diff options
Diffstat (limited to 'generator.cpp')
-rw-r--r-- | generator.cpp | 1520 |
1 files changed, 0 insertions, 1520 deletions
diff --git a/generator.cpp b/generator.cpp deleted file mode 100644 index 305d121..0000000 --- a/generator.cpp +++ /dev/null | |||
@@ -1,1520 +0,0 @@ | |||
1 | #include <libxml/parser.h> | ||
2 | #include <iostream> | ||
3 | #include <dirent.h> | ||
4 | #include <set> | ||
5 | #include <map> | ||
6 | #include <string> | ||
7 | #include <vector> | ||
8 | #include <fstream> | ||
9 | #include <sqlite3.h> | ||
10 | #include <sstream> | ||
11 | #include <regex> | ||
12 | #include <list> | ||
13 | #include "progress.h" | ||
14 | |||
15 | struct verb { | ||
16 | std::string infinitive; | ||
17 | std::string past_tense; | ||
18 | std::string past_participle; | ||
19 | std::string ing_form; | ||
20 | std::string s_form; | ||
21 | }; | ||
22 | |||
23 | struct adjective { | ||
24 | std::string base; | ||
25 | std::string comparative; | ||
26 | std::string superlative; | ||
27 | }; | ||
28 | |||
29 | struct noun { | ||
30 | std::string singular; | ||
31 | std::string plural; | ||
32 | }; | ||
33 | |||
34 | struct group { | ||
35 | std::string id; | ||
36 | std::set<std::string> members; | ||
37 | }; | ||
38 | |||
39 | std::map<std::string, group> groups; | ||
40 | std::map<std::string, verb> verbs; | ||
41 | std::map<std::string, adjective> adjectives; | ||
42 | std::map<std::string, noun> nouns; | ||
43 | std::map<int, std::map<int, int>> wn; | ||
44 | std::map<std::string, std::set<std::string>> pronunciations; | ||
45 | |||
46 | void print_usage() | ||
47 | { | ||
48 | std::cout << "Verbly Datafile Generator" << std::endl; | ||
49 | std::cout << "-------------------------" << std::endl; | ||
50 | std::cout << "Requires exactly six arguments." << std::endl; | ||
51 | std::cout << "1. The path to a VerbNet data directory." << std::endl; | ||
52 | std::cout << "2. The path to a SemLink vnpbMappings file." << std::endl; | ||
53 | std::cout << "3. The path to an AGID infl.txt file." << std::endl; | ||
54 | std::cout << "4. The path to a WordNet prolog data directory." << std::endl; | ||
55 | std::cout << "5. The path to a CMUDICT pronunciation file." << std::endl; | ||
56 | std::cout << "6. Datafile output path." << std::endl; | ||
57 | |||
58 | exit(1); | ||
59 | } | ||
60 | |||
61 | void db_error(sqlite3* ppdb, std::string) | ||
62 | { | ||
63 | std::cout << "Error writing to output database: " << sqlite3_errmsg(ppdb) << std::endl; | ||
64 | sqlite3_close_v2(ppdb); | ||
65 | print_usage(); | ||
66 | } | ||
67 | |||
68 | /* | ||
69 | void parse_group(xmlNodePtr top, std::string filename) | ||
70 | { | ||
71 | xmlChar* key = xmlGetProp(top, (xmlChar*) "ID"); | ||
72 | if (key == 0) | ||
73 | { | ||
74 | std::cout << "Bad VerbNet file format: " << filename << std::endl; | ||
75 | print_usage(); | ||
76 | } | ||
77 | std::string vnid = key; | ||
78 | vnid = vnid.substr(vnid.find_first_of("-")+1); | ||
79 | xmlFree(key); | ||
80 | |||
81 | group g; | ||
82 | g.id = vnid; | ||
83 | |||
84 | for (xmlNodePtr node = top->xmlChildrenNode; node != nullptr; node = node->next) | ||
85 | { | ||
86 | if (!xmlStrcmp(node->name, (const xmlChar*) "MEMBERS")) | ||
87 | { | ||
88 | for (xmlNodePtr member = node->xmlChildrenNode; member != nullptr; member = member->next) | ||
89 | { | ||
90 | if (!xmlStrcmp(member->name, (const xmlChar*) "MEMBER")) | ||
91 | { | ||
92 | key = xmlGetProp(member, (xmlChar*) "name"); | ||
93 | g.members.insert(key); | ||
94 | xmlFree(key); | ||
95 | } | ||
96 | } | ||
97 | } else if (!xmlStrcmp(node->name, (const xmlChar*) "FRAMES")) | ||
98 | { | ||
99 | for (xmlNodePtr frame = node->xmlChildrenNode; frame != nullptr; frame = frame->next) | ||
100 | { | ||
101 | if (!xmlStrcmp(frame->name, (const xmlChar*) "FRAME")) | ||
102 | { | ||
103 | for (xmlNodePtr framenode = frame->xmlChildrenNode; framenode != nullptr; framenode = framenode->next) | ||
104 | { | ||
105 | |||
106 | } | ||
107 | } | ||
108 | } | ||
109 | } | ||
110 | } | ||
111 | }*/ | ||
112 | |||
113 | int main(int argc, char** argv) | ||
114 | { | ||
115 | if (argc != 7) | ||
116 | { | ||
117 | print_usage(); | ||
118 | } | ||
119 | |||
120 | /*DIR* dir; | ||
121 | if ((dir = opendir(argv[1])) == nullptr) | ||
122 | { | ||
123 | std::cout << "Invalid VerbNet data directory." << std::endl; | ||
124 | |||
125 | print_usage(); | ||
126 | } | ||
127 | |||
128 | struct dirent* ent; | ||
129 | while ((ent = readdir(dir)) != nullptr) | ||
130 | { | ||
131 | std::string filename(argv[1]); | ||
132 | if (filename.back() != '/') | ||
133 | { | ||
134 | filename += '/'; | ||
135 | } | ||
136 | |||
137 | filename += ent->d_name; | ||
138 | //std::cout << ent->d_name << std::endl; | ||
139 | |||
140 | if (filename.rfind(".xml") != filename.size() - 4) | ||
141 | { | ||
142 | continue; | ||
143 | } | ||
144 | |||
145 | xmlDocPtr doc = xmlParseFile(filename.c_str()); | ||
146 | if (doc == nullptr) | ||
147 | { | ||
148 | std::cout << "Error opening " << filename << std::endl; | ||
149 | print_usage(); | ||
150 | } | ||
151 | |||
152 | xmlNodePtr top = xmlDocGetRootElement(doc); | ||
153 | if ((top == nullptr) || (xmlStrcmp(top->name, (xmlChar*) "VNCLASS"))) | ||
154 | { | ||
155 | std::cout << "Bad VerbNet file format: " << filename << std::endl; | ||
156 | print_usage(); | ||
157 | } | ||
158 | |||
159 | parse_group(top, filename); | ||
160 | } | ||
161 | |||
162 | closedir(dir);*/ | ||
163 | |||
164 | // Get verbs from AGID | ||
165 | std::cout << "Reading inflections..." << std::endl; | ||
166 | |||
167 | std::ifstream agidfile(argv[3]); | ||
168 | if (!agidfile.is_open()) | ||
169 | { | ||
