c264071c |
1 | /* |
2 | * searchtools.js_t |
3 | * ~~~~~~~~~~~~~~~~ |
4 | * |
5 | * Sphinx JavaScript utilties for the full-text search. |
6 | * |
7 | * :copyright: Copyright 2007-2014 by the Sphinx team, see AUTHORS. |
8 | * :license: BSD, see LICENSE for details. |
9 | * |
10 | */ |
11 | |
12 | |
13 | /** |
14 | * Porter Stemmer |
15 | */ |
16 | var Stemmer = function() { |
17 | |
18 | var step2list = { |
19 | ational: 'ate', |
20 | tional: 'tion', |
21 | enci: 'ence', |
22 | anci: 'ance', |
23 | izer: 'ize', |
24 | bli: 'ble', |
25 | alli: 'al', |
26 | entli: 'ent', |
27 | eli: 'e', |
28 | ousli: 'ous', |
29 | ization: 'ize', |
30 | ation: 'ate', |
31 | ator: 'ate', |
32 | alism: 'al', |
33 | iveness: 'ive', |
34 | fulness: 'ful', |
35 | ousness: 'ous', |
36 | aliti: 'al', |
37 | iviti: 'ive', |
38 | biliti: 'ble', |
39 | logi: 'log' |
40 | }; |
41 | |
42 | var step3list = { |
43 | icate: 'ic', |
44 | ative: '', |
45 | alize: 'al', |
46 | iciti: 'ic', |
47 | ical: 'ic', |
48 | ful: '', |
49 | ness: '' |
50 | }; |
51 | |
52 | var c = "[^aeiou]"; // consonant |
53 | var v = "[aeiouy]"; // vowel |
54 | var C = c + "[^aeiouy]*"; // consonant sequence |
55 | var V = v + "[aeiou]*"; // vowel sequence |
56 | |
57 | var mgr0 = "^(" + C + ")?" + V + C; // [C]VC... is m>0 |
58 | var meq1 = "^(" + C + ")?" + V + C + "(" + V + ")?$"; // [C]VC[V] is m=1 |
59 | var mgr1 = "^(" + C + ")?" + V + C + V + C; // [C]VCVC... is m>1 |
60 | var s_v = "^(" + C + ")?" + v; // vowel in stem |
61 | |
62 | this.stemWord = function (w) { |
63 | var stem; |
64 | var suffix; |
65 | var firstch; |
66 | var origword = w; |
67 | |
68 | if (w.length < 3) |
69 | return w; |
70 | |
71 | var re; |
72 | var re2; |
73 | var re3; |
74 | var re4; |
75 | |
76 | firstch = w.substr(0,1); |
77 | if (firstch == "y") |
78 | w = firstch.toUpperCase() + w.substr(1); |
79 | |
80 | // Step 1a |
81 | re = /^(.+?)(ss|i)es$/; |
82 | re2 = /^(.+?)([^s])s$/; |
83 | |
84 | if (re.test(w)) |
85 | w = w.replace(re,"$1$2"); |
86 | else if (re2.test(w)) |
87 | w = w.replace(re2,"$1$2"); |
88 | |
89 | // Step 1b |
90 | re = /^(.+?)eed$/; |
91 | re2 = /^(.+?)(ed|ing)$/; |
92 | if (re.test(w)) { |
93 | var fp = re.exec(w); |
94 | re = new RegExp(mgr0); |
95 | if (re.test(fp[1])) { |
96 | re = /.$/; |
97 | w = w.replace(re,""); |
98 | } |
99 | } |
100 | else if (re2.test(w)) { |
101 | var fp = re2.exec(w); |
102 | stem = fp[1]; |
103 | re2 = new RegExp(s_v); |
104 | if (re2.test(stem)) { |
105 | w = stem; |
106 | re2 = /(at|bl|iz)$/; |
107 | re3 = new RegExp("([^aeiouylsz])\\1$"); |
108 | re4 = new RegExp("^" + C + v + "[^aeiouwxy]$"); |
109 | if (re2.test(w)) |
110 | w = w + "e"; |
111 | else if (re3.test(w)) { |
112 | re = /.$/; |
113 | w = w.replace(re,""); |
114 | } |
115 | else if (re4.test(w)) |
116 | w = w + "e"; |
117 | } |
118 | } |
119 | |
120 | // Step 1c |
121 | re = /^(.+?)y$/; |
122 | if (re.test(w)) { |
123 | var fp = re.exec(w); |
124 | stem = fp[1]; |
125 | re = new RegExp(s_v); |
126 | if (re.test(stem)) |
127 | w = stem + "i"; |
128 | } |
129 | |
130 | // Step 2 |
131 | re = /^(.+?)(ational|tional|enci|anci|izer|bli|alli|entli|eli|ousli|ization|ation|ator|alism|iveness|fulness|ousness|aliti|iviti|biliti|logi)$/; |
