\c@s"dZddlZddlmZddlmZmZmZmZddl m Z ddl m Z mZmZmZmZmZddlZddlmZmZddlmZdd lmZer'dd lmZmZmZm Z m!Z!m"Z"m#Z#m$Z$dd l%m&Z&dd l'm(Z(nd e)fdYZ*ddl+m,Z,dZ-idd6dd6e,d6dd6dd6dd6dd6dd6d d!6d"d#6d$d%6d&d'6d(d)6d*d+6d,d-6d.d/6d0d16Z.d2e)fd3YZ/e/Z0d4efd5YZ1d6e)fd7YZ2dS(8s sphinx.search ~~~~~~~~~~~~~ Create a full-text search index for offline search. :copyright: Copyright 2007-2019 by the Sphinx team, see AUTHORS. :license: BSD, see LICENSE for details. iN(tpath(t iteritemst itervaluest text_typet string_types(tcPickle(trawtcommentttitletTextt NodeVisitortSkipNode(tjsdumpt rpartition(t htmlescape(t splitter_code(tAnytDicttIOtIterabletListtTupletTypetSet(tnodes(tBuildEnvironmenttSearchLanguagecBskeZdZdZdZeZdZdZ e j dZ dZ dZdZdZdZRS( sI This class is the base class for search natural language preprocessors. If you want to add support for a new language, you should override the methods of this class. You should override `lang` class property too (e.g. 'en', 'fr' and so on). .. attribute:: stopwords This is a set of stop words of the target language. Default `stopwords` is empty. This word is used for building index and embedded in JS. .. attribute:: js_stemmer_code Return stemmer class of JavaScript version. This class' name should be ``Stemmer`` and this class must have ``stemWord`` method. This string is embedded as-is in searchtools.js. This class is used to preprocess search word which Sphinx HTML readers type, before searching index. Default implementation does nothing. s /** * Dummy stemmer for languages without stemming rules. */ var Stemmer = function() { this.stemWord = function(w) { return w; } } s(?u)\w+cCs||_|j|dS(N(toptionstinit(tselfR((s5lib/python2.7/site-packages/sphinx/search/__init__.pyt__init__Fs cCsdS(sK Initialize the class with the options the user has given. N((RR((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRKscCs|jj|S(s This method splits a sentence into words. Default splitter splits input at white spaces, which should be enough for most languages except CJK languages. (t_word_retfindall(Rtinput((s5lib/python2.7/site-packages/sphinx/search/__init__.pytsplitQscCs|S(s This method implements stemming algorithm of the Python version. Default implementation does nothing. You should implement this if the language has any stemming rules. This class is used to preprocess search words before registering them in the search index. The stemming of the Python version and the JS version (given in the js_stemmer_code attribute) must be compatible. ((Rtword((s5lib/python2.7/site-packages/sphinx/search/__init__.pytstemZs cCst|dkpt|dkrJdt|dkoEdknpt|ddkot|dkp||jkp|j S(s Return true if the target word should be registered in the search index. This method is called after stemming. iiiA0i0i(tlentordt stopwordstisdigit(RR#((s5lib/python2.7/site-packages/sphinx/search/__init__.pyt word_filterhs8N(t__name__t __module__t__doc__tNonetlangt language_nametsetR'tjs_stemmer_rawcodetjs_stemmer_codetretcompileRRRR"R$R)(((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRs     (t SearchEnglishcCsJt}x:|jD],}|jdd}|j|jqW|S(ss parse snowball style word list like this: * http://snowball.tartarus.org/algorithms/finnish/stop.txt t|i(R0t splitlinesR"tupdate(tsourcetresulttline((s5lib/python2.7/site-packages/sphinx/search/__init__.pytparse_stop_wordzs  ssphinx.search.da.SearchDanishtdassphinx.search.de.SearchGermantdetenssphinx.search.es.SearchSpanishtesssphinx.search.fi.SearchFinnishtfissphinx.search.fr.SearchFrenchtfrs sphinx.search.hu.SearchHungarianthussphinx.search.it.SearchItaliantitssphinx.search.ja.SearchJapanesetjassphinx.search.nl.SearchDutchtnls