default is SPH_MATCH_ALL) | |
385 | var $_weights; ///< per-field weights (default is 1 for all fields) | |
386 | var $_sort; ///< match sorting mode (default is SPH_SORT_RELEVANCE) | |
387 | var $_sortby; ///< attribute to sort by (defualt is "") | |
388 | var $_min_id; ///< min ID to match (default is 0, which means no limit) | |
389 | var $_max_id; ///< max ID to match (default is 0, which means no limit) | |
390 | var $_filters; ///< search filters | |
391 | var $_groupby; ///< group-by attribute name | |
392 | var $_groupfunc; ///< group-by function (to pre-process group-by attribute value with) | |
393 | var $_groupsort; ///< group-by sorting clause (to sort groups in result set with) | |
394 | var $_groupdistinct;///< group-by count-distinct attribute | |
395 | var $_maxmatches; ///< max matches to retrieve | |
396 | var $_cutoff; ///< cutoff to stop searching at (default is 0) | |
397 | var $_retrycount; ///< distributed retries count | |
398 | var $_retrydelay; ///< distributed retries delay | |
399 | var $_anchor; ///< geographical anchor point | |
400 | var $_indexweights; ///< per-index weights | |
401 | var $_ranker; ///< ranking mode (default is SPH_RANK_PROXIMITY_BM25) | |
402 | var $_maxquerytime; ///< max query time, milliseconds (default is 0, do not limit) | |
403 | var $_fieldweights; ///< per-field-name weights | |
404 | var $_overrides; ///< per-query attribute values overrides | |
405 | var $_select; ///< select-list (attributes or expressions, with optional aliases) | |
406 | ||
407 | var $_error; ///< last error message | |
408 | var $_warning; ///< last warning message | |
409 | var $_connerror; ///< connection error vs remote error flag | |
410 | ||
411 | var $_reqs; ///< requests array for multi-query | |
412 | var $_mbenc; ///< stored mbstring encoding | |
413 | var $_arrayresult; ///< whether $result["matches"] should be a hash or an array | |
414 | var $_timeout; ///< connect timeout | |
415 | ||
416 | ///////////////////////////////////////////////////////////////////////////// | |
417 | // common stuff | |
418 | ///////////////////////////////////////////////////////////////////////////// | |
419 | ||
420 | /// create a new client object and fill defaults | |
421 | function SphinxClient () | |
422 | { | |
423 | // per-client-object settings | |
424 | $this->_host = "localhost"; | |
425 | $this->_port = 9312; | |
426 | $this->_path = false; | |
427 | $this->_socket = false; | |
428 | ||
429 | // per-query settings | |
430 | $this->_offset = 0; | |
431 | $this->_limit = 20; | |
432 | $this->_mode = SPH_MATCH_ALL; | |
433 | $this->_weights = array (); | |
434 | $this->_sort = SPH_SORT_RELEVANCE; | |
435 | $this->_sortby = ""; | |
436 | $this->_min_id = 0; | |
437 | $this->_max_id = 0; | |
438 | $this->_filters = array (); | |
439 | $this->_groupby = ""; | |
440 | $this->_groupfunc = SPH_GROUPBY_DAY; | |
441 | $this->_groupsort = "@group desc"; | |
442 | $this->_groupdistinct= ""; | |
443 | $this->_maxmatches = 1000; | |
444 | $this->_cutoff = 0; | |
445 | $this->_retrycount = 0; | |
446 | $this->_retrydelay = 0; | |
447 | $this->_anchor = array (); | |
448 | $this->_indexweights= array (); | |
449 | $this->_ranker = SPH_RANK_PROXIMITY_BM25; | |
450 | $this->_maxquerytime= 0; | |
451 | $this->_fieldweights= array(); | |
452 | $this->_overrides = array(); | |
453 | $this->_select = "*"; | |
454 | ||
455 | $this->_error = ""; // per-reply fields (for single-query case) | |
456 | $this->_warning = ""; | |
457 | $this->_connerror = false; | |
458 | ||
459 | $this->_reqs = array (); // requests storage (for multi-query case) | |
460 | $this->_mbenc = ""; | |
461 | $this->_arrayresult = false; | |
462 | $this->_timeout = 0; | |
463 | } | |
464 | ||
465 | function __destruct() | |
466 | { | |
467 | if ( $this->_socket !== false ) | |
468 | fclose ( $this->_socket ); | |
469 | } | |
470 | ||
471 | /// get last error message (string) | |
472 | function GetLastError () | |
473 | { | |
474 | return $this->_error; | |
475 | } | |
476 | ||
477 | /// get last warning message (string) | |
478 | function GetLastWarning () | |
479 | { | |
480 | return $this->_warning; | |
481 | } | |
482 | ||
483 | /// get last error flag (to tell network connection errors from searchd errors or broken responses) | |
484 | function IsConnectError() | |
485 | { | |
486 | return $this->_connerror; | |
487 | } | |
488 | ||
489 | /// set searchd host name (string) and port (integer) | |
490 | function SetServer ( $host, $port = 0 ) | |
491 | { | |
492 | assert ( is_string($host) ); | |
493 | if ( $host[0] == '/') | |
494 | { | |
495 | $this->_path = 'unix://' . $host; | |
496 | return; | |
497 | } | |
498 | if ( substr ( $host, 0, 7 )=="unix://" ) | |
499 | { | |
500 | $this->_path = $host; | |
501 | return; | |
502 | } | |
503 | ||
504 | assert ( is_int($port) ); | |
505 | $this->_host = $host; | |
506 | $this->_port = $port; | |
507 | $this->_path = ''; | |
508 | ||
509 | } | |
510 | ||
511 | /// set server connection timeout (0 to remove) | |
512 | function SetConnectTimeout ( $timeout ) | |
513 | { | |
514 | assert ( is_numeric($timeout) ); | |
515 | $this->_timeout = $timeout; | |
516 | } | |
517 | ||
518 | ||
519 | function _Send ( $handle, $data, $length ) | |
520 | { | |
521 | if ( feof($handle) || fwrite ( $handle, $data, $length ) !== $length ) | |
522 | { | |
523 | $this->_error = 'connection unexpectedly closed (timed out?)'; | |
524 | $this->_connerror = true; | |
525 | return false; | |
526 | } | |
527 | return true; | |
528 | } | |
529 | ||
