cwchessboard: PgnGrammar.h Source File

Go to the documentation of this file.
00001 // cwchessboard -- A C++ chessboard tool set
00002 //
00003 //! @file PgnGrammar.h This file contains the declaration of class PgnGrammar.
00004 //
00005 // Copyright (C) 2010, by
00006 // 
00007 // Carlo Wood, Run on IRC <carlo@alinoe.com>
00008 // RSA-1024 0x624ACAD5 1997-01-26                    Sign & Encrypt
00009 // Fingerprint16 = 32 EC A7 B6 AC DB 65 A6  F6 F6 55 DD 1C DC FF 61
00010 // 
00011 // This program is free software: you can redistribute it and/or modify
00012 // it under the terms of the GNU General Public License as published by
00013 // the Free Software Foundation, either version 2 of the License, or
00014 // (at your option) any later version.
00015 // 
00016 // This program is distributed in the hope that it will be useful,
00017 // but WITHOUT ANY WARRANTY; without even the implied warranty of
00018 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00019 // GNU General Public License for more details.
00020 // 
00021 // You should have received a copy of the GNU General Public License
00022 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
00023 
00024 #ifndef PGNGRAMMAR_H
00025 #define PGNGRAMMAR_H
00026 
00027 #ifndef USE_PCH
00028 #include "sys.h"
00029 #endif
00030 
00031 #ifndef CWCHESS_PGN_ANALYSIS
00032 #include <boost/spirit/include/classic_spirit.hpp>
00033 #include <boost/spirit/include/classic_numerics.hpp>
00034 #endif
00035 
00036 namespace cwchess {
00037 namespace pgn {
00038 namespace grammar {
00039 
00040 namespace imp {
00041 }
00042 
00043 // Define whether the rules should generate debug output.
00044 #define TRACE_PGN_GRAMMAR 1
00045 #define TRACE_PGN_WHITESPACE 0
00046 #define TRACE_PGN_PRINTABLE 0
00047 #define TRACE_PGN_ANNOTATION 0
00048 #define TRACE_PGN_INTEGER 0
00049 #define TRACE_PGN_COMMENT 0
00050 #define TRACE_PGN_SAN_MOVE 0
00051 #define TRACE_PGN_SAN_INTERNAL 0
00052 #define TRACE_PGN_PUNCTUATION_JUNK 0
00053 
00054 class PgnGrammar : public boost::spirit::classic::grammar<PgnGrammar> {
00055   private:
00056     typedef PgnGrammar grammar_t;
00057 
00058   public:
00059     PgnGrammar(void)
00060     {
00061       BOOST_SPIRIT_DEBUG_TRACE_GRAMMAR_NAME(*this, "PgnGrammar", 1);
00062     }
00063 
00064   public:
00065     template <typename ScannerT>
00066     struct definition {
00067       // Non-parse_tree generating rule type.
00068       typedef typename ScannerT::iteration_policy_t iteration_policy_t;
00069       typedef boost::spirit::classic::match_policy match_policy_t;
00070       typedef typename ScannerT::action_policy_t action_policy_t;
00071       typedef boost::spirit::classic::scanner_policies<iteration_policy_t, match_policy_t, action_policy_t> policies_t;
00072       typedef boost::spirit::classic::scanner<typename ScannerT::iterator_t, policies_t> non_tree_scanner_t;
00073       typedef boost::spirit::classic::rule<non_tree_scanner_t> no_tree_rule_t;
00074       // Normal (parse_tree generating) rule type.
00075       typedef boost::spirit::classic::rule<ScannerT> rule_t;
00076       // An integer parser for unsigned integers of 1 to 3 digits. 