132 | if (re.test(w)) { |
133 | var fp = re.exec(w); |
134 | stem = fp[1]; |
135 | suffix = fp[2]; |
136 | re = new RegExp(mgr0); |
137 | if (re.test(stem)) |
138 | w = stem + step2list[suffix]; |
139 | } |
140 | |
141 | // Step 3 |
142 | re = /^(.+?)(icate|ative|alize|iciti|ical|ful|ness)$/; |
143 | if (re.test(w)) { |
144 | var fp = re.exec(w); |
145 | stem = fp[1]; |
146 | suffix = fp[2]; |
147 | re = new RegExp(mgr0); |
148 | if (re.test(stem)) |
149 | w = stem + step3list[suffix]; |
150 | } |
151 | |
152 | // Step 4 |
153 | re = /^(.+?)(al|ance|ence|er|ic|able|ible|ant|ement|ment|ent|ou|ism|ate|iti|ous|ive|ize)$/; |
154 | re2 = /^(.+?)(s|t)(ion)$/; |
155 | if (re.test(w)) { |
156 | var fp = re.exec(w); |
157 | stem = fp[1]; |
158 | re = new RegExp(mgr1); |
159 | if (re.test(stem)) |
160 | w = stem; |
161 | } |
162 | else if (re2.test(w)) { |
163 | var fp = re2.exec(w); |
164 | stem = fp[1] + fp[2]; |
165 | re2 = new RegExp(mgr1); |
166 | if (re2.test(stem)) |
167 | w = stem; |
168 | } |
169 | |
170 | // Step 5 |
171 | re = /^(.+?)e$/; |
172 | if (re.test(w)) { |
173 | var fp = re.exec(w); |
174 | stem = fp[1]; |
175 | re = new RegExp(mgr1); |
176 | re2 = new RegExp(meq1); |
177 | re3 = new RegExp("^" + C + v + "[^aeiouwxy]$"); |
178 | if (re.test(stem) || (re2.test(stem) && !(re3.test(stem)))) |
179 | w = stem; |
180 | } |
181 | re = /ll$/; |
182 | re2 = new RegExp(mgr1); |
183 | if (re.test(w) && re2.test(w)) { |
184 | re = /.$/; |
185 | w = w.replace(re,""); |
186 | } |
187 | |
188 | // and turn initial Y back to y |
189 | if (firstch == "y") |
190 | w = firstch.toLowerCase() + w.substr(1); |
191 | return w; |
192 | } |
193 | } |
194 | |
195 | |
196 | |
197 | /** |
198 | * Simple result scoring code. |
199 | */ |
200 | var Scorer = { |
201 | // Implement the following function to further tweak the score for each result |
202 | // The function takes a result array [filename, title, anchor, descr, score] |
203 | // and returns the new score. |
204 | /* |
205 | score: function(result) { |
206 | return result[4]; |
207 | }, |
208 | */ |
209 | |
210 | // query matches the full name of an object |
211 | objNameMatch: 11, |
212 | // or matches in the last dotted part of the object name |
213 | objPartialMatch: 6, |
214 | // Additive scores depending on the priority of the object |
215 | objPrio: {0: 15, // used to be importantResults |
216 | 1: 5, // used to be objectResults |
217 | 2: -5}, // used to be unimportantResults |
218 | // Used when the priority is not in the mapping. |
219 | objPrioDefault: 0, |
220 | |
221 | // query found in title |
222 | title: 15, |
223 | // query found in terms |
224 | term: 5 |
225 | }; |
226 | |
227 | |
228 | /** |
229 | * Search Module |
230 | */ |
231 | var Search = { |
232 | |
233 | _index : null, |
234 | _queued_query : null, |
235 | _pulse_status : -1, |
236 | |
237 | init : function() { |
238 | var params = $.getQueryParameters(); |
239 | if (params.q) { |
240 | var query = params.q[0]; |
241 | $('input[name="q"]')[0].value = query; |
242 | this.performSearch(query); |
243 | } |
244 | }, |
245 | |
246 | loadIndex : function(url) { |
247 | $.ajax({type: "GET", url: url, data: null, |