530 | ///////////////////////////////////////////////////////////////////////////// | |
531 | ||
532 | /// enter mbstring workaround mode | |
533 | function _MBPush () | |
534 | { | |
535 | $this->_mbenc = ""; | |
536 | if ( ini_get ( "mbstring.func_overload" ) & 2 ) | |
537 | { | |
538 | $this->_mbenc = mb_internal_encoding(); | |
539 | mb_internal_encoding ( "latin1" ); | |
540 | } | |
541 | } | |
542 | ||
543 | /// leave mbstring workaround mode | |
544 | function _MBPop () | |
545 | { | |
546 | if ( $this->_mbenc ) | |
547 | mb_internal_encoding ( $this->_mbenc ); | |
548 | } | |
549 | ||
550 | /// connect to searchd server | |
551 | function _Connect () | |
552 | { | |
553 | if ( $this->_socket!==false ) | |
554 | { | |
555 | // we are in persistent connection mode, so we have a socket | |
556 | // however, need to check whether it's still alive | |
557 | if ( !@feof ( $this->_socket ) ) | |
558 | return $this->_socket; | |
559 | ||
560 | // force reopen | |
561 | $this->_socket = false; | |
562 | } | |
563 | ||
564 | $errno = 0; | |
565 | $errstr = ""; | |
566 | $this->_connerror = false; | |
567 | ||
568 | if ( $this->_path ) | |
569 | { | |
570 | $host = $this->_path; | |
571 | $port = 0; | |
572 | } | |
573 | else | |
574 | { | |
575 | $host = $this->_host; | |
576 | $port = $this->_port; | |
577 | } | |
578 | ||
579 | if ( $this->_timeout<=0 ) | |
580 | $fp = @fsockopen ( $host, $port, $errno, $errstr ); | |
581 | else | |
582 | $fp = @fsockopen ( $host, $port, $errno, $errstr, $this->_timeout ); | |
583 | ||
584 | if ( !$fp ) | |
585 | { | |
586 | if ( $this->_path ) | |
587 | $location = $this->_path; | |
588 | else | |
589 | $location = "{$this->_host}:{$this->_port}"; | |
590 | ||
591 | $errstr = trim ( $errstr ); | |
592 | $this->_error = "connection to $location failed (errno=$errno, msg=$errstr)"; | |
593 | $this->_connerror = true; | |
594 | return false; | |
595 | } | |
596 | ||
597 | // send my version | |
598 | // this is a subtle part. we must do it before (!) reading back from searchd. | |
599 | // because otherwise under some conditions (reported on FreeBSD for instance) | |
600 | // TCP stack could throttle write-write-read pattern because of Nagle. | |
601 | if ( !$this->_Send ( $fp, pack ( "N", 1 ), 4 ) ) | |
602 | { | |
603 | fclose ( $fp ); | |
604 | $this->_error = "failed to send client protocol version"; | |
605 | return false; | |
606 | } | |
607 | ||
608 | // check version | |
609 | list(,$v) = unpack ( "N*", fread ( $fp, 4 ) ); | |
610 | $v = (int)$v; | |
611 | if ( $v<1 ) | |
612 | { | |
613 | fclose ( $fp ); | |
614 | $this->_error = "expected searchd protocol version 1+, got version '$v'"; | |
615 | return false; | |
616 | } | |
617 | ||
618 | return $fp; | |
619 | } | |
620 | ||
621 | /// get and check response packet from searchd server | |
622 | function _GetResponse ( $fp, $client_ver ) | |
623 | { | |
624 | $response = ""; | |
625 | $len = 0; | |
626 | ||
627 | $header = fread ( $fp, 8 ); | |
628 | if ( strlen($header)==8 ) | |
629 | { | |
630 | list ( $status, $ver, $len ) = array_values ( unpack ( "n2a/Nb", $header ) ); | |
631 | $left = $len; | |
632 | while ( $left>0 && !feof($fp) ) | |
633 | { | |
634 | $chunk = fread ( $fp, $left ); | |
635 | if ( $chunk ) | |
636 | { | |
637 | $response .= $chunk; | |
638 | $left -= strlen($chunk); | |
639 | } | |
640 | } | |
641 | } | |
642 | if ( $this->_socket === false ) | |
643 | fclose ( $fp ); | |
644 | ||
645 | // check response | |
646 | $read = strlen ( $response ); | |
647 | if ( !$response || $read!=$len ) | |
648 | { | |
649 | $this->_error = $len | |
650 | ? "failed to read searchd response (status=$status, ver=$ver, len=$len, read=$read)" | |
651 | : "received zero-sized searchd response"; | |
652 | return false; | |
653 | } | |
654 | ||
655 | // check status | |
656 | if ( $status==SEARCHD_WARNING ) | |
657 | { | |
658 | list(,$wlen) = unpack ( "N*", substr ( $response, 0, 4 ) ); | |
659 | $this->_warning = substr ( $response, 4, $wlen ); | |
660 | return substr ( $response, 4+$wlen ); | |
661 | } | |
662 | if ( $status==SEARCHD_ERROR ) | |
663 | { | |
664 | $this->_error = "searchd error: " . substr ( $response, 4 ); | |
665 | return false; | |
666 | } | |
667 | if ( $status==SEARCHD_RETRY ) | |
668 | { | |
669 | $this->_error = "temporary searchd error: " . substr ( $response, 4 ); | |
670 | return false; | |
671 | } | |
672 | if ( $status!=SEARCHD_OK ) | |
673 | { | |
674 | $this->_error = "unknown status code '$status'"; | |
675 | return false; | |
676 | } | |
677 | ||
678 | // check version | |
679 | if ( $ver<$client_ver ) | |
680 | { | |
681 | $this->_warning = sprintf ( "searchd command v.%d.%d older than client's v.%d.%d, some options might not work", | |
682 | $ver>>8, $ver&0xff, $client_ver>>8, $client_ver&0xff ); | |
683 | } | |
684 | ||
685 | return $response; | |
686 | } | |
687 | ||
688 | ///////////////////////////////////////////////////////////////////////////// | |
689 | // searching | |
690 | ///////////////////////////////////////////////////////////////////////////// | |
691 | ||
692 | /// set offset and count into result set, | |
693 | /// and optionally set max-matches and cutoff limits | |
694 | function SetLimits ( $offset, $limit, $max=0, $cutoff=0 ) | |
695 | { | |
696 | assert ( is_int($offset) ); | |
697 | assert ( is_int($limit) ); | |
698 | assert ( $offset>=0 ); | |
699 | assert ( $limit>0 ); | |
700 | assert ( $max>=0 ); | |
701 | $this->_offset = $offset; | |
702 | $this->_limit = $limit; | |
703 | if ( $max>0 ) | |
704 | $this->_maxmatches = $max; | |
705 | if ( $cutoff>0 ) | |