00077       boost::spirit::classic::uint_parser<unsigned, 10, 1, 3> uint3_p;
00078 
00079       // RULE DECLARATION START
00080       rule_t leading_junk;
00081       rule_t file;
00082       rule_t rank;
00083       rule_t piece;
00084       rule_t check;
00085       rule_t punctuation_junk;
00086       rule_t tagname_begin;
00087       rule_t tagname_continuation;
00088       rule_t printable;
00089       rule_t printable_string;
00090       rule_t printable_comment;
00091       rule_t promotion;
00092       no_tree_rule_t blank;
00093       no_tree_rule_t white_space;
00094       no_tree_rule_t escaped_data;
00095       no_tree_rule_t eol;
00096       rule_t integer;
00097       rule_t square;
00098       rule_t disambiguation;
00099       rule_t string;
00100       rule_t tag_name;
00101       rule_t tag_value;
00102       rule_t game_termination;
00103       rule_t elipsis;
00104       rule_t move_number_indication;
00105       rule_t SAN_move;
00106       rule_t annotation;
00107       rule_t numeric_annotation_glyph;
00108       rule_t text_annotation1_junk;
00109       rule_t text_annotation2_junk;
00110       rule_t comment;
00111       rule_t parenthesis_junk;
00112       rule_t tag_pair;
00113       rule_t tag_section;
00114       rule_t move;
00115       rule_t white_move;
00116       rule_t last_white_move;
00117       rule_t arena_junk;
00118       rule_t black_move;
00119       rule_t movetext_section_white;
00120       rule_t movetext_section_black;
00121       rule_t white_black_move;
00122       rule_t recursive_variation_white;
00123       rule_t recursive_variation_black;
00124       rule_t wrong_color_white_junk;
00125       rule_t wrong_color_black_junk;
00126       rule_t movetext_section;
00127       rule_t PGN_game;
00128       rule_t PGN_database;
00129       // RULE DECLARATION END
00130 
00131       // Start rule of this grammar.
00132       rule_t const& start(void) const { return PGN_database; }
00133 
00134       definition(PgnGrammar const& self)
00135       {
00136         using namespace boost::spirit::classic;
00137 
00138         leading_junk
00139             =
00140                * ((*~ch_p('\n') >>* ch_p('\n')) - tag_pair)
00141             ;
00142 
00143         file                    // TRACE_PGN_SAN_INTERNAL
00144             =
00145                 range_p('a', 'h')
00146             ;
00147 
00148         rank                    // TRACE_PGN_SAN_INTERNAL
00149             =
00150                 range_p('1', '8')
00151             ;
00152 
00153         piece                   // TRACE_PGN_SAN_INTERNAL
00154             =
00155                 ch_p('R')
00156             |   ch_p('N')
00157             |   ch_p('B')
00158             |   ch_p('Q')
00159             |   ch_p('K')
00160             ;
00161 
00162         check                   // TRACE_PGN_SAN_INTERNAL
00163             =
00164                 ch_p('+')
00165             |   ch_p('#')
00166             ;
00167 
00168         punctuation_junk        // TRACE_PGN_PUNCTUATION_JUNK
00169             =
00170                 ch_p(',')
00171             |   ch_p(';')
00172             ;
00173 
00174         tagname_begin
00175             =
00176                 alnum_p
00177             ;
00178 
00179         tagname_continuation
00180             =
00181                 alnum_p
00182             |   ch_p('_')
00183             ;
00184 
00185         printable               // TRACE_PGN_PRINTABLE
00186             =
00187                 range_p(32, 126)
00188             |   range_p(-96, -1)
00189             ;
00190 
00191         printable_string        // TRACE_PGN_PRINTABLE
00192             =
00193                 (   range_p(35, 91)
00194                 |   ch_p(' ')
00195                 |   range_p(93, 126)
00196                 |   ch_p('!')
00197                 |   range_p(-96, -1)
00198                 )
00199             |   (   ch_p('\\') >> (ch_p('\\') | ch_p('"')))
00200             ;
00201 
00202         printable_comment       // TRACE_PGN_PRINTABLE
00203             =
00204                 range_p(32, 124)
00205             |   ch_p('~')
00206             |   blank
00207             |   eol
00208             |   range_p(-96, -1)
00209             ;
00210 
00211         promotion               // TRACE_PGN_SAN_INTERNAL
00212             =
00213                 ch_p('=') >> piece
00214             ;
00215 
00216         blank                   // TRACE_PGN_WHITESPACE
00217             =
00218                 ch_p(' ')
00219             |   ch_p('\t')
00220             |   ch_p('\v')
00221             ;
00222         
00223         white_space             // TRACE_PGN_WHITESPACE
00224             =
00225                 blank >> !white_space
00226             |   eol >> !white_space
00227             ;
00228 
00229         escaped_data            // TRACE_PGN_WHITESPACE
00230             =
00231                 ch_p('%') >>* (anychar_p - eol_p) >> eol_p
00232             ;
00233 
00234         eol                     // TRACE_PGN_WHITESPACE
00235             =
00236                 eol_p >>* escaped_data
00237             ;
00238 
00239         integer                 // TRACE_PGN_INTEGER
00240             =
00241                 uint3_p
00242             ;
00243 
00244         square                  // TRACE_PGN_SAN_INTERNAL
00245             =
00246                 file >> rank
00247             ;
00248 
00249         disambiguation          // TRACE_PGN_SAN_INTERNAL
00250             =
00251                 file
00252             |   rank
00253             |   square
00254             ;
00255 
00256         string
00257             =
00258                 ch_p('"') >>* printable_string >> ch_p('"')
00259             ;
00260 
00261         tag_name
00262             =
00263                 tagname_begin >>* tagname_continuation
00264             ;
00265 
00266         tag_value
00267             =
00268                 string
00269             ;
00270 
00271         game_termination
00272             =
00273                 str_p("1-0")
00274             |   str_p("0-1")
00275             |   str_p("1/2-1/2")
00276             |   ch_p('*')
00277             ;
00278 
00279         elipsis
00280             =
00281                 str_p("..") >> !ch_p('.')