170 | std::cout << "Could not open AGID file: " << argv[3] << std::endl; | ||
171 | print_usage(); | ||
172 | } | ||
173 | |||
174 | for (;;) | ||
175 | { | ||
176 | std::string line; | ||
177 | if (!getline(agidfile, line)) | ||
178 | { | ||
179 | break; | ||
180 | } | ||
181 | |||
182 | if (line.back() == '\r') | ||
183 | { | ||
184 | line.pop_back(); | ||
185 | } | ||
186 | |||
187 | int divider = line.find_first_of(" "); | ||
188 | std::string word = line.substr(0, divider); | ||
189 | line = line.substr(divider+1); | ||
190 | char type = line[0]; | ||
191 | |||
192 | if (line[1] == '?') | ||
193 | { | ||
194 | line.erase(0, 4); | ||
195 | } else { | ||
196 | line.erase(0, 3); | ||
197 | } | ||
198 | |||
199 | std::vector<std::string> forms; | ||
200 | while (!line.empty()) | ||
201 | { | ||
202 | std::string inflection; | ||
203 | if ((divider = line.find(" | ")) != std::string::npos) | ||
204 | { | ||
205 | inflection = line.substr(0, divider); | ||
206 | line = line.substr(divider + 3); | ||
207 | } else { | ||
208 | inflection = line; | ||
209 | line = ""; | ||
210 | } | ||
211 | |||
212 | if ((divider = inflection.find_first_of(",?")) != std::string::npos) | ||
213 | { | ||
214 | inflection = inflection.substr(0, divider); | ||
215 | } | ||
216 | |||
217 | forms.push_back(inflection); | ||
218 | } | ||
219 | |||
220 | switch (type) | ||
221 | { | ||
222 | case 'V': | ||
223 | { | ||
224 | verb v; | ||
225 | v.infinitive = word; | ||
226 | if (forms.size() == 4) | ||
227 | { | ||
228 | v.past_tense = forms[0]; | ||
229 | v.past_participle = forms[1]; | ||
230 | v.ing_form = forms[2]; | ||
231 | v.s_form = forms[3]; | ||
232 | } else if (forms.size() == 3) | ||
233 | { | ||
234 | v.past_tense = forms[0]; | ||
235 | v.past_participle = forms[0]; | ||
236 | v.ing_form = forms[1]; | ||
237 | v.s_form = forms[2]; | ||
238 | } else if (forms.size() == 8) | ||
239 | { | ||
240 | // As of AGID 2014.08.11, this is only "to be" | ||
241 | v.past_tense = forms[0]; | ||
242 | v.past_participle = forms[2]; | ||
243 | v.ing_form = forms[3]; | ||
244 | v.s_form = forms[4]; | ||
245 | } else { | ||
246 | // Words that don't fit the cases above as of AGID 2014.08.11: | ||
247 | // - may and shall do not conjugate the way we want them to | ||
248 | // - methinks only has a past tense and is an outlier | ||
249 | // - wit has five forms, and is archaic/obscure enough that we can ignore it for now | ||
250 | std::cout << "Ignoring verb \"" << word << "\" due to non-standard number of forms." << std::endl; | ||
251 | } | ||
252 | |||
253 | verbs[word] = v; | ||
254 | |||
255 | break; | ||
256 | } | ||
257 | |||
258 | case 'A': | ||
259 | { | ||
260 | adjective adj; | ||
261 | adj.base = word; | ||
262 | if (forms.size() == 2) | ||
263 | { | ||
264 | adj.comparative = forms[0]; | ||
265 | adj.superlative = forms[1]; | ||
266 | } else { | ||
267 | // As of AGID 2014.08.11, this is only "only", which has only the form "onliest" | ||
268 | std::cout << "Ignoring adjective/adverb \"" << word << "\" due to non-standard number of forms." << std::endl; | ||
269 | } | ||
270 | |||
271 | adjectives[word] = adj; | ||
272 | |||
273 | break; | ||
274 | } | ||
275 | |||
276 | case 'N': | ||
277 | { | ||
278 | noun n; | ||
279 | n.singular = word; | ||
280 | if (forms.size() == 1) | ||
281 | { | ||
282 | n.plural = forms[0]; | ||
283 | } else { | ||
284 | // As of AGID 2014.08.11, this is non-existent. | ||
285 | std::cout << "Ignoring noun \"" << word << "\" due to non-standard number of forms." << std::endl; | ||
286 | } | ||
287 | |||
288 | nouns[word] = n; | ||
289 | |||
290 | break; | ||
291 | } | ||
292 | } | ||
293 | } | ||
294 | |||
295 | // Pronounciations | ||
296 | std::cout << "Reading pronunciations..." << std::endl; | ||
297 | |||
298 | std::ifstream pronfile(argv[5]); | ||
299 | if (!pronfile.is_open()) | ||
300 | { | ||
301 | std::cout << "Could not open CMUDICT file: " << argv[5] << std::endl; | ||
302 | print_usage(); | ||
303 | } | ||
304 | |||
305 | for (;;) | ||
306 | { | ||
307 | std::string line; | ||
308 | if (!getline(pronfile, line)) | ||
309 | { | ||
310 | break; | ||
311 | } | ||
312 | |||
313 | if (line.back() == '\r') | ||
314 | { | ||
315 | line.pop_back(); | ||
316 | } | ||
317 | |||
318 | std::regex phoneme("([A-Z][^ \\(]*)(?:\\(\\d+\\))? ([A-Z 0-9]+)"); | ||
319 | std::smatch phoneme_data; | ||
320 | if (std::regex_search(line, phoneme_data, phoneme)) | ||
321 | { | ||
322 | std::string canonical(phoneme_data[1]); | ||
323 | std::transform(std::begin(canonical), std::end(canonical), std::begin(canonical), ::tolower); | ||
324 | |||
325 | pronunciations[canonical].insert(phoneme_data[2]); | ||
326 | } | ||
327 | } | ||
328 | |||
329 | // Start writing output | ||
330 | std::cout << "Writing schema..." << std::endl; | ||
331 | |||
332 | sqlite3* ppdb; | ||
333 | if (sqlite3_open_v2(argv[6], &ppdb, SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE, NULL) != SQLITE_OK) | ||
334 | { | ||
335 | std::cout << "Error opening output datafile: " << sqlite3_errmsg(ppdb) << std::endl; | ||
336 | print_usage(); | ||
337 | } | ||
338 | |||
339 | std::ifstream schemafile("schema.sql"); | ||
340 | if (!schemafile.is_open()) | ||
341 | { | ||
342 | std::cout << "Could not find schema file" << std::endl; | ||
343 | print_usage(); | ||
344 | } | ||
345 | |||
346 | std::stringstream schemabuilder; | ||
347 | for (;;) | ||
348 | { | ||
349 | std::string line; | ||
350 | if (!getline(schemafile, line)) | ||
351 | { | ||
352 | break; | ||
353 | } | ||
354 | |||
355 | if (line.back() == '\r') | ||
356 | { | ||
357 | line.pop_back(); | ||
358 | } | ||
359 | |||
