The world's most popular open source database

MySQL 5.1: mysql/src/5.1-dbg/storage/myisam/ft_static.c Source File
mysql/src/5.1-dbg/storage/myisam/ft_static.c

00001 /* Copyright (C) 2000 MySQL AB & MySQL Finland AB & TCX DataKonsult AB
00002 
00003    This program is free software; you can redistribute it and/or modify
00004    it under the terms of the GNU General Public License as published by
00005    the Free Software Foundation; either version 2 of the License, or
00006    (at your option) any later version.
00007 
00008    This program is distributed in the hope that it will be useful,
00009    but WITHOUT ANY WARRANTY; without even the implied warranty of
00010    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00011    GNU General Public License for more details.
00012 
00013    You should have received a copy of the GNU General Public License
00014    along with this program; if not, write to the Free Software
00015    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA */
00016 
00017 /* Written by Sergei A. Golubchik, who has a shared copyright to this code */
00018 
00019 #include "ftdefs.h"
00020 
00021 ulong ft_min_word_len=4;
00022 ulong ft_max_word_len=HA_FT_MAXCHARLEN;
00023 ulong ft_query_expansion_limit=5;
00024 char ft_boolean_syntax[]="+ -><()~*:\"\"&|";
00025 
00026 const HA_KEYSEG ft_keysegs[FT_SEGS]={
00027 {
00028   0,                                            /* charset  */
00029   HA_FT_WLEN,                                   /* start */
00030   0,                                            /* null_pos */
00031   0,                                            /* Bit pos */
00032   HA_VAR_LENGTH_PART | HA_PACK_KEY,             /* flag */
00033   HA_FT_MAXBYTELEN,                             /* length */
00034   HA_KEYTYPE_VARTEXT2,                          /* type */
00035   63,                                           /* language (will be overwritten) */
00036   0,                                            /* null_bit */
00037   2, 0, 0                                       /* bit_start, bit_end, bit_length */
00038 },
00039 {
00040   /*
00041       Note, this (and the last HA_KEYTYPE_END) segment should NOT
00042       be packed in any way, otherwise w_search() won't be able to
00043       update key entry 'in vivo'
00044     */
00045   0, 0, 0, 0, HA_NO_SORT, HA_FT_WLEN, HA_FT_WTYPE, 63, 0, 0, 0, 0
00046 }
00047 };
00048 
00049 const struct _ft_vft _ft_vft_nlq = {
00050   ft_nlq_read_next, ft_nlq_find_relevance, ft_nlq_close_search,
00051   ft_nlq_get_relevance,  ft_nlq_reinit_search
00052 };
00053 const struct _ft_vft _ft_vft_boolean = {
00054   ft_boolean_read_next, ft_boolean_find_relevance, ft_boolean_close_search,
00055   ft_boolean_get_relevance,  ft_boolean_reinit_search
00056 };
00057 
00058 
00059 FT_INFO *ft_init_search(uint flags, void *info, uint keynr,
00060                         byte *query, uint query_len, CHARSET_INFO *cs,
00061                         byte *record)
00062 {
00063   FT_INFO *res;
00064   if (flags & FT_BOOL)
00065     res= ft_init_boolean_search((MI_INFO *)info, keynr, query, query_len,cs);
00066   else
00067     res= ft_init_nlq_search((MI_INFO *)info, keynr, query, query_len, flags,
00068                             record);
00069   return res;
00070 }
00071 
00072 const char *ft_stopword_file = 0;
00073 const char *ft_precompiled_stopwords[] = {
00074 
00075 #ifdef COMPILE_STOPWORDS_IN
00076 
00077 /* This particular stopword list was taken from SMART distribution
00078    ftp://ftp.cs.cornell.edu/pub/smart/smart.11.0.tar.Z
00079    it was slightly modified to my taste, though
00080  */
00081 
00082   "a's",
00083   "able",
00084   "about",
00085   "above",
00086   "according",
00087   "accordingly",
00088   "across",
00089   "actually",
00090   "after",
00091   "afterwards",
00092   "again",
00093   "against",
00094   "ain't",
00095   "all",
00096   "allow",
00097   "allows",
00098   "almost",
00099   "alone",
00100   "along",
00101   "already",
00102   "also",
00103   "although",
00104   "always",
00105   "am",
00106   "among",
00107   "amongst",
00108   "an",
00109   "and",
00110   "another",
00111   "any",
00112   "anybody",
00113   "anyhow",
00114   "anyone",
00115   "anything",
00116   "anyway",
00117   "anyways",
00118   "anywhere",
00119   "apart",
00120   "appear",
00121   "appreciate",
00122   "appropriate",
00123   "are",
00124   "aren't",
00125   "around",
00126   "as",
00127   "aside",
00128   "ask",
00129   "asking",
00130   "associated",
00131   "at",