00282             ;
00283 
00284         move_number_indication
00285             =
00286                 integer >> (ch_p('.') |  ch_p(' '))
00287             ;
00288 
00289         SAN_move                // TRACE_PGN_SAN_MOVE
00290             =
00291                 (   piece >> !ch_p('x') >> square
00292                 |   piece >> disambiguation >> !ch_p('x') >> square
00293                 |   file >> ch_p('x') >> square >> !promotion
00294                 |   square >> !promotion
00295                 |   str_p("O-O-O")
00296                 |   str_p("O-O")
00297                 |   str_p("0-0-0")
00298                 |   str_p("0-0")
00299                 )
00300                 >>  !check
00301             ;
00302 
00303         annotation
00304             =
00305                 !(text_annotation1_junk >> !white_space) >>* (!comment >> (numeric_annotation_glyph | text_annotation2_junk) >> !white_space)
00306             ;
00307 
00308         numeric_annotation_glyph// TRACE_PGN_ANNOTATION
00309             =
00310                 ch_p('$') >> uint3_p
00311             ;
00312 
00313         text_annotation1_junk   // TRACE_PGN_ANNOTATION
00314             =
00315                 (   ch_p('!')
00316                 |   ch_p('?')
00317                 ) >>
00318                !(   ch_p('!')
00319                 |   ch_p('?')
00320                 )
00321             ;
00322 
00323         text_annotation2_junk   // TRACE_PGN_ANNOTATION
00324             =
00325                 ch_p('=') >> !(   ch_p('-')
00326                               |   ch_p('+')
00327                               )
00328             |   ch_p('-') >>  (    ch_p('=')
00329                               |    (!ch_p('/') >> ch_p('+'))
00330                               )
00331             |   ch_p('+') >>  (    (!ch_p('/') >> ch_p('-'))
00332                               | ch_p('=')
00333                               )
00334             |   ch_p('~')
00335             ;
00336 
00337         comment                 // TRACE_PGN_COMMENT
00338             =
00339                 ch_p('{') >>* printable_comment >> ch_p('}') >> !white_space >> !comment
00340             |   ch_p(';') >>* printable >> eol >> !white_space >> !comment
00341             |   parenthesis_junk >> !white_space >> !comment
00342             ;
00343 
00344         // Allow '(...)' to be be empty, if it contains a brace comment.
00345         parenthesis_junk        // TRACE_PGN_COMMENT
00346             =
00347                 ch_p('(') >> !white_space >> ch_p('{') >>* printable_comment >> ch_p('}') >> !white_space >> ch_p(')')
00348             ;
00349 
00350         tag_pair
00351             =
00352                 ch_p('[') >> !white_space >> tag_name >> !white_space >> tag_value >> !white_space >> ch_p(']') >> !white_space
00353             ;
00354 
00355         tag_section
00356             =
00357                 tag_pair >> !tag_section
00358             ;
00359 
00360         // Allow a SAN move to be directly followed by some punctuation,
00361         // although such punctuation should really be part of a comment.
00362         // Most notably this will cause a semicolon to NOT result the
00363         // rest of line to be considered as a comment if the semi colon
00364         // directly follows a SAN move.