706 | $this->_cutoff = $cutoff; | |
707 | } | |
708 | ||
709 | /// set maximum query time, in milliseconds, per-index | |
710 | /// integer, 0 means "do not limit" | |
711 | function SetMaxQueryTime ( $max ) | |
712 | { | |
713 | assert ( is_int($max) ); | |
714 | assert ( $max>=0 ); | |
715 | $this->_maxquerytime = $max; | |
716 | } | |
717 | ||
718 | /// set matching mode | |
719 | function SetMatchMode ( $mode ) | |
720 | { | |
721 | assert ( $mode==SPH_MATCH_ALL | |
722 | || $mode==SPH_MATCH_ANY | |
723 | || $mode==SPH_MATCH_PHRASE | |
724 | || $mode==SPH_MATCH_BOOLEAN | |
725 | || $mode==SPH_MATCH_EXTENDED | |
726 | || $mode==SPH_MATCH_FULLSCAN | |
727 | || $mode==SPH_MATCH_EXTENDED2 ); | |
728 | $this->_mode = $mode; | |
729 | } | |
730 | ||
731 | /// set ranking mode | |
732 | function SetRankingMode ( $ranker ) | |
733 | { | |
734 | assert ( $ranker==SPH_RANK_PROXIMITY_BM25 | |
735 | || $ranker==SPH_RANK_BM25 | |
736 | || $ranker==SPH_RANK_NONE | |
737 | || $ranker==SPH_RANK_WORDCOUNT | |
738 | || $ranker==SPH_RANK_PROXIMITY ); | |
739 | $this->_ranker = $ranker; | |
740 | } | |
741 | ||
742 | /// set matches sorting mode | |
743 | function SetSortMode ( $mode, $sortby="" ) | |
744 | { | |
745 | assert ( | |
746 | $mode==SPH_SORT_RELEVANCE || | |
747 | $mode==SPH_SORT_ATTR_DESC || | |
748 | $mode==SPH_SORT_ATTR_ASC || | |
749 | $mode==SPH_SORT_TIME_SEGMENTS || | |
750 | $mode==SPH_SORT_EXTENDED || | |
751 | $mode==SPH_SORT_EXPR ); | |
752 | assert ( is_string($sortby) ); | |
753 | assert ( $mode==SPH_SORT_RELEVANCE || strlen($sortby)>0 ); | |
754 | ||
755 | $this->_sort = $mode; | |
756 | $this->_sortby = $sortby; | |
757 | } | |
758 | ||
759 | /// bind per-field weights by order | |
760 | /// DEPRECATED; use SetFieldWeights() instead | |
761 | function SetWeights ( $weights ) | |
762 | { | |
763 | assert ( is_array($weights) ); | |
764 | foreach ( $weights as $weight ) | |
765 | assert ( is_int($weight) ); | |
766 | ||
767 | $this->_weights = $weights; | |
768 | } | |
769 | ||
770 | /// bind per-field weights by name | |
771 | function SetFieldWeights ( $weights ) | |
772 | { | |
773 | assert ( is_array($weights) ); | |
774 | foreach ( $weights as $name=>$weight ) | |
775 | { | |
776 | assert ( is_string($name) ); | |
777 | assert ( is_int($weight) ); | |
778 | } | |
779 | $this->_fieldweights = $weights; | |
780 | } | |
781 | ||
782 | /// bind per-index weights by name | |
783 | function SetIndexWeights ( $weights ) | |
784 | { | |
785 | assert ( is_array($weights) ); | |
786 | foreach ( $weights as $index=>$weight ) | |
787 | { | |
788 | assert ( is_string($index) ); | |
789 | assert ( is_int($weight) ); | |
790 | } | |
791 | $this->_indexweights = $weights; | |
792 | } | |
793 | ||
794 | /// set IDs range to match | |
795 | /// only match records if document ID is beetwen $min and $max (inclusive) | |
796 | function SetIDRange ( $min, $max ) | |
797 | { | |
798 | assert ( is_numeric($min) ); | |
799 | assert ( is_numeric($max) ); | |
800 | assert ( $min<=$max ); | |
801 | $this->_min_id = $min; | |
802 | $this->_max_id = $max; | |
803 | } | |
804 | ||
805 | /// set values set filter | |
806 | /// only match records where $attribute value is in given set | |
807 | function SetFilter ( $attribute, $values, $exclude=false ) | |
808 | { | |
809 | assert ( is_string($attribute) ); | |
810 | assert ( is_array($values) ); | |
811 | assert ( count($values) ); | |
812 | ||
813 | if ( is_array($values) && count($values) ) | |
814 | { | |
815 | foreach ( $values as $value ) | |
816 | assert ( is_numeric($value) ); | |
817 | ||
818 | $this->_filters[] = array ( "type"=>SPH_FILTER_VALUES, "attr"=>$attribute, "exclude"=>$exclude, "values"=>$values ); | |
819 | } | |
820 | } | |
821 | ||
822 | /// set range filter | |
823 | /// only match records if $attribute value is beetwen $min and $max (inclusive) | |
824 | function SetFilterRange ( $attribute, $min, $max, $exclude=false ) | |
825 | { | |
826 | assert ( is_string($attribute) ); | |
827 | assert ( is_numeric($min) ); | |
828 | assert ( is_numeric($max) ); | |
829 | assert ( $min<=$max ); | |
830 | ||
831 | $this->_filters[] = array ( "type"=>SPH_FILTER_RANGE, "attr"=>$attribute, "exclude"=>$exclude, "min"=>$min, "max"=>$max ); | |
832 | } | |
833 | ||
834 | /// set float range filter | |
835 | /// only match records if $attribute value is beetwen $min and $max (inclusive) | |
836 | function SetFilterFloatRange ( $attribute, $min, $max, $exclude=false ) | |
837 | { | |
838 | assert ( is_string($attribute) ); | |
839 | assert ( is_float($min) ); | |
840 | assert ( is_float($max) ); | |
841 | assert ( $min<=$max ); | |
842 | ||
843 | $this->_filters[] = array ( "type"=>SPH_FILTER_FLOATRANGE, "attr"=>$attribute, "exclude"=>$exclude, "min"=>$min, "max"=>$max ); | |
844 | } | |
845 | ||
846 | /// setup anchor point for geosphere distance calculations | |
847 | /// required to use @geodist in filters and sorting | |
848 | /// latitude and longitude must be in radians | |
849 | function SetGeoAnchor ( $attrlat, $attrlong, $lat, $long ) | |
850 | { | |
851 | assert ( is_string($attrlat) ); | |
852 | assert ( is_string($attrlong) ); | |
853 | assert ( is_float($lat) ); | |
854 | assert ( is_float($long) ); | |
855 | ||
856 | $this->_anchor = array ( "attrlat"=>$attrlat, "attrlong"=>$attrlong, "lat"=>$lat, "long"=>$long ); | |
857 | } | |
858 | ||
859 | /// set grouping attribute and function | |
860 | function SetGroupBy ( $attribute, $func, $groupsort="@group desc" ) | |
861 | { | |
862 | assert ( is_string($attribute) ); | |
863 | assert ( is_string($groupsort) ); | |