00132   "available",
00133   "away",
00134   "awfully",
00135   "be",
00136   "became",
00137   "because",
00138   "become",
00139   "becomes",
00140   "becoming",
00141   "been",
00142   "before",
00143   "beforehand",
00144   "behind",
00145   "being",
00146   "believe",
00147   "below",
00148   "beside",
00149   "besides",
00150   "best",
00151   "better",
00152   "between",
00153   "beyond",
00154   "both",
00155   "brief",
00156   "but",
00157   "by",
00158   "c'mon",
00159   "c's",
00160   "came",
00161   "can",
00162   "can't",
00163   "cannot",
00164   "cant",
00165   "cause",
00166   "causes",
00167   "certain",
00168   "certainly",
00169   "changes",
00170   "clearly",
00171   "co",
00172   "com",
00173   "come",
00174   "comes",
00175   "concerning",
00176   "consequently",
00177   "consider",
00178   "considering",
00179   "contain",
00180   "containing",
00181   "contains",
00182   "corresponding",
00183   "could",
00184   "couldn't",
00185   "course",
00186   "currently",
00187   "definitely",
00188   "described",
00189   "despite",
00190   "did",
00191   "didn't",
00192   "different",
00193   "do",
00194   "does",
00195   "doesn't",
00196   "doing",
00197   "don't",
00198   "done",
00199   "down",
00200   "downwards",
00201   "during",
00202   "each",
00203   "edu",
00204   "eg",
00205   "eight",
00206   "either",
00207   "else",
00208   "elsewhere",
00209   "enough",
00210   "entirely",
00211   "especially",
00212   "et",
00213   "etc",
00214   "even",
00215   "ever",
00216   "every",
00217   "everybody",
00218   "everyone",
00219   "everything",
00220   "everywhere",
00221   "ex",
00222   "exactly",
00223   "example",
00224   "except",
00225   "far",
00226   "few",
00227   "fifth",
00228   "first",
00229   "five",
00230   "followed",
00231   "following",
00232   "follows",
00233   "for",
00234   "former",
00235   "formerly",
00236   "forth",
00237   "four",
00238   "from",
00239   "further",
00240   "furthermore",
00241   "get",
00242   "gets",
00243   "getting",
00244   "given",
00245   "gives",
00246   "go",
00247   "goes",
00248   "going",
00249   "gone",
00250   "got",
00251   "gotten",
00252   "greetings",
00253   "had",
00254   "hadn't",
00255   "happens",
00256   "hardly",
00257   "has",
00258   "hasn't",
00259   "have",
00260   "haven't",
00261   "having",
00262   "he",
00263   "he's",
00264   "hello",
00265   "help",
00266   "hence",
00267   "her",
00268   "here",
00269   "here's",
00270   "hereafter",
00271   "hereby",
00272   "herein",
00273   "hereupon",
00274   "hers",
00275   "herself",
00276   "hi",
00277   "him",
00278   "himself",
00279   "his",
00280   "hither",
00281   "hopefully",
00282   "how",
00283   "howbeit",
00284   "however",
00285   "i'd",
00286   "i'll",
00287   "i'm",
00288   "i've",
00289   "ie",
00290   "if",
00291   "ignored",
00292   "immediate",
00293   "in",
00294   "inasmuch",
00295   "inc",
00296   "indeed",
00297   "indicate",
00298   "indicated",
00299   "indicates",
00300   "inner",
00301   "insofar",
00302   "instead",
00303   "into",
00304   "inward",
00305   "is",
00306   "isn't",
00307   "it",
00308   "it'd",
00309   "it'll",
00310   "it's",
00311   "its",
00312   "itself",
00313   "just",
00314   "keep",
00315   "keeps",
00316   "kept",
00317   "know",
00318   "knows",
00319   "known",
00320   "last",
00321   "lately",
00322   "later",
00323   "latter",
00324   "latterly",
00325   "least",
00326   "less",
00327   "lest",
00328   "let",
00329   "let's",
00330   "like",
00331   "liked",
00332   "likely",
00333   "little",
00334   "look",
00335   "looking",
00336   "looks",
00337   "ltd",
00338   "mainly",
00339   "many",
00340   "may",
00341   "maybe",
00342   "me",
00343   "mean",
00344   "meanwhile",
00345   "merely",
00346   "might",
00347   "more",
00348   "moreover",
00349   "most",
00350   "mostly",
00351   "much",
00352   "must",
00353   "my",
00354   "myself",
00355   "name",
00356   "namely",
00357   "nd",
00358   "near",
00359   "nearly",
00360   "necessary",
00361   "need",
00362   "needs",
00363   "neither",
00364   "never",
00365   "nevertheless",
00366   "new",
00367   "next",
00368   "nine",
00369   "no",
00370   "nobody",
00371   "non",
00372   "none",
00373   "noone",
00374   "nor",
00375   "normally",
00376   "not",
00377   "nothing",
00378   "novel",
00379   "now",
00380   "nowhere",
00381   "obviously",
00382   "of",
00383   "off",
00384   "often",
00385   "oh",
00386   "ok",
00387   "okay",
00388   "old",
00389   "on",
00390   "once",
00391   "one",