248 | dataType: "script", cache: true, |
249 | complete: function(jqxhr, textstatus) { |
250 | if (textstatus != "success") { |
251 | document.getElementById("searchindexloader").src = url; |
252 | } |
253 | }}); |
254 | }, |
255 | |
256 | setIndex : function(index) { |
257 | var q; |
258 | this._index = index; |
259 | if ((q = this._queued_query) !== null) { |
260 | this._queued_query = null; |
261 | Search.query(q); |
262 | } |
263 | }, |
264 | |
265 | hasIndex : function() { |
266 | return this._index !== null; |
267 | }, |
268 | |
269 | deferQuery : function(query) { |
270 | this._queued_query = query; |
271 | }, |
272 | |
273 | stopPulse : function() { |
274 | this._pulse_status = 0; |
275 | }, |
276 | |
277 | startPulse : function() { |
278 | if (this._pulse_status >= 0) |
279 | return; |
280 | function pulse() { |
281 | var i; |
282 | Search._pulse_status = (Search._pulse_status + 1) % 4; |
283 | var dotString = ''; |
284 | for (i = 0; i < Search._pulse_status; i++) |
285 | dotString += '.'; |
286 | Search.dots.text(dotString); |
287 | if (Search._pulse_status > -1) |
288 | window.setTimeout(pulse, 500); |
289 | } |
290 | pulse(); |
291 | }, |
292 | |
293 | /** |
294 | * perform a search for something (or wait until index is loaded) |
295 | */ |
296 | performSearch : function(query) { |
297 | // create the required interface elements |
298 | this.out = $('#search-results'); |
299 | this.title = $('<h2>' + _('Searching') + '</h2>').appendTo(this.out); |
300 | this.dots = $('<span></span>').appendTo(this.title); |
301 | this.status = $('<p style="display: none"></p>').appendTo(this.out); |
302 | this.output = $('<ul class="search"/>').appendTo(this.out); |
303 | |
304 | $('#search-progress').text(_('Preparing search...')); |
305 | this.startPulse(); |
306 | |
307 | // index already loaded, the browser was quick! |
308 | if (this.hasIndex()) |
309 | this.query(query); |
310 | else |
311 | this.deferQuery(query); |
312 | }, |
313 | |
314 | /** |
315 | * execute search (requires search index to be loaded) |
316 | */ |
317 | query : function(query) { |
318 | var i; |
319 | var stopwords = ["a","and","are","as","at","be","but","by","for","if","in","into","is","it","near","no","not","of","on","or","such","that","the","their","then","there","these","they","this","to","was","will","with"]; |
320 | |
321 | // stem the searchterms and add them to the correct list |
322 | var stemmer = new Stemmer(); |
323 | var searchterms = []; |
324 | var excluded = []; |
325 | var hlterms = []; |
326 | var tmp = query.split(/\s+/); |
327 | var objectterms = []; |
328 | for (i = 0; i < tmp.length; i++) { |
329 | if (tmp[i] !== "") { |
330 | objectterms.push(tmp[i].toLowerCase()); |
331 | } |
332 | |
333 | if ($u.indexOf(stopwords, tmp[i].toLowerCase()) != -1 || tmp[i].match(/^\d+$/) || |
334 | tmp[i] === "") { |
335 | // skip this "word" |
336 | continue; |
337 | } |
338 | // stem the word |
339 | var word = stemmer.stemWord(tmp[i].toLowerCase()); |
340 | var toAppend; |
341 | // select the correct list |
342 | if (word[0] == '-') { |
343 | toAppend = excluded; |
344 | word = word.substr(1); |
345 | } |
346 | else { |
347 | toAppend = searchterms; |