864 | assert ( $func==SPH_GROUPBY_DAY | |
865 | || $func==SPH_GROUPBY_WEEK | |
866 | || $func==SPH_GROUPBY_MONTH | |
867 | || $func==SPH_GROUPBY_YEAR | |
868 | || $func==SPH_GROUPBY_ATTR | |
869 | || $func==SPH_GROUPBY_ATTRPAIR ); | |
870 | ||
871 | $this->_groupby = $attribute; | |
872 | $this->_groupfunc = $func; | |
873 | $this->_groupsort = $groupsort; | |
874 | } | |
875 | ||
876 | /// set count-distinct attribute for group-by queries | |
877 | function SetGroupDistinct ( $attribute ) | |
878 | { | |
879 | assert ( is_string($attribute) ); | |
880 | $this->_groupdistinct = $attribute; | |
881 | } | |
882 | ||
883 | /// set distributed retries count and delay | |
884 | function SetRetries ( $count, $delay=0 ) | |
885 | { | |
886 | assert ( is_int($count) && $count>=0 ); | |
887 | assert ( is_int($delay) && $delay>=0 ); | |
888 | $this->_retrycount = $count; | |
889 | $this->_retrydelay = $delay; | |
890 | } | |
891 | ||
892 | /// set result set format (hash or array; hash by default) | |
893 | /// PHP specific; needed for group-by-MVA result sets that may contain duplicate IDs | |
894 | function SetArrayResult ( $arrayresult ) | |
895 | { | |
896 | assert ( is_bool($arrayresult) ); | |
897 | $this->_arrayresult = $arrayresult; | |
898 | } | |
899 | ||
900 | /// set attribute values override | |
901 | /// there can be only one override per attribute | |
902 | /// $values must be a hash that maps document IDs to attribute values | |
903 | function SetOverride ( $attrname, $attrtype, $values ) | |
904 | { | |
905 | assert ( is_string ( $attrname ) ); | |
906 | assert ( in_array ( $attrtype, array ( SPH_ATTR_INTEGER, SPH_ATTR_TIMESTAMP, SPH_ATTR_BOOL, SPH_ATTR_FLOAT, SPH_ATTR_BIGINT ) ) ); | |
907 | assert ( is_array ( $values ) ); | |
908 | ||
909 | $this->_overrides[$attrname] = array ( "attr"=>$attrname, "type"=>$attrtype, "values"=>$values ); | |
910 | } | |
911 | ||
912 | /// set select-list (attributes or expressions), SQL-like syntax | |
913 | function SetSelect ( $select ) | |
914 | { | |
915 | assert ( is_string ( $select ) ); | |
916 | $this->_select = $select; | |
917 | } | |
918 | ||
919 | ////////////////////////////////////////////////////////////////////////////// | |
920 | ||
921 | /// clear all filters (for multi-queries) | |
922 | function ResetFilters () | |
923 | { | |
924 | $this->_filters = array(); | |
925 | $this->_anchor = array(); | |
926 | } | |
927 | ||
928 | /// clear groupby settings (for multi-queries) | |
929 | function ResetGroupBy () | |
930 | { | |
931 | $this->_groupby = ""; | |
932 | $this->_groupfunc = SPH_GROUPBY_DAY; | |
933 | $this->_groupsort = "@group desc"; | |
934 | $this->_groupdistinct= ""; | |
935 | } | |
936 | ||
937 | /// clear all attribute value overrides (for multi-queries) | |
938 | function ResetOverrides () | |
939 | { | |
940 | $this->_overrides = array (); | |
941 | } | |
942 | ||
943 | ////////////////////////////////////////////////////////////////////////////// | |
944 | ||
945 | /// connect to searchd server, run given search query through given indexes, | |
946 | /// and return the search results | |
947 | function Query ( $query, $index="*", $comment="" ) | |
948 | { | |
949 | assert ( empty($this->_reqs) ); | |
950 | ||
951 | $this->AddQuery ( $query, $index, $comment ); | |
952 | $results = $this->RunQueries (); | |
953 | $this->_reqs = array (); // just in case it failed too early | |
954 | ||
955 | if ( !is_array($results) ) | |
956 | return false; // probably network error; error message should be already filled | |
957 | ||
958 | $this->_error = $results[0]["error"]; | |
959 | $this->_warning = $results[0]["warning"]; | |
960 | if ( $results[0]["status"]==SEARCHD_ERROR ) | |
961 | return false; | |
962 | else | |
963 | return $results[0]; | |
964 | } | |
965 | ||
966 | /// helper to pack floats in network byte order | |
967 | function _PackFloat ( $f ) | |
968 | { | |
969 | $t1 = pack ( "f", $f ); // machine order | |
970 | list(,$t2) = unpack ( "L*", $t1 ); // int in machine order | |
971 | return pack ( "N", $t2 ); | |
972 | } | |
973 | ||
974 | /// add query to multi-query batch | |
975 | /// returns index into results array from RunQueries() call | |
976 | function AddQuery ( $query, $index="*", $comment="" ) | |
977 | { | |
978 | // mbstring workaround | |
979 | $this->_MBPush (); | |
980 | ||
981 | // build request | |
982 | $req = pack ( "NNNNN", $this->_offset, $this->_limit, $this->_mode, $this->_ranker, $this->_sort ); // mode and limits | |
983 | $req .= pack ( "N", strlen($this->_sortby) ) . $this->_sortby; | |
984 | $req .= pack ( "N", strlen($query) ) . $query; // query itself | |
985 | $req .= pack ( "N", count($this->_weights) ); // weights | |
986 | foreach ( $this->_weights as $weight ) | |
987 | $req .= pack ( "N", (int)$weight ); | |
988 | $req .= pack ( "N", strlen($index) ) . $index; // indexes | |
989 | $req .= pack ( "N", 1 ); // id64 range marker | |
990 | $req .= sphPackU64 ( $this->_min_id ) . sphPackU64 ( $this->_max_id ); // id64 range | |
991 | ||
992 | // filters | |
993 | $req .= pack ( "N", count($this->_filters) ); | |
994 | foreach ( $this->_filters as $filter ) | |
995 | { | |
996 | $req .= pack ( "N", strlen($filter["attr"]) ) . $filter["attr"]; | |
997 | $req .= pack ( "N", $filter["type"] ); | |
998 | switch ( $filter["type"] ) | |
999 | { | |
1000 | case SPH_FILTER_VALUES: | |
1001 | $req .= pack ( "N", count($filter["values"]) ); | |
1002 | foreach ( $filter["values"] as $value ) | |
1003 | $req .= sphPackI64 ( $value ); | |
1004 | break; | |
1005 | ||