360 | schemabuilder << line << std::endl; | ||
361 | } | ||
362 | |||
363 | std::string schema = schemabuilder.str(); | ||
364 | while (!schema.empty()) | ||
365 | { | ||
366 | std::string query; | ||
367 | int divider = schema.find(";"); | ||
368 | if (divider != std::string::npos) | ||
369 | { | ||
370 | query = schema.substr(0, divider+1); | ||
371 | schema = schema.substr(divider+2); | ||
372 | } else { | ||
373 | break; | ||
374 | } | ||
375 | |||
376 | sqlite3_stmt* schmstmt; | ||
377 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &schmstmt, NULL) != SQLITE_OK) | ||
378 | { | ||
379 | db_error(ppdb, query); | ||
380 | } | ||
381 | |||
382 | if (sqlite3_step(schmstmt) != SQLITE_DONE) | ||
383 | { | ||
384 | db_error(ppdb, query); | ||
385 | } | ||
386 | |||
387 | sqlite3_finalize(schmstmt); | ||
388 | } | ||
389 | |||
390 | { | ||
391 | progress ppgs("Writing verbs...", verbs.size()); | ||
392 | for (auto& mapping : verbs) | ||
393 | { | ||
394 | sqlite3_stmt* ppstmt; | ||
395 | std::string query("INSERT INTO verbs (infinitive, past_tense, past_participle, ing_form, s_form) VALUES (?, ?, ?, ?, ?)"); | ||
396 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
397 | { | ||
398 | db_error(ppdb, query); | ||
399 | } | ||
400 | |||
401 | sqlite3_bind_text(ppstmt, 1, mapping.second.infinitive.c_str(), mapping.second.infinitive.length(), SQLITE_STATIC); | ||
402 | sqlite3_bind_text(ppstmt, 2, mapping.second.past_tense.c_str(), mapping.second.past_tense.length(), SQLITE_STATIC); | ||
403 | sqlite3_bind_text(ppstmt, 3, mapping.second.past_participle.c_str(), mapping.second.past_participle.length(), SQLITE_STATIC); | ||
404 | sqlite3_bind_text(ppstmt, 4, mapping.second.ing_form.c_str(), mapping.second.ing_form.length(), SQLITE_STATIC); | ||
405 | sqlite3_bind_text(ppstmt, 5, mapping.second.s_form.c_str(), mapping.second.s_form.length(), SQLITE_STATIC); | ||
406 | |||
407 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
408 | { | ||
409 | db_error(ppdb, query); | ||
410 | } | ||
411 | |||
412 | sqlite3_finalize(ppstmt); | ||
413 | |||
414 | std::string canonical(mapping.second.infinitive); | ||
415 | std::transform(std::begin(canonical), std::end(canonical), std::begin(canonical), ::tolower); | ||
416 | if (pronunciations.count(canonical) == 1) | ||
417 | { | ||
418 | query = "SELECT last_insert_rowid()"; | ||
419 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
420 | { | ||
421 | db_error(ppdb, query); | ||
422 | } | ||
423 | |||
424 | if (sqlite3_step(ppstmt) != SQLITE_ROW) | ||
425 | { | ||
426 | db_error(ppdb, query); | ||
427 | } | ||
428 | |||
429 | int rowid = sqlite3_column_int(ppstmt, 0); | ||
430 | |||
431 | sqlite3_finalize(ppstmt); | ||
432 | |||
433 | for (auto pronunciation : pronunciations[canonical]) | ||
434 | { | ||
435 | query = "INSERT INTO verb_pronunciations (verb_id, pronunciation) VALUES (?, ?)"; | ||
436 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
437 | { | ||
438 | db_error(ppdb, query); | ||
439 | } | ||
440 | |||
441 | sqlite3_bind_int(ppstmt, 1, rowid); | ||
442 | sqlite3_bind_text(ppstmt, 2, pronunciation.c_str(), pronunciation.length(), SQLITE_STATIC); | ||
443 | |||
444 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
445 | { | ||
446 | db_error(ppdb, query); | ||
447 | } | ||
448 | |||
449 | sqlite3_finalize(ppstmt); | ||
450 | } | ||
451 | } | ||
452 | |||
453 | ppgs.update(); | ||
454 | } | ||
455 | } | ||
456 | |||
457 | // Get nouns/adjectives/adverbs from WordNet | ||
458 | // Useful relations: | ||
459 | // - s: master list | ||
460 | // - ant: antonymy (e.g. happy/sad, sad/happy, happiness/sadness) | ||
461 | // - at: variation (e.g. a measurement can be standard or nonstandard) | ||
462 | // - hyp: hypernymy/hyponymy (e.g. color/red, color/blue) | ||
463 | // - ins: instantiation (do we need this? let's see) | ||
464 | // - mm: member meronymy/holonymy (e.g. family/mother, family/child) | ||
465 | // - mp: part meronymy/holonymy (e.g. wheel/spoke, wheel/tire) | ||
466 | // - ms: substance meronymy/holonymy (e.g. tire/rubber, doorstop/rubber) | ||
467 | // - per: pertainymy (e.g. something that is Alaskan pertains to Alaska) | ||
468 | // mannernymy (e.g. something done quickly is done in a manner that is quick) | ||
469 | // - sa: specification (e.g. inaccurate (general) can mean imprecise or incorrect (specific)) | ||
470 | // - sim: synonymy (e.g. cheerful/happy, happy/cheerful) | ||
471 | // - syntax: positioning flags for some adjectives | ||
472 | std::string wnpref {argv[4]}; | ||
473 | if (wnpref.back() != '/') | ||
474 | { | ||
475 | wnpref += '/'; | ||
476 | } | ||
477 | |||
478 | // s table | ||
479 | { | ||
480 | std::ifstream wnsfile(wnpref + "wn_s.pl"); | ||
481 | if (!wnsfile.is_open()) | ||
482 | { | ||
483 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
484 | print_usage(); | ||
485 | } | ||
486 | |||
487 | std::list<std::string> lines; | ||
488 | for (;;) | ||
489 | { | ||
490 | std::string line; | ||
491 | if (!getline(wnsfile, line)) | ||
492 | { | ||
493 | break; | ||
494 | } | ||
495 | |||
496 | if (line.back() == '\r') | ||
497 | { | ||
498 | line.pop_back(); | ||
499 | } | ||
500 | |||
501 | lines.push_back(line); | ||
502 | } | ||
503 | |||
504 | progress ppgs("Writing nouns, adjectives, and adverbs...", lines.size()); | ||
505 | for (auto line : lines) | ||
506 | { | ||
507 | ppgs.update(); | ||
508 | |||
509 | std::regex relation("^s\\(([134]\\d{8}),(\\d+),'([\\w ]+)',"); | ||
510 | std::smatch relation_data; | ||
511 | if (!std::regex_search(line, relation_data, relation)) | ||
512 | { | ||
513 | continue; | ||
514 | } | ||
515 | |||
516 | int synset_id = stoi(relation_data[1]); | ||