348 | hlterms.push(tmp[i].toLowerCase()); |
349 | } |
350 | // only add if not already in the list |
351 | if (!$u.contains(toAppend, word)) |
352 | toAppend.push(word); |
353 | } |
354 | var highlightstring = '?highlight=' + $.urlencode(hlterms.join(" ")); |
355 | |
356 | // console.debug('SEARCH: searching for:'); |
357 | // console.info('required: ', searchterms); |
358 | // console.info('excluded: ', excluded); |
359 | |
360 | // prepare search |
361 | var terms = this._index.terms; |
362 | var titleterms = this._index.titleterms; |
363 | |
364 | // array of [filename, title, anchor, descr, score] |
365 | var results = []; |
366 | $('#search-progress').empty(); |
367 | |
368 | // lookup as object |
369 | for (i = 0; i < objectterms.length; i++) { |
370 | var others = [].concat(objectterms.slice(0, i), |
371 | objectterms.slice(i+1, objectterms.length)); |
372 | results = results.concat(this.performObjectSearch(objectterms[i], others)); |
373 | } |
374 | |
375 | // lookup as search terms in fulltext |
376 | results = results.concat(this.performTermsSearch(searchterms, excluded, terms, Scorer.term)) |
377 | .concat(this.performTermsSearch(searchterms, excluded, titleterms, Scorer.title)); |
378 | |
379 | // let the scorer override scores with a custom scoring function |
380 | if (Scorer.score) { |
381 | for (i = 0; i < results.length; i++) |
382 | results[i][4] = Scorer.score(results[i]); |
383 | } |
384 | |
385 | // now sort the results by score (in opposite order of appearance, since the |
386 | // display function below uses pop() to retrieve items) and then |
387 | // alphabetically |
388 | results.sort(function(a, b) { |
389 | var left = a[4]; |
390 | var right = b[4]; |
391 | if (left > right) { |
392 | return 1; |
393 | } else if (left < right) { |
394 | return -1; |
395 | } else { |
396 | // same score: sort alphabetically |
397 | left = a[1].toLowerCase(); |
398 | right = b[1].toLowerCase(); |
399 | return (left > right) ? -1 : ((left < right) ? 1 : 0); |
400 | } |
401 | }); |
402 | |
403 | // for debugging |
404 | //Search.lastresults = results.slice(); // a copy |
405 | //console.info('search results:', Search.lastresults); |
406 | |
407 | // print the results |
408 | var resultCount = results.length; |
409 | function displayNextItem() { |
410 | // results left, load the summary and display it |
411 | if (results.length) { |
412 | var item = results.pop(); |
413 | var listItem = $('<li style="display:none"></li>'); |
414 | if (DOCUMENTATION_OPTIONS.FILE_SUFFIX === '') { |
415 | // dirhtml builder |
416 | var dirname = item[0] + '/'; |
417 | if (dirname.match(/\/index\/$/)) { |
418 | dirname = dirname.substring(0, dirname.length-6); |
419 | } else if (dirname == 'index/') { |
420 | dirname = ''; |
421 | } |
422 | listItem.append($('<a/>').attr('href', |
423 | DOCUMENTATION_OPTIONS.URL_ROOT + dirname + |
424 | highlightstring + item[2]).html(item[1])); |
425 | } else { |
426 | // normal html builders |
427 | listItem.append($('<a/>').attr('href', |
428 | item[0] + DOCUMENTATION_OPTIONS.FILE_SUFFIX + |
429 | highlightstring + item[2]).html(item[1])); |
430 | } |
431 | if (item[3]) { |
432 | listItem.append($('<span> (' + item[3] + ')</span>')); |