1006 | case SPH_FILTER_RANGE: | |
1007 | $req .= sphPackI64 ( $filter["min"] ) . sphPackI64 ( $filter["max"] ); | |
1008 | break; | |
1009 | ||
1010 | case SPH_FILTER_FLOATRANGE: | |
1011 | $req .= $this->_PackFloat ( $filter["min"] ) . $this->_PackFloat ( $filter["max"] ); | |
1012 | break; | |
1013 | ||
1014 | default: | |
1015 | assert ( 0 && "internal error: unhandled filter type" ); | |
1016 | } | |
1017 | $req .= pack ( "N", $filter["exclude"] ); | |
1018 | } | |
1019 | ||
1020 | // group-by clause, max-matches count, group-sort clause, cutoff count | |
1021 | $req .= pack ( "NN", $this->_groupfunc, strlen($this->_groupby) ) . $this->_groupby; | |
1022 | $req .= pack ( "N", $this->_maxmatches ); | |
1023 | $req .= pack ( "N", strlen($this->_groupsort) ) . $this->_groupsort; | |
1024 | $req .= pack ( "NNN", $this->_cutoff, $this->_retrycount, $this->_retrydelay ); | |
1025 | $req .= pack ( "N", strlen($this->_groupdistinct) ) . $this->_groupdistinct; | |
1026 | ||
1027 | // anchor point | |
1028 | if ( empty($this->_anchor) ) | |
1029 | { | |
1030 | $req .= pack ( "N", 0 ); | |
1031 | } else | |
1032 | { | |
1033 | $a =& $this->_anchor; | |
1034 | $req .= pack ( "N", 1 ); | |
1035 | $req .= pack ( "N", strlen($a["attrlat"]) ) . $a["attrlat"]; | |
1036 | $req .= pack ( "N", strlen($a["attrlong"]) ) . $a["attrlong"]; | |
1037 | $req .= $this->_PackFloat ( $a["lat"] ) . $this->_PackFloat ( $a["long"] ); | |
1038 | } | |
1039 | ||
1040 | // per-index weights | |
1041 | $req .= pack ( "N", count($this->_indexweights) ); | |
1042 | foreach ( $this->_indexweights as $idx=>$weight ) | |
1043 | $req .= pack ( "N", strlen($idx) ) . $idx . pack ( "N", $weight ); | |
1044 | ||
1045 | // max query time | |
1046 | $req .= pack ( "N", $this->_maxquerytime ); | |
1047 | ||
1048 | // per-field weights | |
1049 | $req .= pack ( "N", count($this->_fieldweights) ); | |
1050 | foreach ( $this->_fieldweights as $field=>$weight ) | |
1051 | $req .= pack ( "N", strlen($field) ) . $field . pack ( "N", $weight ); | |
1052 | ||
1053 | // comment | |
1054 | $req .= pack ( "N", strlen($comment) ) . $comment; | |
1055 | ||
1056 | // attribute overrides | |
1057 | $req .= pack ( "N", count($this->_overrides) ); | |
1058 | foreach ( $this->_overrides as $key => $entry ) | |
1059 | { | |
1060 | $req .= pack ( "N", strlen($entry["attr"]) ) . $entry["attr"]; | |
1061 | $req .= pack ( "NN", $entry["type"], count($entry["values"]) ); | |
1062 | foreach ( $entry["values"] as $id=>$val ) | |
1063 | { | |
1064 | assert ( is_numeric($id) ); | |
1065 | assert ( is_numeric($val) ); | |
1066 | ||
1067 | $req .= sphPackU64 ( $id ); | |
1068 | switch ( $entry["type"] ) | |
1069 | { | |
1070 | case SPH_ATTR_FLOAT: $req .= $this->_PackFloat ( $val ); break; | |
1071 | case SPH_ATTR_BIGINT: $req .= sphPackI64 ( $val ); break; | |
1072 | default: $req .= pack ( "N", $val ); break; | |
1073 | } | |
1074 | } | |
1075 | } | |
1076 | ||
1077 | // select-list | |
1078 | $req .= pack ( "N", strlen($this->_select) ) . $this->_select; | |
1079 | ||
1080 | // mbstring workaround | |
1081 | $this->_MBPop (); | |
1082 | ||
1083 | // store request to requests array | |
1084 | $this->_reqs[] = $req; | |
1085 | return count($this->_reqs)-1; | |
1086 | } | |
1087 | ||
1088 | /// connect to searchd, run queries batch, and return an array of result sets | |
1089 | function RunQueries () | |
1090 | { | |
1091 | if ( empty($this->_reqs) ) | |
1092 | { | |
1093 | $this->_error = "no queries defined, issue AddQuery() first"; | |
1094 | return false; | |
1095 | } | |
1096 | ||
1097 | // mbstring workaround | |
1098 | $this->_MBPush (); | |
1099 | ||
1100 | if (!( $fp = $this->_Connect() )) | |
1101 | { | |
1102 | $this->_MBPop (); | |
1103 | return false; | |
1104 | } | |
1105 | ||
1106 | // send query, get response | |
1107 | $nreqs = count($this->_reqs); | |
1108 | $req = join ( "", $this->_reqs ); | |
1109 | $len = 4+strlen($req); | |
1110 | $req = pack ( "nnNN", SEARCHD_COMMAND_SEARCH, VER_COMMAND_SEARCH, $len, $nreqs ) . $req; // add header | |
1111 | ||
1112 | if ( !( $this->_Send ( $fp, $req, $len+8 ) ) || | |
1113 | !( $response = $this->_GetResponse ( $fp, VER_COMMAND_SEARCH ) ) ) | |
1114 | { | |
1115 | $this->_MBPop (); | |
1116 | return false; | |
1117 | } | |
1118 | ||
1119 | // query sent ok; we can reset reqs now | |
1120 | $this->_reqs = array (); | |
1121 | ||
1122 | // parse and return response | |
1123 | return $this->_ParseSearchResponse ( $response, $nreqs ); | |
1124 | } | |
1125 | ||
1126 | /// parse and return search query (or queries) response | |
1127 | function _ParseSearchResponse ( $response, $nreqs ) | |
1128 | { | |
1129 | $p = 0; // current position | |
1130 | $max = strlen($response); // max position for checks, to protect against broken responses | |
1131 | ||
1132 | $results = array (); | |
1133 | for ( $ires=0; $ires<$nreqs && $p<$max; $ires++ ) | |
1134 | { | |
1135 | $results[] = array(); | |
1136 | $result =& $results[$ires]; | |
1137 | ||
1138 | $result["error"] = ""; | |
1139 | $result["warning"] = ""; | |
1140 | ||
1141 | // extract status | |
1142 | list(,$status) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1143 | $result["status"] = $status; | |
1144 | if ( $status!=SEARCHD_OK ) | |
1145 | { | |
1146 | list(,$len) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1147 | $message = substr ( $response, $p, $len ); $p += $len; | |
1148 | ||
1149 | if ( $status==SEARCHD_WARNING ) | |
1150 | { | |
1151 | $result["warning"] = $message; | |
1152 | } else | |
1153 | { | |
1154 | $result["error"] = $message; | |