517 | int wnum = stoi(relation_data[2]); | ||
518 | std::string word = relation_data[3]; | ||
519 | |||
520 | std::string query; | ||
521 | switch (synset_id / 100000000) | ||
522 | { | ||
523 | case 1: // Noun | ||
524 | { | ||
525 | if (nouns.count(word) == 1) | ||
526 | { | ||
527 | query = "INSERT INTO nouns (singular, plural) VALUES (?, ?)"; | ||
528 | } else { | ||
529 | query = "INSERT INTO nouns (singular) VALUES (?)"; | ||
530 | } | ||
531 | |||
532 | break; | ||
533 | } | ||
534 | |||
535 | case 2: // Verb | ||
536 | { | ||
537 | // Ignore | ||
538 | |||
539 | break; | ||
540 | } | ||
541 | |||
542 | case 3: // Adjective | ||
543 | { | ||
544 | if (adjectives.count(word) == 1) | ||
545 | { | ||
546 | query = "INSERT INTO adjectives (base_form, comparative, superlative) VALUES (?, ?, ?)"; | ||
547 | } else { | ||
548 | query = "INSERT INTO adjectives (base_form) VALUES (?)"; | ||
549 | } | ||
550 | |||
551 | break; | ||
552 | } | ||
553 | |||
554 | case 4: // Adverb | ||
555 | { | ||
556 | if (adjectives.count(word) == 1) | ||
557 | { | ||
558 | query = "INSERT INTO adverbs (base_form, comparative, superlative) VALUES (?, ?, ?)"; | ||
559 | } else { | ||
560 | query = "INSERT INTO adverbs (base_form) VALUES (?)"; | ||
561 | } | ||
562 | |||
563 | break; | ||
564 | } | ||
565 | } | ||
566 | |||
567 | sqlite3_stmt* ppstmt; | ||
568 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
569 | { | ||
570 | db_error(ppdb, query); | ||
571 | } | ||
572 | |||
573 | sqlite3_bind_text(ppstmt, 1, word.c_str(), word.length(), SQLITE_STATIC); | ||
574 | switch (synset_id / 100000000) | ||
575 | { | ||
576 | case 1: // Noun | ||
577 | { | ||
578 | if (nouns.count(word) == 1) | ||
579 | { | ||
580 | sqlite3_bind_text(ppstmt, 2, nouns[word].plural.c_str(), nouns[word].plural.length(), SQLITE_STATIC); | ||
581 | } | ||
582 | |||
583 | break; | ||
584 | } | ||
585 | |||
586 | case 3: // Adjective | ||
587 | case 4: // Adverb | ||
588 | { | ||
589 | if (adjectives.count(word) == 1) | ||
590 | { | ||
591 | sqlite3_bind_text(ppstmt, 2, adjectives[word].comparative.c_str(), adjectives[word].comparative.length(), SQLITE_STATIC); | ||
592 | sqlite3_bind_text(ppstmt, 3, adjectives[word].superlative.c_str(), adjectives[word].superlative.length(), SQLITE_STATIC); | ||
593 | } | ||
594 | |||
595 | break; | ||
596 | } | ||
597 | } | ||
598 | |||
599 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
600 | { | ||
601 | db_error(ppdb, query); | ||
602 | } | ||
603 | |||
604 | sqlite3_finalize(ppstmt); | ||
605 | |||
606 | query = "SELECT last_insert_rowid()"; | ||
607 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
608 | { | ||
609 | db_error(ppdb, query); | ||
610 | } | ||
611 | |||
612 | if (sqlite3_step(ppstmt) != SQLITE_ROW) | ||
613 | { | ||
614 | db_error(ppdb, query); | ||
615 | } | ||
616 | |||
617 | int rowid = sqlite3_column_int(ppstmt, 0); | ||
618 | wn[synset_id][wnum] = rowid; | ||
619 | |||
620 | sqlite3_finalize(ppstmt); | ||
621 | |||
622 | std::string canonical(word); | ||
623 | std::transform(std::begin(canonical), std::end(canonical), std::begin(canonical), ::tolower); | ||
624 | if (pronunciations.count(canonical) == 1) | ||
625 | { | ||
626 | for (auto pronunciation : pronunciations[canonical]) | ||
627 | { | ||
628 | switch (synset_id / 100000000) | ||
629 | { | ||
630 | case 1: // Noun | ||
631 | { | ||
632 | query = "INSERT INTO noun_pronunciations (noun_id, pronunciation) VALUES (?, ?)"; | ||
633 | |||
634 | break; | ||
635 | } | ||
636 | |||
637 | case 3: // Adjective | ||
638 | { | ||
639 | query = "INSERT INTO adjective_pronunciations (adjective_id, pronunciation) VALUES (?, ?)"; | ||
640 | |||
641 | break; | ||
642 | } | ||
643 | |||
644 | case 4: // Adverb | ||
645 | { | ||
646 | query = "INSERT INTO adverb_pronunciations (adverb_id, pronunciation) VALUES (?, ?)"; | ||
647 | |||
648 | break; | ||
649 | } | ||
650 | } | ||
651 | |||
652 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
653 | { | ||
654 | db_error(ppdb, query); | ||
655 | } | ||
656 | |||
657 | sqlite3_bind_int(ppstmt, 1, rowid); | ||
658 | sqlite3_bind_text(ppstmt, 2, pronunciation.c_str(), pronunciation.length(), SQLITE_STATIC); | ||
659 | |||
660 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
661 | { | ||
662 | db_error(ppdb, query); | ||
663 | } | ||
664 | |||
665 | sqlite3_finalize(ppstmt); | ||
666 | } | ||
667 | } | ||
668 | } | ||
669 | } | ||
670 | |||
671 | // While we're working on s | ||
672 | { | ||
673 | progress ppgs("Writing word synonyms...", wn.size()); | ||
674 | for (auto sense : wn) | ||
675 | { | ||
676 | ppgs.update(); | ||
677 | |||
678 | for (auto word1 : sense.second) | ||
679 | { | ||
680 | for (auto word2 : sense.second) | ||
681 | { | ||
682 | if (word1 != word2) | ||
683 | { | ||
684 | std::string query; | ||
685 | switch (sense.first / 100000000) | ||
686 | { | ||
687 | case 1: // Noun | ||
688 | { | ||
689 | query = "INSERT INTO noun_synonymy (noun_1_id, noun_2_id) VALUES (?, ?)"; | ||
690 | |||
691 | break; | ||
692 | } | ||
693 | |||
694 | case 2: // Verb | ||
695 | { | ||
696 | // Ignore | ||
697 | |||
698 | break; | ||
699 | } | ||
700 | |||
701 | case 3: // Adjective | ||
702 | { | ||
703 | query = "INSERT INTO adjective_synonymy (adjective_1_id, adjective_2_id) VALUES (?, ?)"; | ||
704 | |||
705 | break; | ||
706 | } | ||
707 | |||
708 | case 4: // Adverb | ||
709 | { | ||
710 | query = "INSERT INTO adverb_synonymy (adverb_1_id, adverb_2_id) VALUES (?, ?)"; | ||
711 | |||
712 | break; | ||
713 | } | ||
714 | } | ||
715 | |||
716 | sqlite3_stmt* ppstmt; | ||
717 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
718 | { | ||
719 | db_error(ppdb, query); | ||
