2 Copyright 2013 Karel Matas
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>.
20 #include "parsers.hxx"
21 #include "sqlite3.hxx"
27 using aoi::ElementKanji
;
28 using aoi::ElementReading
;
29 using aoi::ElementSense
;
33 using utils::to_string
;
35 const char *TEMP_JMDICT
= "gztemp.jmdict";
36 const char *TEMP_KANJIDIC
= "gztemp.kanjidic";
39 * Parses JMDict. File may be GZipped.
40 * \see parsers::JmdictParser
41 * \exception utils::ParsingError
42 * \param fname path to JMDict file
44 void parse_jmdict ( const char *fname
)
46 printf("Loading JMdict file '%s'.\n", fname
);
48 ss
<< "BEGIN TRANSACTION;\n";
52 printf("Decompressing file: %s -> %s\n", fname
, TEMP_JMDICT
);
53 utils::gzip_decompress_file( fname
, TEMP_JMDICT
);
54 parsers::JmdictParser
jmp(TEMP_JMDICT
);
55 const char *SEP
= aoi::SEPARATOR_SQL
;
58 for ( auto &mi
: aoi_config::db_tables
.at("main") ){
59 if ( strncmp( mi
.first
, "d_",2 ) !=0 && strcmp( mi
.first
, "aoi" )!=0 )
61 ss
<< "DROP TABLE IF EXISTS " << mi
.first
<< ";\n"
62 << "CREATE TABLE " << mi
.first
<< " ( ";
63 for ( size_t i
=0; i
<mi
.second
.size(); i
++ )
64 ss
<< mi
.second
[i
].name
<< " " << mi
.second
[i
].type
65 << ((i
==mi
.second
.size()-1) ? "":",");
70 ss
<< "INSERT INTO aoi (key,val) VALUES ('jmdict_version', '"
71 << jmp
.get_version() << "');\n";
72 printf("jmdict_version: %s\n", jmp
.get_version().c_str());
75 for ( std::pair
<string
,string
> elt
: jmp
.get_entities() ){
76 ss
<< "INSERT INTO d_entities (abbr,desc) VALUES ('" << SQLite3::escape(elt
.first
)
77 << "','" << SQLite3::escape(elt
.second
) << "');\n";
81 DicWord entry
= jmp
.get_entry();
82 while ( entry
.did() != -1 ) {
83 for ( ElementReading
&rele
: entry
.r_ele() )
84 ss
<< rele
.sql(entry
.did(),SEP
);
85 for ( ElementKanji
&kele
: entry
.k_ele() )
86 ss
<< kele
.sql(entry
.did(),SEP
);
87 for ( ElementSense
&sele
: entry
.s_ele() )
88 ss
<< sele
.sql(entry
.did(),SEP
);
91 entry
= jmp
.get_entry();
93 printf("%d entries processed.\n", n_entries
);
95 catch ( utils::ParsingError
&e
){
96 std::string msg
= "App::parse_jmdict(): ParsingError: ";
98 printf("parse_jmdict(): ParsingError: %s\n", e
.what() );
102 ss
<< "END TRANSACTION;\n";
103 printf("Writing file 'script.jmdict.sql'...\n");
104 std::ofstream
f ("script.jmdict.sql");
113 * Parses kanjidic2. Works in the same way as parse_jmdict().
114 * \see parsers::KanjidicParser
116 void parse_kanjidic ( const char *fname
)
118 printf("Loading kanjidic file: %s\n", fname
);
119 const char *SEP
= aoi::SEPARATOR_SQL
;
123 printf("Decompressing file: %s -> %s\n", fname
, TEMP_KANJIDIC
);
124 utils::gzip_decompress_file( fname
, TEMP_KANJIDIC
);
125 parsers::KanjidicParser
p(TEMP_KANJIDIC
);
126 auto kanji
= p
.get_entry();
128 std::stringstream ss
;
129 ss
<< "BEGIN TRANSACTION;\n";
132 for ( auto &mi
: aoi_config::db_tables
.at("main") ){
133 if ( strncmp( mi
.first
, "k_",2 ) !=0 )
135 ss
<< "DROP TABLE IF EXISTS " << mi
.first
<< ";\n"
136 << "CREATE TABLE " << mi
.first
<< " ( ";
137 for ( size_t i
=0; i
<mi
.second
.size(); i
++ )
138 ss
<< mi
.second
[i
].name
<< " " << mi
.second
[i
].type
139 << ((i
==mi
.second
.size()-1) ? "":",");
144 ss
<< "REPLACE INTO aoi (key,val) VALUES ('kanjidic_version','"
145 << p
.get_version() << "');\n";
147 while ( kanji
.kanji() != "" ){
149 ss
<< "INSERT INTO k_kanji "
150 << "(kanji,ucs,onyomi,kunyomi,meaning,nanori,flags,jlpt,grade,freq,strokes,"