00365         //
00366         // For example:
00367         //
00368         // {At which point he did} 11. b4; {and went to the bathroom.} 11. ... Rb8
00369         //
00370         // will be read as:
00371         //
00372         // {At which point he did} 11. b4 {; and went to the bathroom.} 11. ... Rb8
00373         //
00374         // Rather than,
00375         //
00376         // {At which point he did} 11. b4 { and went to the bathroom. 11. ... Rb8}
00377         move
00378             =
00379                 SAN_move >> !punctuation_junk >> !white_space >> !(annotation >> !white_space)
00380             ;
00381 
00382         white_move
00383             =
00384                 move_number_indication >> !white_space >> move >> !comment >>* recursive_variation_white
00385             ;
00386 
00387         last_white_move
00388             =
00389                 white_move
00390             |   arena_junk
00391             ;
00392 
00393         arena_junk
00394             =
00395                 move_number_indication >> !white_space >> elipsis >> !white_space >> comment
00396             ;
00397 
00398         black_move
00399             =
00400                 move_number_indication >> !white_space >> elipsis >> !white_space >> move >> !comment >>* recursive_variation_black
00401             ;
00402 
00403         movetext_section_white
00404             =
00405                 white_black_move >>* white_black_move >> !last_white_move
00406             |   last_white_move
00407             ;
00408 
00409         movetext_section_black
00410             =
00411                 black_move >> !movetext_section_white
00412             ;
00413 
00414         white_black_move
00415             =
00416                 move_number_indication >> !white_space >> move >> !comment >> move >> !comment >>* recursive_variation_black
00417             |   white_move >> black_move
00418             ;
00419 
00420         recursive_variation_white
00421             =
00422                 ch_p('(') >> !white_space >> !comment >> (movetext_section_white | wrong_color_black_junk) >> ch_p(')') >> !white_space >> !comment
00423             ;
00424 
00425         recursive_variation_black
00426             =
00427                 ch_p('(') >> !white_space >> !comment >> (movetext_section_black | wrong_color_white_junk) >> ch_p(')') >> !white_space >> !comment
00428             ;
00429 
00430         wrong_color_white_junk
00431             =
00432                 movetext_section_white
00433             ;
00434 
00435         wrong_color_black_junk
00436             =
00437                 movetext_section_black
00438             ;
00439 
00440         movetext_section
00441             =
00442                 !comment >> (movetext_section_white | movetext_section_black)
00443             ;
00444 
00445         PGN_game
00446             =
00447                 !tag_section >> !movetext_section >> game_termination >> !white_space >> !comment
00448             ;
00449 
00450         PGN_database
00451             =
00452                 !white_space >> leading_junk >>* PGN_game;
00453             ;
00454 
00455         // DEBUG MACROS START
00456         BOOST_SPIRIT_DEBUG_TRACE_RULE(leading_junk, TRACE_PGN_GRAMMAR);
00457         BOOST_SPIRIT_DEBUG_TRACE_RULE(file, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00458         BOOST_SPIRIT_DEBUG_TRACE_RULE(rank, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00459         BOOST_SPIRIT_DEBUG_TRACE_RULE(piece, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00460         BOOST_SPIRIT_DEBUG_TRACE_RULE(check, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00461         BOOST_SPIRIT_DEBUG_TRACE_RULE(punctuation_junk, TRACE_PGN_GRAMMAR & TRACE_PGN_PUNCTUATION_JUNK);
00462         BOOST_SPIRIT_DEBUG_TRACE_RULE(tagname_begin, TRACE_PGN_GRAMMAR);