720 | } | ||
721 | |||
722 | sqlite3_bind_int(ppstmt, 1, word1.second); | ||
723 | sqlite3_bind_int(ppstmt, 2, word2.second); | ||
724 | |||
725 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
726 | { | ||
727 | db_error(ppdb, query); | ||
728 | } | ||
729 | |||
730 | sqlite3_finalize(ppstmt); | ||
731 | } | ||
732 | } | ||
733 | } | ||
734 | } | ||
735 | } | ||
736 | |||
737 | // ant table | ||
738 | { | ||
739 | std::ifstream wnantfile(wnpref + "wn_ant.pl"); | ||
740 | if (!wnantfile.is_open()) | ||
741 | { | ||
742 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
743 | print_usage(); | ||
744 | } | ||
745 | |||
746 | std::list<std::string> lines; | ||
747 | for (;;) | ||
748 | { | ||
749 | std::string line; | ||
750 | if (!getline(wnantfile, line)) | ||
751 | { | ||
752 | break; | ||
753 | } | ||
754 | |||
755 | if (line.back() == '\r') | ||
756 | { | ||
757 | line.pop_back(); | ||
758 | } | ||
759 | |||
760 | lines.push_back(line); | ||
761 | } | ||
762 | |||
763 | progress ppgs("Writing antonyms...", lines.size()); | ||
764 | for (auto line : lines) | ||
765 | { | ||
766 | ppgs.update(); | ||
767 | |||
768 | std::regex relation("^ant\\(([134]\\d{8}),(\\d+),([134]\\d{8}),(\\d+)\\)\\."); | ||
769 | std::smatch relation_data; | ||
770 | if (!std::regex_search(line, relation_data, relation)) | ||
771 | { | ||
772 | continue; | ||
773 | } | ||
774 | |||
775 | int synset_id_1 = stoi(relation_data[1]); | ||
776 | int wnum_1 = stoi(relation_data[2]); | ||
777 | int synset_id_2 = stoi(relation_data[3]); | ||
778 | int wnum_2 = stoi(relation_data[4]); | ||
779 | |||
780 | std::string query; | ||
781 | switch (synset_id_1 / 100000000) | ||
782 | { | ||
783 | case 1: // Noun | ||
784 | { | ||
785 | query = "INSERT INTO noun_antonymy (noun_1_id, noun_2_id) VALUES (?, ?)"; | ||
786 | |||
787 | break; | ||
788 | } | ||
789 | |||
790 | case 2: // Verb | ||
791 | { | ||
792 | // Ignore | ||
793 | |||
794 | break; | ||
795 | } | ||
796 | |||
797 | case 3: // Adjective | ||
798 | { | ||
799 | query = "INSERT INTO adjective_antonymy (adjective_1_id, adjective_2_id) VALUES (?, ?)"; | ||
800 | |||
801 | break; | ||
802 | } | ||
803 | |||
804 | case 4: // Adverb | ||
805 | { | ||
806 | query = "INSERT INTO adverb_antonymy (adverb_1_id, adverb_2_id) VALUES (?, ?)"; | ||
807 | |||
808 | break; | ||
809 | } | ||
810 | } | ||
811 | |||
812 | sqlite3_stmt* ppstmt; | ||
813 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
814 | { | ||
815 | db_error(ppdb, query); | ||
816 | } | ||
817 | |||
818 | sqlite3_bind_int(ppstmt, 1, wn[synset_id_1][wnum_1]); | ||
819 | sqlite3_bind_int(ppstmt, 2, wn[synset_id_2][wnum_2]); | ||
820 | |||
821 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
822 | { | ||
823 | db_error(ppdb, query); | ||
824 | } | ||
825 | |||
826 | sqlite3_finalize(ppstmt); | ||
827 | } | ||
828 | } | ||
829 | |||
830 | // at table | ||
831 | { | ||
832 | std::ifstream wnatfile(wnpref + "wn_at.pl"); | ||
833 | if (!wnatfile.is_open()) | ||
834 | { | ||
835 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
836 | print_usage(); | ||
837 | } | ||
838 | |||
839 | std::list<std::string> lines; | ||
840 | for (;;) | ||
841 | { | ||
842 | std::string line; | ||
843 | if (!getline(wnatfile, line)) | ||
844 | { | ||
845 | break; | ||
846 | } | ||
847 | |||
848 | if (line.back() == '\r') | ||
849 | { | ||
850 | line.pop_back(); | ||
851 | } | ||
852 | |||
853 | lines.push_back(line); | ||
854 | } | ||
855 | |||
856 | progress ppgs("Writing variations...", lines.size()); | ||
857 | for (auto line : lines) | ||
858 | { | ||
859 | ppgs.update(); | ||
860 | |||
861 | std::regex relation("^at\\((1\\d{8}),(3\\d{8})\\)\\."); | ||
862 | std::smatch relation_data; | ||
863 | if (!std::regex_search(line, relation_data, relation)) | ||
864 | { | ||
865 | continue; | ||
866 | } | ||
867 | |||
868 | int synset_id_1 = stoi(relation_data[1]); | ||
869 | int synset_id_2 = stoi(relation_data[2]); | ||
870 | std::string query("INSERT INTO variation (noun_id, adjective_id) VALUES (?, ?)"); | ||
871 | |||
872 | for (auto mapping1 : wn[synset_id_1]) | ||
873 | { | ||
874 | for (auto mapping2 : wn[synset_id_2]) | ||
875 | { | ||
876 | sqlite3_stmt* ppstmt; | ||
877 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
878 | { | ||
879 | db_error(ppdb, query); | ||
880 | } | ||
881 | |||
882 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
883 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
884 | |||
885 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
886 | { | ||
887 | db_error(ppdb, query); | ||
888 | } | ||
889 | |||
890 | sqlite3_finalize(ppstmt); | ||
891 | } | ||
892 | } | ||
893 | } | ||
894 | } | ||
895 | |||
896 | // hyp table | ||
897 | { | ||
898 | std::ifstream wnhypfile(wnpref + "wn_hyp.pl"); | ||
899 | if (!wnhypfile.is_open()) | ||
900 | { | ||
901 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
902 | print_usage(); | ||
903 | } | ||
904 | |||
905 | std::list<std::string> lines; | ||
906 | for (;;) | ||
907 | { | ||
908 | std::string line; | ||
909 | if (!getline(wnhypfile, line)) | ||
910 | { | ||
911 | break; | ||
912 | } | ||
913 | |||
914 | if (line.back() == '\r') | ||
915 | { | ||
916 | line.pop_back(); | ||
917 | } | ||
918 | |||
919 | lines.push_back(line); | ||
920 | } | ||
921 | |||
922 | progress ppgs("Writing hypernyms...", lines.size()); | ||
923 | for (auto line : lines) | ||
924 | { | ||
925 | ppgs.update(); | ||
926 | |||
927 | std::regex relation("^hyp\\((1\\d{8}),(1\\d{8})\\)\\."); | ||
928 | std::smatch relation_data; | ||
929 | if (!std::regex_search(line, relation_data, relation)) | ||
930 | { | ||
931 | continue; | ||