151 << "rad_classic,rad_nelson,components)"
153 << kanji
.kanji() << "','"
154 << kanji
.ucs() << "','"
155 << to_string(kanji
.onyomi(),SEP
) << "','"
156 << to_string(kanji
.kunyomi(),SEP
) << "','"
157 << SQLite3::escape(to_string(kanji
.meaning(),SEP
)) << "','"
158 << to_string(kanji
.nanori(),SEP
) << "','"
159 << to_string(kanji
.flags(),SEP
) << "',"
160 << kanji
.jlpt() << ","
161 << kanji
.grade() << ","
162 << kanji
.freq() << ","
163 << kanji
.strokes() << ","
164 << kanji
.rad_classic() << ","
165 << kanji
.rad_nelson() << ","
168 for ( aoi::SKIP
&s
: kanji
.skip() ) {
169 ss
<< "INSERT INTO k_skip (kanji,skip1,skip2,skip3,misclass) VALUES('"
170 << kanji
.kanji() << "'," << s
.s1
<< "," << s
.s2
<< "," << s
.s3
171 << ",'" << s
.misclass
<< "');\n";
173 kanji
= p
.get_entry();
175 ss
<< "END TRANSACTION\n;";
177 printf("Writing file 'script.kanjidic.sql'...\n");
178 std::ofstream
f ("script.kanjidic.sql");
182 catch ( utils::ParsingError
&e
){
183 printf("parse_kanjidic(): ParsingError: %s\n", e
.what());
186 remove( TEMP_KANJIDIC
);
193 printf("USAGE: aoi [OPTIONS]\n");
194 printf("OPTIONS:\n");
195 printf(" -geometry W*H+X+Y \n");
196 printf(" -scheme none|GTK+|plastic \n");
197 printf(" -config config string\n");
198 printf(" -parse jmdict|kanjidic parse either jmdict or kanjidic\n");
202 //////////////////////////////////////////////////////////////////////////
205 int main ( int argc
, char **argv
)
209 char **fltk_argv
= &argv
[0];
211 // COmmandline options
213 for ( int i
=1; i
< argc
; i
++ ){
215 if ( !strcmp( argv
[i
], "-config") ){
216 string opts
= argv
[++i
];
217 if ( opts
== "help" ){
218 for ( auto mi
: App::get()->get_config_map() )
219 printf("%s [%s]\n %s\n\n",
220 mi
.first
.c_str(), mi
.second
.val
.c_str(), mi
.second
.desc
.c_str() );
223 for ( string
&s
: utils::split_string( opts
, ":" ) ){
224 vector
<string
> kv
= utils::split_string(s
, "=");
225 App::get()->log("Config override: " + kv
[0] + "=" + kv
[1]);
227 App::get()->config_override( kv
[0], kv
[1] ); }
228 catch ( std::runtime_error
&e
) {
229 App::get()->log_w("Unknown option: " + kv
[0] ); }
231 App::get()->apply_config();
234 else if ( !strcmp( argv
[i
], "-parse" ) ) {
236 printf("-parse: Missing parameter: either kanjidic or jmdict\n");
240 if ( !strcmp( argv
[i
], "jmdict") ){
241 if ( utils::file_exists("JMdict_e.gz") )
242 parse_jmdict( "JMdict_e.gz" );
243 else if ( utils::file_exists("JMdict_e") )
244 parse_jmdict( "JMdict_e" );
246 printf("File not found: JMdict_e or JMdict_e.gz\n");
250 else if ( !strcmp( argv
[i
], "kanjidic") ){
251 if ( utils::file_exists("kanjidic2.xml.gz") )
252 parse_kanjidic( "kanjidic2.xml.gz" );
253 else if ( utils::file_exists("kanjidic2.xml") )
254 parse_kanjidic( "kanjidic2.xml" );
256 printf("File not found: kanjidic2.xml or kanjidic2.xml.gz\n");
261 printf("-parse: wrong parameter '%s'.\n", argv
[i
] );
262 printf("Possible parameters: kanjidic or jmdict\n");
267 printf("argv[%d] : %s -> FLTK\n", i
, argv
[i
]);
268 // pass unparsed arguments to FLTK
269 fltk_argv
[fltk_argc
++] = argv
[i
];
272 ret
= App::get()->run(fltk_argc
,fltk_argv
);
275 catch ( std::exception
&e
){
276 // This should never happen ...
277 string msg
= std::string("Something went wrong...\n")
278 + std::string(typeid(e
).name())
279 + std::string(": ") + std::string(e
.what());
280 App::get()->alert(msg
);