1155 | continue; | |
1156 | } | |
1157 | } | |
1158 | ||
1159 | // read schema | |
1160 | $fields = array (); | |
1161 | $attrs = array (); | |
1162 | ||
1163 | list(,$nfields) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1164 | while ( $nfields-->0 && $p<$max ) | |
1165 | { | |
1166 | list(,$len) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1167 | $fields[] = substr ( $response, $p, $len ); $p += $len; | |
1168 | } | |
1169 | $result["fields"] = $fields; | |
1170 | ||
1171 | list(,$nattrs) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1172 | while ( $nattrs-->0 && $p<$max ) | |
1173 | { | |
1174 | list(,$len) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1175 | $attr = substr ( $response, $p, $len ); $p += $len; | |
1176 | list(,$type) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1177 | $attrs[$attr] = $type; | |
1178 | } | |
1179 | $result["attrs"] = $attrs; | |
1180 | ||
1181 | // read match count | |
1182 | list(,$count) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1183 | list(,$id64) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1184 | ||
1185 | // read matches | |
1186 | $idx = -1; | |
1187 | while ( $count-->0 && $p<$max ) | |
1188 | { | |
1189 | // index into result array | |
1190 | $idx++; | |
1191 | ||
1192 | // parse document id and weight | |
1193 | if ( $id64 ) | |
1194 | { | |
1195 | $doc = sphUnpackU64 ( substr ( $response, $p, 8 ) ); $p += 8; | |
1196 | list(,$weight) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1197 | } | |
1198 | else | |
1199 | { | |
1200 | list ( $doc, $weight ) = array_values ( unpack ( "N*N*", | |
1201 | substr ( $response, $p, 8 ) ) ); | |
1202 | $p += 8; | |
1203 | $doc = sphFixUint($doc); | |
1204 | } | |
1205 | $weight = sprintf ( "%u", $weight ); | |
1206 | ||
1207 | // create match entry | |
1208 | if ( $this->_arrayresult ) | |
1209 | $result["matches"][$idx] = array ( "id"=>$doc, "weight"=>$weight ); | |
1210 | else | |
1211 | $result["matches"][$doc]["weight"] = $weight; | |
1212 | ||
1213 | // parse and create attributes | |
1214 | $attrvals = array (); | |
1215 | foreach ( $attrs as $attr=>$type ) | |
1216 | { | |
1217 | // handle 64bit ints | |
1218 | if ( $type==SPH_ATTR_BIGINT ) | |
1219 | { | |
1220 | $attrvals[$attr] = sphUnpackI64 ( substr ( $response, $p, 8 ) ); $p += 8; | |
1221 | continue; | |
1222 | } | |
1223 | ||
1224 | // handle floats | |
1225 | if ( $type==SPH_ATTR_FLOAT ) | |
1226 | { | |
1227 | list(,$uval) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1228 | list(,$fval) = unpack ( "f*", pack ( "L", $uval ) ); | |
1229 | $attrvals[$attr] = $fval; | |
1230 | continue; | |
1231 | } | |
1232 | ||
1233 | // handle everything else as unsigned ints | |
1234 | list(,$val) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1235 | if ( $type & SPH_ATTR_MULTI ) | |
1236 | { | |
1237 | $attrvals[$attr] = array (); | |
1238 | $nvalues = $val; | |
1239 | while ( $nvalues-->0 && $p<$max ) | |
1240 | { | |
1241 | list(,$val) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1242 | $attrvals[$attr][] = sphFixUint($val); | |
1243 | } | |
1244 | } else | |
1245 | { | |
1246 | $attrvals[$attr] = sphFixUint($val); | |
1247 | } | |
1248 | } | |
1249 | ||
1250 | if ( $this->_arrayresult ) | |
1251 | $result["matches"][$idx]["attrs"] = $attrvals; | |
1252 | else | |
1253 | $result["matches"][$doc]["attrs"] = $attrvals; | |
1254 | } | |
1255 | ||
1256 | list ( $total, $total_found, $msecs, $words ) = | |
1257 | array_values ( unpack ( "N*N*N*N*", substr ( $response, $p, 16 ) ) ); | |
1258 | $result["total"] = sprintf ( "%u", $total ); | |
1259 | $result["total_found"] = sprintf ( "%u", $total_found ); | |
1260 | $result["time"] = sprintf ( "%.3f", $msecs/1000 ); | |
1261 | $p += 16; | |
1262 | ||
1263 | while ( $words-->0 && $p<$max ) | |
1264 | { | |
1265 | list(,$len) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1266 | $word = substr ( $response, $p, $len ); $p += $len; | |
1267 | list ( $docs, $hits ) = array_values ( unpack ( "N*N*", substr ( $response, $p, 8 ) ) ); $p += 8; | |
1268 | $result["words"][$word] = array ( | |
1269 | "docs"=>sprintf ( "%u", $docs ), | |
1270 | "hits"=>sprintf ( "%u", $hits ) ); | |
1271 | } | |
1272 | } | |
1273 | ||
1274 | $this->_MBPop (); | |
1275 | return $results; | |
1276 | } | |
1277 | ||
1278 | ///////////////////////////////////////////////////////////////////////////// | |
1279 | // excerpts generation | |
1280 | ///////////////////////////////////////////////////////////////////////////// | |
1281 | ||
1282 | /// connect to searchd server, and generate exceprts (snippets) | |
1283 | /// of given documents for given query. returns false on failure, | |
1284 | /// an array of snippets on success | |
1285 | function BuildExcerpts ( $docs, $index, $words, $opts=array() ) | |
1286 | { | |
1287 | assert ( is_array($docs) ); | |
1288 | assert ( is_string($index) ); | |
1289 | assert ( is_string($words) ); | |
1290 | assert ( is_array($opts) ); | |
1291 | ||
1292 | $this->_MBPush (); | |
1293 | ||
1294 | if (!( $fp = $this->_Connect() )) | |
1295 | { | |
1296 | $this->_MBPop(); | |
1297 | return false; | |
1298 | } | |
1299 | ||
1300 | ///////////////// | |
1301 | // fixup options | |
1302 | ///////////////// | |
1303 | ||
1304 | if ( !isset($opts["before_match"]) ) $opts["before_match"] = "<b>"; | |
1305 | if ( !isset($opts["after_match"]) ) $opts["after_match"] = "</b>"; | |
1306 | if ( !isset($opts["chunk_separator"]) ) $opts["chunk_separator"] = " ... "; | |
1307 | if ( !isset($opts["limit"]) ) $opts["limit"] = 256; | |