932 | } | ||
933 | |||
934 | int synset_id_1 = stoi(relation_data[1]); | ||
935 | int synset_id_2 = stoi(relation_data[2]); | ||
936 | std::string query("INSERT INTO hypernymy (hyponym_id, hypernym_id) VALUES (?, ?)"); | ||
937 | |||
938 | for (auto mapping1 : wn[synset_id_1]) | ||
939 | { | ||
940 | for (auto mapping2 : wn[synset_id_2]) | ||
941 | { | ||
942 | sqlite3_stmt* ppstmt; | ||
943 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
944 | { | ||
945 | db_error(ppdb, query); | ||
946 | } | ||
947 | |||
948 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
949 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
950 | |||
951 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
952 | { | ||
953 | db_error(ppdb, query); | ||
954 | } | ||
955 | |||
956 | sqlite3_finalize(ppstmt); | ||
957 | } | ||
958 | } | ||
959 | } | ||
960 | } | ||
961 | |||
962 | // ins table | ||
963 | { | ||
964 | std::ifstream wninsfile(wnpref + "wn_ins.pl"); | ||
965 | if (!wninsfile.is_open()) | ||
966 | { | ||
967 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
968 | print_usage(); | ||
969 | } | ||
970 | |||
971 | std::list<std::string> lines; | ||
972 | for (;;) | ||
973 | { | ||
974 | std::string line; | ||
975 | if (!getline(wninsfile, line)) | ||
976 | { | ||
977 | break; | ||
978 | } | ||
979 | |||
980 | if (line.back() == '\r') | ||
981 | { | ||
982 | line.pop_back(); | ||
983 | } | ||
984 | |||
985 | lines.push_back(line); | ||
986 | } | ||
987 | |||
988 | progress ppgs("Writing instantiations...", lines.size()); | ||
989 | for (auto line : lines) | ||
990 | { | ||
991 | ppgs.update(); | ||
992 | |||
993 | std::regex relation("^ins\\((1\\d{8}),(1\\d{8})\\)\\."); | ||
994 | std::smatch relation_data; | ||
995 | if (!std::regex_search(line, relation_data, relation)) | ||
996 | { | ||
997 | continue; | ||
998 | } | ||
999 | |||
1000 | int synset_id_1 = stoi(relation_data[1]); | ||
1001 | int synset_id_2 = stoi(relation_data[2]); | ||
1002 | std::string query("INSERT INTO instantiation (instance_id, class_id) VALUES (?, ?)"); | ||
1003 | |||
1004 | for (auto mapping1 : wn[synset_id_1]) | ||
1005 | { | ||
1006 | for (auto mapping2 : wn[synset_id_2]) | ||
1007 | { | ||
1008 | sqlite3_stmt* ppstmt; | ||
1009 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1010 | { | ||
1011 | db_error(ppdb, query); | ||
1012 | } | ||
1013 | |||
1014 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
1015 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
1016 | |||
1017 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1018 | { | ||
1019 | db_error(ppdb, query); | ||
1020 | } | ||
1021 | |||
1022 | sqlite3_finalize(ppstmt); | ||
1023 | } | ||
1024 | } | ||
1025 | } | ||
1026 | } | ||
1027 | |||
1028 | // mm table | ||
1029 | { | ||
1030 | std::ifstream wnmmfile(wnpref + "wn_mm.pl"); | ||
1031 | if (!wnmmfile.is_open()) | ||
1032 | { | ||
1033 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1034 | print_usage(); | ||
1035 | } | ||
1036 | |||
1037 | std::list<std::string> lines; | ||
1038 | for (;;) | ||
1039 | { | ||
1040 | std::string line; | ||
1041 | if (!getline(wnmmfile, line)) | ||
1042 | { | ||
1043 | break; | ||
1044 | } | ||
1045 | |||
1046 | if (line.back() == '\r') | ||
1047 | { | ||
1048 | line.pop_back(); | ||
1049 | } | ||
1050 | |||
1051 | lines.push_back(line); | ||
1052 | } | ||
1053 | |||
1054 | progress ppgs("Writing member meronyms...", lines.size()); | ||
1055 | for (auto line : lines) | ||
1056 | { | ||
1057 | ppgs.update(); | ||
1058 | |||
1059 | std::regex relation("^mm\\((1\\d{8}),(1\\d{8})\\)\\."); | ||
1060 | std::smatch relation_data; | ||
1061 | if (!std::regex_search(line, relation_data, relation)) | ||
1062 | { | ||
1063 | continue; | ||
1064 | } | ||
1065 | |||
1066 | int synset_id_1 = stoi(relation_data[1]); | ||
1067 | int synset_id_2 = stoi(relation_data[2]); | ||
1068 | std::string query("INSERT INTO member_meronymy (holonym_id, meronym_id) VALUES (?, ?)"); | ||
1069 | |||
1070 | for (auto mapping1 : wn[synset_id_1]) | ||
1071 | { | ||
1072 | for (auto mapping2 : wn[synset_id_2]) | ||
1073 | { | ||
1074 | sqlite3_stmt* ppstmt; | ||
1075 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1076 | { | ||
1077 | db_error(ppdb, query); | ||
1078 | } | ||
1079 | |||
1080 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
1081 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
1082 | |||
1083 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1084 | { | ||
1085 | db_error(ppdb, query); | ||
1086 | } | ||
1087 | |||
1088 | sqlite3_finalize(ppstmt); | ||
1089 | } | ||
1090 | } | ||
1091 | } | ||
1092 | } | ||
1093 | |||
1094 | // ms table | ||
1095 | { | ||
1096 | std::ifstream wnmsfile(wnpref + "wn_ms.pl"); | ||
1097 | if (!wnmsfile.is_open()) | ||
1098 | { | ||
1099 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1100 | print_usage(); | ||
1101 | } | ||
1102 | |||
1103 | std::list<std::string> lines; | ||
1104 | for (;;) | ||
1105 | { | ||
1106 | std::string line; | ||
1107 | if (!getline(wnmsfile, line)) | ||
1108 | { | ||
1109 | break; | ||
1110 | } | ||
1111 | |||
1112 | if (line.back() == '\r') | ||
1113 | { | ||
1114 | line.pop_back(); | ||
1115 | } | ||
1116 | |||
1117 | lines.push_back(line); | ||
1118 | } | ||
1119 | |||
1120 | progress ppgs("Writing substance meronyms...", lines.size()); | ||
1121 | for (auto line : lines) | ||
1122 | { | ||
1123 | ppgs.update(); | ||
1124 | |||
1125 | std::regex relation("^ms\\((1\\d{8}),(1\\d{8})\\)\\."); | ||
1126 | std::smatch relation_data; | ||
1127 | if (!std::regex_search(line, relation_data, relation)) | ||
1128 | { | ||
1129 | continue; | ||
1130 | } | ||
1131 | |||