433 | Search.output.append(listItem); |
434 | listItem.slideDown(5, function() { |
435 | displayNextItem(); |
436 | }); |
437 | } else if (DOCUMENTATION_OPTIONS.HAS_SOURCE) { |
438 | $.ajax({url: DOCUMENTATION_OPTIONS.URL_ROOT + '_sources/' + item[0] + '.txt', |
439 | dataType: "text", |
440 | complete: function(jqxhr, textstatus) { |
441 | var data = jqxhr.responseText; |
442 | if (data !== '') { |
443 | listItem.append(Search.makeSearchSummary(data, searchterms, hlterms)); |
444 | } |
445 | Search.output.append(listItem); |
446 | listItem.slideDown(5, function() { |
447 | displayNextItem(); |
448 | }); |
449 | }}); |
450 | } else { |
451 | // no source available, just display title |
452 | Search.output.append(listItem); |
453 | listItem.slideDown(5, function() { |
454 | displayNextItem(); |
455 | }); |
456 | } |
457 | } |
458 | // search finished, update title and status message |
459 | else { |
460 | Search.stopPulse(); |
461 | Search.title.text(_('Search Results')); |
462 | if (!resultCount) |
463 | Search.status.text(_('Your search did not match any documents. Please make sure that all words are spelled correctly and that you\'ve selected enough categories.')); |
464 | else |
465 | Search.status.text(_('Search finished, found %s page(s) matching the search query.').replace('%s', resultCount)); |
466 | Search.status.fadeIn(500); |
467 | } |
468 | } |
469 | displayNextItem(); |
470 | }, |
471 | |
472 | /** |
473 | * search for object names |
474 | */ |
475 | performObjectSearch : function(object, otherterms) { |
476 | var filenames = this._index.filenames; |
477 | var objects = this._index.objects; |
478 | var objnames = this._index.objnames; |
479 | var titles = this._index.titles; |
480 | |
481 | var i; |
482 | var results = []; |
483 | |
484 | for (var prefix in objects) { |
485 | for (var name in objects[prefix]) { |
486 | var fullname = (prefix ? prefix + '.' : '') + name; |
487 | if (fullname.toLowerCase().indexOf(object) > -1) { |
488 | var score = 0; |
489 | var parts = fullname.split('.'); |
490 | // check for different match types: exact matches of full name or |
491 | // "last name" (i.e. last dotted part) |
492 | if (fullname == object || parts[parts.length - 1] == object) { |
493 | score += Scorer.objNameMatch; |
494 | // matches in last name |
495 | } else if (parts[parts.length - 1].indexOf(object) > -1) { |
496 | score += Scorer.objPartialMatch; |
497 | } |
498 | var match = objects[prefix][name]; |
499 | var objname = objnames[match[1]][2]; |
500 | var title = titles[match[0]]; |
501 | // If more than one term searched for, we require other words to be |
502 | // found in the name/title/description |
503 | if (otherterms.length > 0) { |
504 | var haystack = (prefix + ' ' + name + ' ' + |
505 | objname + ' ' + title).toLowerCase(); |
506 | var allfound = true; |
507 | for (i = 0; i < otherterms.length; i++) { |
508 | if (haystack.indexOf(otherterms[i]) == -1) { |
509 | allfound = false; |
510 | break; |
511 | } |
512 | } |
513 | if (!allfound) { |
514 | continue; |
515 | } |
516 | } |
517 | var descr = objname + _(', in ') + title; |
518 | |
519 | var anchor = match[3]; |
520 | if (anchor === '') |
521 | anchor = fullname; |
522 | else if (anchor == '-') |