00392   "ones",
00393   "only",
00394   "onto",
00395   "or",
00396   "other",
00397   "others",
00398   "otherwise",
00399   "ought",
00400   "our",
00401   "ours",
00402   "ourselves",
00403   "out",
00404   "outside",
00405   "over",
00406   "overall",
00407   "own",
00408   "particular",
00409   "particularly",
00410   "per",
00411   "perhaps",
00412   "placed",
00413   "please",
00414   "plus",
00415   "possible",
00416   "presumably",
00417   "probably",
00418   "provides",
00419   "que",
00420   "quite",
00421   "qv",
00422   "rather",
00423   "rd",
00424   "re",
00425   "really",
00426   "reasonably",
00427   "regarding",
00428   "regardless",
00429   "regards",
00430   "relatively",
00431   "respectively",
00432   "right",
00433   "said",
00434   "same",
00435   "saw",
00436   "say",
00437   "saying",
00438   "says",
00439   "second",
00440   "secondly",
00441   "see",
00442   "seeing",
00443   "seem",
00444   "seemed",
00445   "seeming",
00446   "seems",
00447   "seen",
00448   "self",
00449   "selves",
00450   "sensible",
00451   "sent",
00452   "serious",
00453   "seriously",
00454   "seven",
00455   "several",
00456   "shall",
00457   "she",
00458   "should",
00459   "shouldn't",
00460   "since",
00461   "six",
00462   "so",
00463   "some",
00464   "somebody",
00465   "somehow",
00466   "someone",
00467   "something",
00468   "sometime",
00469   "sometimes",
00470   "somewhat",
00471   "somewhere",
00472   "soon",
00473   "sorry",
00474   "specified",
00475   "specify",
00476   "specifying",
00477   "still",
00478   "sub",
00479   "such",
00480   "sup",
00481   "sure",
00482   "t's",
00483   "take",
00484   "taken",
00485   "tell",
00486   "tends",
00487   "th",
00488   "than",
00489   "thank",
00490   "thanks",
00491   "thanx",
00492   "that",
00493   "that's",
00494   "thats",
00495   "the",
00496   "their",
00497   "theirs",
00498   "them",
00499   "themselves",
00500   "then",
00501   "thence",
00502   "there",
00503   "there's",
00504   "thereafter",
00505   "thereby",
00506   "therefore",
00507   "therein",
00508   "theres",
00509   "thereupon",
00510   "these",
00511   "they",
00512   "they'd",
00513   "they'll",
00514   "they're",
00515   "they've",
00516   "think",
00517   "third",
00518   "this",
00519   "thorough",
00520   "thoroughly",
00521   "those",
00522   "though",
00523   "three",
00524   "through",
00525   "throughout",
00526   "thru",
00527   "thus",
00528   "to",
00529   "together",
00530   "too",
00531   "took",
00532   "toward",
00533   "towards",
00534   "tried",
00535   "tries",
00536   "truly",
00537   "try",
00538   "trying",
00539   "twice",
00540   "two",
00541   "un",
00542   "under",
00543   "unfortunately",
00544   "unless",
00545   "unlikely",
00546   "until",
00547   "unto",
00548   "up",
00549   "upon",
00550   "us",
00551   "use",
00552   "used",
00553   "useful",
00554   "uses",
00555   "using",
00556   "usually",
00557   "value",
00558   "various",
00559   "very",
00560   "via",
00561   "viz",
00562   "vs",
00563   "want",
00564   "wants",
00565   "was",
00566   "wasn't",
00567   "way",
00568   "we",
00569   "we'd",
00570   "we'll",
00571   "we're",
00572   "we've",
00573   "welcome",
00574   "well",
00575   "went",
00576   "were",
00577   "weren't",
00578   "what",
00579   "what's",
00580   "whatever",
00581   "when",
00582   "whence",
00583   "whenever",
00584   "where",
00585   "where's",
00586   "whereafter",
00587   "whereas",
00588   "whereby",
00589   "wherein",
00590   "whereupon",
00591   "wherever",
00592   "whether",
00593   "which",
00594   "while",
00595   "whither",
00596   "who",
00597   "who's",
00598   "whoever",
00599   "whole",
00600   "whom",
00601   "whose",
00602   "why",
00603   "will",
00604   "willing",
00605   "wish",
00606   "with",
00607   "within",
00608   "without",
00609   "won't",
00610   "wonder",
00611   "would",
00612   "would",
00613   "wouldn't",
00614   "yes",
00615   "yet",
00616   "you",
00617   "you'd",
00618   "you'll",
00619   "you're",
00620   "you've",
00621   "your",
00622   "yours",
00623   "yourself",
00624   "yourselves",
00625   "zero",
00626 #endif
00627 
00628   NULL };
00629 
00630 static int ft_default_parser_parse(MYSQL_FTPARSER_PARAM *param)
00631 {
00632   return param->mysql_parse(param, param->doc, param->length);
00633 }
00634 
00635 struct st_mysql_ftparser ft_default_parser=
00636 {
00637   MYSQL_FTPARSER_INTERFACE_VERSION, ft_default_parser_parse, 0, 0
00638 };
00639
Generated on Sat Sep 9 00:50:30 2006 for MySQL 5.1 by
1.4.7