1132 | int synset_id_1 = stoi(relation_data[1]); | ||
1133 | int synset_id_2 = stoi(relation_data[2]); | ||
1134 | std::string query("INSERT INTO substance_meronymy (holonym_id, meronym_id) VALUES (?, ?)"); | ||
1135 | |||
1136 | for (auto mapping1 : wn[synset_id_1]) | ||
1137 | { | ||
1138 | for (auto mapping2 : wn[synset_id_2]) | ||
1139 | { | ||
1140 | sqlite3_stmt* ppstmt; | ||
1141 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1142 | { | ||
1143 | db_error(ppdb, query); | ||
1144 | } | ||
1145 | |||
1146 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
1147 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
1148 | |||
1149 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1150 | { | ||
1151 | db_error(ppdb, query); | ||
1152 | } | ||
1153 | |||
1154 | sqlite3_finalize(ppstmt); | ||
1155 | } | ||
1156 | } | ||
1157 | } | ||
1158 | } | ||
1159 | |||
1160 | // mm table | ||
1161 | { | ||
1162 | std::ifstream wnmpfile(wnpref + "wn_mp.pl"); | ||
1163 | if (!wnmpfile.is_open()) | ||
1164 | { | ||
1165 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1166 | print_usage(); | ||
1167 | } | ||
1168 | |||
1169 | std::list<std::string> lines; | ||
1170 | for (;;) | ||
1171 | { | ||
1172 | std::string line; | ||
1173 | if (!getline(wnmpfile, line)) | ||
1174 | { | ||
1175 | break; | ||
1176 | } | ||
1177 | |||
1178 | if (line.back() == '\r') | ||
1179 | { | ||
1180 | line.pop_back(); | ||
1181 | } | ||
1182 | |||
1183 | lines.push_back(line); | ||
1184 | } | ||
1185 | |||
1186 | progress ppgs("Writing part meronyms...", lines.size()); | ||
1187 | for (auto line : lines) | ||
1188 | { | ||
1189 | ppgs.update(); | ||
1190 | |||
1191 | std::regex relation("^mp\\((1\\d{8}),(1\\d{8})\\)\\."); | ||
1192 | std::smatch relation_data; | ||
1193 | if (!std::regex_search(line, relation_data, relation)) | ||
1194 | { | ||
1195 | continue; | ||
1196 | } | ||
1197 | |||
1198 | int synset_id_1 = stoi(relation_data[1]); | ||
1199 | int synset_id_2 = stoi(relation_data[2]); | ||
1200 | std::string query("INSERT INTO part_meronymy (holonym_id, meronym_id) VALUES (?, ?)"); | ||
1201 | |||
1202 | for (auto mapping1 : wn[synset_id_1]) | ||
1203 | { | ||
1204 | for (auto mapping2 : wn[synset_id_2]) | ||
1205 | { | ||
1206 | sqlite3_stmt* ppstmt; | ||
1207 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1208 | { | ||
1209 | db_error(ppdb, query); | ||
1210 | } | ||
1211 | |||
1212 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
1213 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
1214 | |||
1215 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1216 | { | ||
1217 | db_error(ppdb, query); | ||
1218 | } | ||
1219 | |||
1220 | sqlite3_finalize(ppstmt); | ||
1221 | } | ||
1222 | } | ||
1223 | } | ||
1224 | } | ||
1225 | |||
1226 | // per table | ||
1227 | { | ||
1228 | std::ifstream wnperfile(wnpref + "wn_per.pl"); | ||
1229 | if (!wnperfile.is_open()) | ||
1230 | { | ||
1231 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1232 | print_usage(); | ||
1233 | } | ||
1234 | |||
1235 | std::list<std::string> lines; | ||
1236 | for (;;) | ||
1237 | { | ||
1238 | std::string line; | ||
1239 | if (!getline(wnperfile, line)) | ||
1240 | { | ||
1241 | break; | ||
1242 | } | ||
1243 | |||
1244 | if (line.back() == '\r') | ||
1245 | { | ||
1246 | line.pop_back(); | ||
1247 | } | ||
1248 | |||
1249 | lines.push_back(line); | ||
1250 | } | ||
1251 | |||
1252 | progress ppgs("Writing pertainyms and mannernyms...", lines.size()); | ||
1253 | for (auto line : lines) | ||
1254 | { | ||
1255 | ppgs.update(); | ||
1256 | |||
1257 | std::regex relation("^per\\(([34]\\d{8}),(\\d+),([13]\\d{8}),(\\d+)\\)\\."); | ||
1258 | std::smatch relation_data; | ||
1259 | if (!std::regex_search(line, relation_data, relation)) | ||
1260 | { | ||
1261 | continue; | ||
1262 | } | ||
1263 | |||
1264 | int synset_id_1 = stoi(relation_data[1]); | ||
1265 | int wnum_1 = stoi(relation_data[2]); | ||
1266 | int synset_id_2 = stoi(relation_data[3]); | ||
1267 | int wnum_2 = stoi(relation_data[4]); | ||
1268 | std::string query; | ||
1269 | switch (synset_id_1 / 100000000) | ||
1270 | { | ||
1271 | case 3: // Adjective | ||
1272 | { | ||
1273 | // This is a pertainym, the second word should be a noun | ||
1274 | // Technically it can be an adjective but we're ignoring that | ||
1275 | if (synset_id_2 / 100000000 != 1) | ||
1276 | { | ||
1277 | continue; | ||
1278 | } | ||
1279 | |||
1280 | query = "INSERT INTO pertainymy (pertainym_id, noun_id) VALUES (?, ?)"; | ||
1281 | |||
1282 | break; | ||
1283 | } | ||
1284 | |||
1285 | case 4: // Adverb | ||
1286 | { | ||
1287 | // This is a mannernym, the second word should be an adjective | ||
1288 | if (synset_id_2 / 100000000 != 3) | ||
1289 | { | ||
1290 | continue; | ||
1291 | } | ||
1292 | |||
1293 | query = "INSERT INTO mannernymy (mannernym_id, adjective_id) VALUES (?, ?)"; | ||
1294 | |||
1295 | break; | ||
1296 | } | ||
1297 | } | ||
1298 | |||
1299 | sqlite3_stmt* ppstmt; | ||
1300 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
1301 | { | ||
1302 | db_error(ppdb, query); | ||
1303 | } | ||
1304 | |||
1305 | sqlite3_bind_int(ppstmt, 1, wn[synset_id_1][wnum_1]); | ||
1306 | sqlite3_bind_int(ppstmt, 2, wn[synset_id_2][wnum_2]); | ||
1307 | |||
1308 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1309 | { | ||
1310 | db_error(ppdb, query); | ||
1311 | } | ||
1312 | |||
1313 | sqlite3_finalize(ppstmt); | ||
1314 | } | ||
1315 | } | ||
1316 | |||
1317 | // sa table | ||
1318 | { | ||
1319 | std::ifstream wnsafile(wnpref + "wn_sa.pl"); | ||
1320 | if (!wnsafile.is_open()) | ||
1321 | { | ||
1322 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1323 | print_usage(); | ||