523 | anchor = objnames[match[1]][1] + '-' + fullname; |
524 | // add custom score for some objects according to scorer |
525 | if (Scorer.objPrio.hasOwnProperty(match[2])) { |
526 | score += Scorer.objPrio[match[2]]; |
527 | } else { |
528 | score += Scorer.objPrioDefault; |
529 | } |
530 | results.push([filenames[match[0]], fullname, '#'+anchor, descr, score]); |
531 | } |
532 | } |
533 | } |
534 | |
535 | return results; |
536 | }, |
537 | |
538 | /** |
539 | * search for full-text terms in the index |
540 | */ |
541 | performTermsSearch : function(searchterms, excluded, terms, score) { |
542 | var filenames = this._index.filenames; |
543 | var titles = this._index.titles; |
544 | |
545 | var i, j, file, files; |
546 | var fileMap = {}; |
547 | var results = []; |
548 | |
549 | // perform the search on the required terms |
550 | for (i = 0; i < searchterms.length; i++) { |
551 | var word = searchterms[i]; |
552 | // no match but word was a required one |
553 | if ((files = terms[word]) === undefined) |
554 | break; |
555 | if (files.length === undefined) { |
556 | files = [files]; |
557 | } |
558 | // create the mapping |
559 | for (j = 0; j < files.length; j++) { |
560 | file = files[j]; |
561 | if (file in fileMap) |
562 | fileMap[file].push(word); |
563 | else |
564 | fileMap[file] = [word]; |
565 | } |
566 | } |
567 | |
568 | // now check if the files don't contain excluded terms |
569 | for (file in fileMap) { |
570 | var valid = true; |
571 | |
572 | // check if all requirements are matched |
573 | if (fileMap[file].length != searchterms.length) |
574 | continue; |
575 | |
576 | // ensure that none of the excluded terms is in the search result |
577 | for (i = 0; i < excluded.length; i++) { |
578 | if (terms[excluded[i]] == file || |
579 | $u.contains(terms[excluded[i]] || [], file)) { |
580 | valid = false; |
581 | break; |
582 | } |
583 | } |
584 | |
585 | // if we have still a valid result we can add it to the result list |
586 | if (valid) { |
587 | results.push([filenames[file], titles[file], '', null, score]); |
588 | } |
589 | } |
590 | return results; |
591 | }, |
592 | |
593 | /** |
594 | * helper function to return a node containing the |
595 | * search summary for a given text. keywords is a list |
596 | * of stemmed words, hlwords is the list of normal, unstemmed |
597 | * words. the first one is used to find the occurance, the |
598 | * latter for highlighting it. |
599 | */ |
600 | makeSearchSummary : function(text, keywords, hlwords) { |
601 | var textLower = text.toLowerCase(); |
602 | var start = 0; |
603 | $.each(keywords, function() { |
604 | var i = textLower.indexOf(this.toLowerCase()); |
605 | if (i > -1) |
606 | start = i; |
607 | }); |
608 | start = Math.max(start - 120, 0); |
609 | var excerpt = ((start > 0) ? '...' : '') + |
610 | $.trim(text.substr(start, 240)) + |
611 | ((start + 240 - text.length) ? '...' : ''); |
612 | var rv = $('<div class="context"></div>').text(excerpt); |
613 | $.each(hlwords, function() { |
614 | rv = rv.highlightText(this, 'highlighted'); |
615 | }); |
616 | return rv; |
617 | } |
618 | }; |
619 | |
620 | $(document).ready(function() { |
621 | Search.init(); |
622 | }); |