00463         BOOST_SPIRIT_DEBUG_TRACE_RULE(tagname_continuation, TRACE_PGN_GRAMMAR);
00464         BOOST_SPIRIT_DEBUG_TRACE_RULE(printable, TRACE_PGN_GRAMMAR & TRACE_PGN_PRINTABLE);
00465         BOOST_SPIRIT_DEBUG_TRACE_RULE(printable_string, TRACE_PGN_GRAMMAR & TRACE_PGN_PRINTABLE);
00466         BOOST_SPIRIT_DEBUG_TRACE_RULE(printable_comment, TRACE_PGN_GRAMMAR & TRACE_PGN_PRINTABLE);
00467         BOOST_SPIRIT_DEBUG_TRACE_RULE(promotion, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00468         BOOST_SPIRIT_DEBUG_TRACE_RULE(blank, TRACE_PGN_GRAMMAR & TRACE_PGN_WHITESPACE);
00469         BOOST_SPIRIT_DEBUG_TRACE_RULE(white_space, TRACE_PGN_GRAMMAR & TRACE_PGN_WHITESPACE);
00470         BOOST_SPIRIT_DEBUG_TRACE_RULE(escaped_data, TRACE_PGN_GRAMMAR & TRACE_PGN_WHITESPACE);
00471         BOOST_SPIRIT_DEBUG_TRACE_RULE(eol, TRACE_PGN_GRAMMAR & TRACE_PGN_WHITESPACE);
00472         BOOST_SPIRIT_DEBUG_TRACE_RULE(integer, TRACE_PGN_GRAMMAR & TRACE_PGN_INTEGER);
00473         BOOST_SPIRIT_DEBUG_TRACE_RULE(square, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00474         BOOST_SPIRIT_DEBUG_TRACE_RULE(disambiguation, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_INTERNAL);
00475         BOOST_SPIRIT_DEBUG_TRACE_RULE(string, TRACE_PGN_GRAMMAR);
00476         BOOST_SPIRIT_DEBUG_TRACE_RULE(tag_name, TRACE_PGN_GRAMMAR);
00477         BOOST_SPIRIT_DEBUG_TRACE_RULE(tag_value, TRACE_PGN_GRAMMAR);
00478         BOOST_SPIRIT_DEBUG_TRACE_RULE(game_termination, TRACE_PGN_GRAMMAR);
00479         BOOST_SPIRIT_DEBUG_TRACE_RULE(elipsis, TRACE_PGN_GRAMMAR);
00480         BOOST_SPIRIT_DEBUG_TRACE_RULE(move_number_indication, TRACE_PGN_GRAMMAR);
00481         BOOST_SPIRIT_DEBUG_TRACE_RULE(SAN_move, TRACE_PGN_GRAMMAR & TRACE_PGN_SAN_MOVE);
00482         BOOST_SPIRIT_DEBUG_TRACE_RULE(annotation, TRACE_PGN_GRAMMAR);
00483         BOOST_SPIRIT_DEBUG_TRACE_RULE(numeric_annotation_glyph, TRACE_PGN_GRAMMAR & TRACE_PGN_ANNOTATION);
00484         BOOST_SPIRIT_DEBUG_TRACE_RULE(text_annotation1_junk, TRACE_PGN_GRAMMAR & TRACE_PGN_ANNOTATION);
00485         BOOST_SPIRIT_DEBUG_TRACE_RULE(text_annotation2_junk, TRACE_PGN_GRAMMAR & TRACE_PGN_ANNOTATION);
00486         BOOST_SPIRIT_DEBUG_TRACE_RULE(comment, TRACE_PGN_GRAMMAR & TRACE_PGN_COMMENT);
00487         BOOST_SPIRIT_DEBUG_TRACE_RULE(parenthesis_junk, TRACE_PGN_GRAMMAR & TRACE_PGN_COMMENT);
00488         BOOST_SPIRIT_DEBUG_TRACE_RULE(tag_pair, TRACE_PGN_GRAMMAR);
00489         BOOST_SPIRIT_DEBUG_TRACE_RULE(tag_section, TRACE_PGN_GRAMMAR);
00490         BOOST_SPIRIT_DEBUG_TRACE_RULE(move, TRACE_PGN_GRAMMAR);
00491         BOOST_SPIRIT_DEBUG_TRACE_RULE(white_move, TRACE_PGN_GRAMMAR);
00492         BOOST_SPIRIT_DEBUG_TRACE_RULE(last_white_move, TRACE_PGN_GRAMMAR);
00493         BOOST_SPIRIT_DEBUG_TRACE_RULE(arena_junk, TRACE_PGN_GRAMMAR);
00494         BOOST_SPIRIT_DEBUG_TRACE_RULE(black_move, TRACE_PGN_GRAMMAR);
00495         BOOST_SPIRIT_DEBUG_TRACE_RULE(movetext_section_white, TRACE_PGN_GRAMMAR);
00496         BOOST_SPIRIT_DEBUG_TRACE_RULE(movetext_section_black, TRACE_PGN_GRAMMAR);
00497         BOOST_SPIRIT_DEBUG_TRACE_RULE(white_black_move, TRACE_PGN_GRAMMAR);
00498         BOOST_SPIRIT_DEBUG_TRACE_RULE(recursive_variation_white, TRACE_PGN_GRAMMAR);
00499         BOOST_SPIRIT_DEBUG_TRACE_RULE(recursive_variation_black, TRACE_PGN_GRAMMAR);
00500         BOOST_SPIRIT_DEBUG_TRACE_RULE(wrong_color_white_junk, TRACE_PGN_GRAMMAR);
00501         BOOST_SPIRIT_DEBUG_TRACE_RULE(wrong_color_black_junk, TRACE_PGN_GRAMMAR);
00502         BOOST_SPIRIT_DEBUG_TRACE_RULE(movetext_section, TRACE_PGN_GRAMMAR);
00503         BOOST_SPIRIT_DEBUG_TRACE_RULE(PGN_game, TRACE_PGN_GRAMMAR);
00504         BOOST_SPIRIT_DEBUG_TRACE_RULE(PGN_database, TRACE_PGN_GRAMMAR);
00505         // DEBUG MACROS END
00506       }
00507     };
00508 };
00509 
00510 } // namespace grammar
00511 } // namespace pgn
00512 } // namespace cwchess
00513 
00514 #endif  // PGNGRAMMAR_H