sphinx.search.no.SearchNorwegiantnos!sphinx.search.pt.SearchPortuguesetptssphinx.search.ro.SearchRomaniantrossphinx.search.ru.SearchRussiantrussphinx.search.sv.SearchSwedishtsvssphinx.search.tr.SearchTurkishttrssphinx.search.zh.SearchChinesetzht_JavaScriptIndexcBs>eZdZdZdZdZdZdZdZRS(s The search index as javascript file that calls a function on the documentation search object to register the index. sSearch.setIndex(t)cCs|jtj||jS(N(tPREFIXR tdumpstSUFFIX(Rtdata((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRQscCsi|t|jt|j !}| sM|j|j sM|j|j r\tdntj|S(Ns invalid data(R%RPRRt startswithtendswitht ValueErrorR tloads(RtsRS((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRWs  cCs|j|j|dS(N(twriteRQ(RRStf((s5lib/python2.7/site-packages/sphinx/search/__init__.pytdumpscCs|j|jS(N(RWtread(RRZ((s5lib/python2.7/site-packages/sphinx/search/__init__.pytloads( R*R+R,RPRRRQRWR[R](((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRNs   t WordCollectorcBs)eZdZdZdZdZRS(sG A special visitor that collects words for the `IndexBuilder`. cCs/tj||g|_g|_||_dS(N(R Rt found_wordstfound_title_wordsR.(RtdocumentR.((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRs  cCsft|tjjrb|jddkrb|jd}|dkrItS||jjkrbtSntS(NtnametkeywordsR.( t isinstancetsphinxtaddnodestmetatgetR-tTrueR.tFalse(Rtnodetnodetypet meta_lang((s5lib/python2.7/site-packages/sphinx/search/__init__.pytis_meta_keywordss* cCst|}t|tr$tnt|trd|jddjkrtjdd|j }tjdd|}tjdd|}|j j |j j|ntnt|t r|j j |j j|j nt|tr&|jj |j j|j nW|j||r}|d}g|jdD]}|j^qR}|j j |ndS( Nthtmltformatts(?is)s(?is)s<[^<]+?>tcontentt,(ttypet issubclassRR RRhR"R3tsubtastextR_textendR.R RR`Rntstrip(RRkRltnodetextRctkeyword((s5lib/python2.7/site-packages/sphinx/search/__init__.pytdispatch_visits$   %% ((R*R+R,RRnR|(((s5lib/python2.7/site-packages/sphinx/search/__init__.pyR^s  t IndexBuildercBseZdZied6ed6ZdZdZdZdZ dZ dZ d Z d Z d Zd Zd ZRS(sh Helper class that creates a searchindex based on the doctrees passed to the `feed` method. R tpicklec Csd||_i|_i|_i|_i|_i|_i|_i|_tj |}|dkrd|krtj |j dd}n|dkrt ||_ nit|tr|jdd\}}tt|dd|g|}|||_ n|||_ |rNt|d}|jjd|_WdQXn d|_t|_dS(Nt_it.itrbsutf-8u(tenvt_titlest _filenamest_mappingt_title_mappingt _stem_cachet _objtypest _objnamest languagesRhR-R"R5R.Rdtstrtrsplittgetattrt __import__topenR\tdecodetjs_scorer_codeRtjs_splitter_code( RRR.Rtscoringt lang_classtmodulet classnametfp((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRs0          ! cst|tr|j|}n|j|}t|t sY|jd|jjkrhtdn|dtt |d|_ tt |d|_ fd}||d|_ ||d|_ d S( sReconstruct from frozen data.t envversions old formattdocnamest filenamesttitlescspi}xct|D]U\}}t|trHt|g||8s(RRdtintR0(tmappingtrvtktv(R(s5lib/python2.7/site-packages/sphinx/search/__init__.pyt load_terms1s $ttermst titletermsN(RdRtformatsR]tdictRhRtversionRVtzipRRRR(RtstreamRptfrozenR((Rs5lib/python2.7/site-packages/sphinx/search/__init__.pyR]#s  cCs9t|tr|j|}n|j|j|dS(s"Dump the frozen index to a stream.N(RdRRR[tfreeze(RRRp((s5lib/python2.7/site-packages/sphinx/search/__init__.pyR[?scCsi}|j}|j}xtt|jjD]\}}xut|jD]a\}}} } } } | |krzqPn| dkrqPnt|}t|}t|d\} }|j | i}y||| f}Wn}t k r^t |}|||| f<|j j | }|rH|| t|j|f||  %(         "cCsiif}xt||j|jfD]\}}xt|D]z\}}t|dkr|\}||kr||||}scss1|]'\}}||dd|dfVqdS(it:iN((RRR((s5lib/python2.7/site-packages/sphinx/search/__init__.pys sRRRRtobjectstobjtypestobjnamesRR(RRRtitemsRRhRt enumerateRRRRRRR( RRRRRRRt title_termsRRR((s5lib/python2.7/site-packages/sphinx/search/__init__.pyRxs!%  cCsd|jj|jjfS(Ns %s (code: %s)(R.R/(R((s5lib/python2.7/site-packages/sphinx/search/__init__.pytlabelscCsi}i}xB|D]:}||jkr|j|||<|j||| sL ". :X   0