1308 | if ( !isset($opts["around"]) ) $opts["around"] = 5; | |
1309 | if ( !isset($opts["exact_phrase"]) ) $opts["exact_phrase"] = false; | |
1310 | if ( !isset($opts["single_passage"]) ) $opts["single_passage"] = false; | |
1311 | if ( !isset($opts["use_boundaries"]) ) $opts["use_boundaries"] = false; | |
1312 | if ( !isset($opts["weight_order"]) ) $opts["weight_order"] = false; | |
1313 | ||
1314 | ///////////////// | |
1315 | // build request | |
1316 | ///////////////// | |
1317 | ||
1318 | // v.1.0 req | |
1319 | $flags = 1; // remove spaces | |
1320 | if ( $opts["exact_phrase"] ) $flags |= 2; | |
1321 | if ( $opts["single_passage"] ) $flags |= 4; | |
1322 | if ( $opts["use_boundaries"] ) $flags |= 8; | |
1323 | if ( $opts["weight_order"] ) $flags |= 16; | |
1324 | $req = pack ( "NN", 0, $flags ); // mode=0, flags=$flags | |
1325 | $req .= pack ( "N", strlen($index) ) . $index; // req index | |
1326 | $req .= pack ( "N", strlen($words) ) . $words; // req words | |
1327 | ||
1328 | // options | |
1329 | $req .= pack ( "N", strlen($opts["before_match"]) ) . $opts["before_match"]; | |
1330 | $req .= pack ( "N", strlen($opts["after_match"]) ) . $opts["after_match"]; | |
1331 | $req .= pack ( "N", strlen($opts["chunk_separator"]) ) . $opts["chunk_separator"]; | |
1332 | $req .= pack ( "N", (int)$opts["limit"] ); | |
1333 | $req .= pack ( "N", (int)$opts["around"] ); | |
1334 | ||
1335 | // documents | |
1336 | $req .= pack ( "N", count($docs) ); | |
1337 | foreach ( $docs as $doc ) | |
1338 | { | |
1339 | assert ( is_string($doc) ); | |
1340 | $req .= pack ( "N", strlen($doc) ) . $doc; | |
1341 | } | |
1342 | ||
1343 | //////////////////////////// | |
1344 | // send query, get response | |
1345 | //////////////////////////// | |
1346 | ||
1347 | $len = strlen($req); | |
1348 | $req = pack ( "nnN", SEARCHD_COMMAND_EXCERPT, VER_COMMAND_EXCERPT, $len ) . $req; // add header | |
1349 | if ( !( $this->_Send ( $fp, $req, $len+8 ) ) || | |
1350 | !( $response = $this->_GetResponse ( $fp, VER_COMMAND_EXCERPT ) ) ) | |
1351 | { | |
1352 | $this->_MBPop (); | |
1353 | return false; | |
1354 | } | |
1355 | ||
1356 | ////////////////// | |
1357 | // parse response | |
1358 | ////////////////// | |
1359 | ||
1360 | $pos = 0; | |
1361 | $res = array (); | |
1362 | $rlen = strlen($response); | |
1363 | for ( $i=0; $i<count($docs); $i++ ) | |
1364 | { | |
1365 | list(,$len) = unpack ( "N*", substr ( $response, $pos, 4 ) ); | |
1366 | $pos += 4; | |
1367 | ||
1368 | if ( $pos+$len > $rlen ) | |
1369 | { | |
1370 | $this->_error = "incomplete reply"; | |
1371 | $this->_MBPop (); | |
1372 | return false; | |
1373 | } | |
1374 | $res[] = $len ? substr ( $response, $pos, $len ) : ""; | |
1375 | $pos += $len; | |
1376 | } | |
1377 | ||
1378 | $this->_MBPop (); | |
1379 | return $res; | |
1380 | } | |
1381 | ||
1382 | ||
1383 | ///////////////////////////////////////////////////////////////////////////// | |
1384 | // keyword generation | |
1385 | ///////////////////////////////////////////////////////////////////////////// | |
1386 | ||
1387 | /// connect to searchd server, and generate keyword list for a given query | |
1388 | /// returns false on failure, | |
1389 | /// an array of words on success | |
1390 | function BuildKeywords ( $query, $index, $hits ) | |
1391 | { | |
1392 | assert ( is_string($query) ); | |
1393 | assert ( is_string($index) ); | |
1394 | assert ( is_bool($hits) ); | |
1395 | ||
1396 | $this->_MBPush (); | |
1397 | ||
1398 | if (!( $fp = $this->_Connect() )) | |
1399 | { | |
1400 | $this->_MBPop(); | |
1401 | return false; | |
1402 | } | |
1403 | ||
1404 | ///////////////// | |
1405 | // build request | |
1406 | ///////////////// | |
1407 | ||
1408 | // v.1.0 req | |
1409 | $req = pack ( "N", strlen($query) ) . $query; // req query | |
1410 | $req .= pack ( "N", strlen($index) ) . $index; // req index | |
1411 | $req .= pack ( "N", (int)$hits ); | |
1412 | ||
1413 | //////////////////////////// | |
1414 | // send query, get response | |
1415 | //////////////////////////// | |
1416 | ||
1417 | $len = strlen($req); | |
1418 | $req = pack ( "nnN", SEARCHD_COMMAND_KEYWORDS, VER_COMMAND_KEYWORDS, $len ) . $req; // add header | |
1419 | if ( !( $this->_Send ( $fp, $req, $len+8 ) ) || | |
1420 | !( $response = $this->_GetResponse ( $fp, VER_COMMAND_KEYWORDS ) ) ) | |
1421 | { | |
1422 | $this->_MBPop (); | |
1423 | return false; | |
1424 | } | |
1425 | ||
1426 | ////////////////// | |
1427 | // parse response | |
1428 | ////////////////// | |
1429 | ||
1430 | $pos = 0; | |
1431 | $res = array (); | |
1432 | $rlen = strlen($response); | |
1433 | list(,$nwords) = unpack ( "N*", substr ( $response, $pos, 4 ) ); | |
1434 | $pos += 4; | |
1435 | for ( $i=0; $i<$nwords; $i++ ) | |
1436 | { | |
1437 | list(,$len) = unpack ( "N*", substr ( $response, $pos, 4 ) ); $pos += 4; | |
1438 | $tokenized = $len ? substr ( $response, $pos, $len ) : ""; | |
1439 | $pos += $len; | |
1440 | ||
1441 | list(,$len) = unpack ( "N*", substr ( $response, $pos, 4 ) ); $pos += 4; | |
1442 | $normalized = $len ? substr ( $response, $pos, $len ) : ""; | |
1443 | $pos += $len; | |
1444 | ||
1445 | $res[] = array ( "tokenized"=>$tokenized, "normalized"=>$normalized ); | |
1446 | ||
1447 | if ( $hits ) | |
1448 | { | |
1449 | list($ndocs,$nhits) = array_values ( unpack ( "N*N*", substr ( $response, $pos, 8 ) ) ); | |
1450 | $pos += 8; | |
1451 | $res [$i]["docs"] = $ndocs; | |
1452 | $res [$i]["hits"] = $nhits; | |
1453 | } | |
1454 | ||
1455 | if ( $pos > $rlen ) | |
1456 | { | |
1457 | $this->_error = "incomplete reply"; | |
1458 | $this->_MBPop (); | |
1459 | return false; | |
1460 | } | |
1461 | } | |
1462 | ||
1463 | $this->_MBPop (); | |
1464 | return $res; | |
1465 | } | |
1466 | ||
1467 | function EscapeString ( $string ) | |
1468 | { | |
1469 | $from = array ( '\\', '(',')','|','-','!','@','~','"','&', '/', '^', '$', '=' ); | |
1470 | $to = array ( '\\\\', '\(','\)','\|','\-','\!','\@','\~','\"', '\&', '\/', '\^', '\$', '\=' ); | |
1471 | ||
1472 | return str_replace ( $from, $to, $string ); | |
1473 | } | |
1474 | ||
1475 | ///////////////////////////////////////////////////////////////////////////// | |
1476 | // attribute updates | |
1477 | ///////////////////////////////////////////////////////////////////////////// | |
1478 | ||
1479 | /// batch update given attributes in given rows in given indexes | |
1480 | /// returns amount of updated documents (0 or more) on success, or -1 on failure | |
1481 | function UpdateAttributes ( $index, $attrs, $values, $mva=false ) | |
1482 | { | |
1483 | // verify everything | |
1484 | assert ( is_string($index) ); | |
1485 | assert ( is_bool($mva) ); | |
1486 | ||
1487 | assert ( is_array($attrs) ); | |
1488 | foreach ( $attrs as $attr ) | |
1489 | assert ( is_string($attr) ); | |
1490 | ||
1491 | assert ( is_array($values) ); | |
1492 | foreach ( $values as $id=>$entry ) | |
1493 | { | |
1494 | assert ( is_numeric($id) ); | |
1495 | assert ( is_array($entry) ); | |
1496 | assert ( count($entry)==count($attrs) ); | |
1497 | foreach ( $entry as $v ) | |
1498 | { | |
1499 | if ( $mva ) | |
1500 | { | |
1501 | assert ( is_array($v) ); | |
1502 | foreach ( $v as $vv ) | |
1503 | assert ( is_int($vv) ); | |
1504 | } else | |
1505 | assert ( is_int($v) ); | |
1506 | } | |
1507 | } | |
1508 | ||
1509 | // build request | |
1510 | $req = pack ( "N", strlen($index) ) . $index; | |
1511 | ||
1512 | $req .= pack ( "N", count($attrs) ); | |
1513 | foreach ( $attrs as $attr ) | |
1514 | { | |
1515 | $req .= pack ( "N", strlen($attr) ) . $attr; | |
1516 | $req .= pack ( "N", $mva ? 1 : 0 ); | |
1517 | } | |
1518 | ||
1519 | $req .= pack ( "N", count($values) ); | |
1520 | foreach ( $values as $id=>$entry ) | |
1521 | { | |
1522 | $req .= sphPackU64 ( $id ); | |
1523 | foreach ( $entry as $v ) | |
1524 | { | |
1525 | $req .= pack ( "N", $mva ? count($v) : $v ); | |
1526 | if ( $mva ) | |
1527 | foreach ( $v as $vv ) | |
1528 | $req .= pack ( "N", $vv ); | |
1529 | } | |
1530 | } | |
1531 | ||
1532 | // connect, send query, get response | |
1533 | if (!( $fp = $this->_Connect() )) | |
1534 | return -1; | |
1535 | ||
1536 | $len = strlen($req); | |
1537 | $req = pack ( "nnN", SEARCHD_COMMAND_UPDATE, VER_COMMAND_UPDATE, $len ) . $req; // add header | |
1538 | if ( !$this->_Send ( $fp, $req, $len+8 ) ) | |
1539 | return -1; | |
1540 | ||
1541 | if (!( $response = $this->_GetResponse ( $fp, VER_COMMAND_UPDATE ) )) | |
1542 | return -1; | |
1543 | ||
1544 | // parse response | |
1545 | list(,$updated) = unpack ( "N*", substr ( $response, 0, 4 ) ); | |
1546 | return $updated; | |
1547 | } | |
1548 | ||
1549 | ///////////////////////////////////////////////////////////////////////////// | |
1550 | // persistent connections | |
1551 | ///////////////////////////////////////////////////////////////////////////// | |
1552 | ||
1553 | function Open() | |
1554 | { | |
1555 | if ( $this->_socket !== false ) | |
1556 | { | |
1557 | $this->_error = 'already connected'; | |
1558 | return false; | |
1559 | } | |
1560 | if ( !$fp = $this->_Connect() ) | |
1561 | return false; | |
1562 | ||
1563 | // command, command version = 0, body length = 4, body = 1 | |
1564 | $req = pack ( "nnNN", SEARCHD_COMMAND_PERSIST, 0, 4, 1 ); | |
1565 | if ( !$this->_Send ( $fp, $req, 12 ) ) | |
1566 | return false; | |
1567 | ||
1568 | $this->_socket = $fp; | |
1569 | return true; | |
1570 | } | |
1571 | ||
1572 | function Close() | |
1573 | { | |
1574 | if ( $this->_socket === false ) | |
1575 | { | |
1576 | $this->_error = 'not connected'; | |
1577 | return false; | |
1578 | } | |
1579 | ||
1580 | fclose ( $this->_socket ); | |
1581 | $this->_socket = false; | |
1582 | ||
1583 | return true; | |
1584 | } | |
1585 | ||
1586 | ////////////////////////////////////////////////////////////////////////// | |
1587 | // status | |
1588 | ////////////////////////////////////////////////////////////////////////// | |
1589 | ||
1590 | function Status () | |
1591 | { | |
1592 | $this->_MBPush (); | |
1593 | if (!( $fp = $this->_Connect() )) | |
1594 | { | |
1595 | $this->_MBPop(); | |
1596 | return false; | |
1597 | } | |
1598 | ||
1599 | $req = pack ( "nnNN", SEARCHD_COMMAND_STATUS, VER_COMMAND_STATUS, 4, 1 ); // len=4, body=1 | |
1600 | if ( !( $this->_Send ( $fp, $req, 12 ) ) || | |
1601 | !( $response = $this->_GetResponse ( $fp, VER_COMMAND_STATUS ) ) ) | |
1602 | { | |
1603 | $this->_MBPop (); | |
1604 | return false; | |
1605 | } | |
1606 | ||
1607 | $res = substr ( $response, 4 ); // just ignore length, error handling, etc | |
1608 | $p = 0; | |
1609 | list ( $rows, $cols ) = array_values ( unpack ( "N*N*", substr ( $response, $p, 8 ) ) ); $p += 8; | |
1610 | ||
1611 | $res = array(); | |
1612 | for ( $i=0; $i<$rows; $i++ ) | |
1613 | for ( $j=0; $j<$cols; $j++ ) | |
1614 | { | |
1615 | list(,$len) = unpack ( "N*", substr ( $response, $p, 4 ) ); $p += 4; | |
1616 | $res[$i][] = substr ( $response, $p, $len ); $p += $len; | |
1617 | } | |
1618 | ||
1619 | $this->_MBPop (); | |
1620 | return $res; | |
1621 | } | |
1622 | } | |
1623 | ||
1624 | // | |
1625 | // $Id: sphinxapi.php 2055 2009-11-06 23:09:58Z shodan $ | |
1626 | // |