1324 | } | ||
1325 | |||
1326 | std::list<std::string> lines; | ||
1327 | for (;;) | ||
1328 | { | ||
1329 | std::string line; | ||
1330 | if (!getline(wnsafile, line)) | ||
1331 | { | ||
1332 | break; | ||
1333 | } | ||
1334 | |||
1335 | if (line.back() == '\r') | ||
1336 | { | ||
1337 | line.pop_back(); | ||
1338 | } | ||
1339 | |||
1340 | lines.push_back(line); | ||
1341 | } | ||
1342 | |||
1343 | progress ppgs("Writing specifications...", lines.size()); | ||
1344 | for (auto line : lines) | ||
1345 | { | ||
1346 | ppgs.update(); | ||
1347 | |||
1348 | std::regex relation("^per\\((3\\d{8}),(\\d+),(3\\d{8}),(\\d+)\\)\\."); | ||
1349 | std::smatch relation_data; | ||
1350 | if (!std::regex_search(line, relation_data, relation)) | ||
1351 | { | ||
1352 | continue; | ||
1353 | } | ||
1354 | |||
1355 | int synset_id_1 = stoi(relation_data[1]); | ||
1356 | int wnum_1 = stoi(relation_data[2]); | ||
1357 | int synset_id_2 = stoi(relation_data[3]); | ||
1358 | int wnum_2 = stoi(relation_data[4]); | ||
1359 | std::string query("INSERT INTO specification (general_id, specific_id) VALUES (?, ?)"); | ||
1360 | |||
1361 | sqlite3_stmt* ppstmt; | ||
1362 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
1363 | { | ||
1364 | db_error(ppdb, query); | ||
1365 | } | ||
1366 | |||
1367 | sqlite3_bind_int(ppstmt, 1, wn[synset_id_1][wnum_1]); | ||
1368 | sqlite3_bind_int(ppstmt, 2, wn[synset_id_2][wnum_2]); | ||
1369 | |||
1370 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1371 | { | ||
1372 | db_error(ppdb, query); | ||
1373 | } | ||
1374 | |||
1375 | sqlite3_finalize(ppstmt); | ||
1376 | } | ||
1377 | } | ||
1378 | /* | ||
1379 | // sim table | ||
1380 | { | ||
1381 | std::ifstream wnsimfile(wnpref + "wn_sim.pl"); | ||
1382 | if (!wnsimfile.is_open()) | ||
1383 | { | ||
1384 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1385 | print_usage(); | ||
1386 | } | ||
1387 | |||
1388 | std::list<std::string> lines; | ||
1389 | for (;;) | ||
1390 | { | ||
1391 | std::string line; | ||
1392 | if (!getline(wnsimfile, line)) | ||
1393 | { | ||
1394 | break; | ||
1395 | } | ||
1396 | |||
1397 | if (line.back() == '\r') | ||
1398 | { | ||
1399 | line.pop_back(); | ||
1400 | } | ||
1401 | |||
1402 | lines.push_back(line); | ||
1403 | } | ||
1404 | |||
1405 | progress ppgs("Writing sense synonyms...", lines.size()); | ||
1406 | for (auto line : lines) | ||
1407 | { | ||
1408 | ppgs.update(); | ||
1409 | |||
1410 | std::regex relation("^sim\\((3\\d{8}),(3\\d{8})\\)\\."); | ||
1411 | std::smatch relation_data; | ||
1412 | if (!std::regex_search(line, relation_data, relation)) | ||
1413 | { | ||
1414 | continue; | ||
1415 | } | ||
1416 | |||
1417 | int synset_id_1 = stoi(relation_data[1]); | ||
1418 | int synset_id_2 = stoi(relation_data[2]); | ||
1419 | std::string query("INSERT INTO adjective_synonymy (adjective_1_id, adjective_2_id) VALUES (?, ?)"); | ||
1420 | |||
1421 | for (auto mapping1 : wn[synset_id_1]) | ||
1422 | { | ||
1423 | for (auto mapping2 : wn[synset_id_2]) | ||
1424 | { | ||
1425 | sqlite3_stmt* ppstmt; | ||
1426 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1427 | { | ||
1428 | db_error(ppdb, query); | ||
1429 | } | ||
1430 | |||
1431 | sqlite3_bind_int(ppstmt, 1, mapping1.second); | ||
1432 | sqlite3_bind_int(ppstmt, 2, mapping2.second); | ||
1433 | |||
1434 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1435 | { | ||
1436 | db_error(ppdb, query); | ||
1437 | } | ||
1438 | |||
1439 | sqlite3_reset(ppstmt); | ||
1440 | sqlite3_clear_bindings(ppstmt); | ||
1441 | |||
1442 | sqlite3_bind_int(ppstmt, 1, mapping2.second); | ||
1443 | sqlite3_bind_int(ppstmt, 2, mapping1.second); | ||
1444 | |||
1445 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1446 | { | ||
1447 | db_error(ppdb, query); | ||
1448 | } | ||
1449 | |||
1450 | sqlite3_finalize(ppstmt); | ||
1451 | } | ||
1452 | } | ||
1453 | } | ||
1454 | } | ||
1455 | */ | ||
1456 | // syntax table | ||
1457 | { | ||
1458 | std::ifstream wnsyntaxfile(wnpref + "wn_syntax.pl"); | ||
1459 | if (!wnsyntaxfile.is_open()) | ||
1460 | { | ||
1461 | std::cout << "Invalid WordNet data directory." << std::endl; | ||
1462 | print_usage(); | ||
1463 | } | ||
1464 | |||
1465 | std::list<std::string> lines; | ||
1466 | for (;;) | ||
1467 | { | ||
1468 | std::string line; | ||
1469 | if (!getline(wnsyntaxfile, line)) | ||
1470 | { | ||
1471 | break; | ||
1472 | } | ||
1473 | |||
1474 | if (line.back() == '\r') | ||
1475 | { | ||
1476 | line.pop_back(); | ||
1477 | } | ||
1478 | |||
1479 | lines.push_back(line); | ||
1480 | } | ||
1481 | |||
1482 | progress ppgs("Writing adjective syntax markers...", lines.size()); | ||
1483 | for (auto line : lines) | ||
1484 | { | ||
1485 | ppgs.update(); | ||
1486 | |||
1487 | std::regex relation("^syntax\\((3\\d{8}),(\\d+),([ipa])p?\\)\\."); | ||
1488 | std::smatch relation_data; | ||
1489 | if (!std::regex_search(line, relation_data, relation)) | ||
1490 | { | ||
1491 | continue; | ||
1492 | } | ||
1493 | |||
1494 | int synset_id = stoi(relation_data[1]); | ||
1495 | int wnum = stoi(relation_data[2]); | ||
1496 | std::string syn = relation_data[3]; | ||
1497 | std::string query("UPDATE adjectives SET position = ? WHERE adjective_id = ?"); | ||
1498 | |||
1499 | sqlite3_stmt* ppstmt; | ||
1500 | if (sqlite3_prepare_v2(ppdb, query.c_str(), query.size(), &ppstmt, NULL) != SQLITE_OK) | ||
1501 | { | ||
1502 | db_error(ppdb, query); | ||
1503 | } | ||
1504 | |||
1505 | sqlite3_bind_text(ppstmt, 1, syn.c_str(), 1, SQLITE_STATIC); | ||
1506 | sqlite3_bind_int(ppstmt, 2, wn[synset_id][wnum]); | ||
1507 | |||
1508 | if (sqlite3_step(ppstmt) != SQLITE_DONE) | ||
1509 | { | ||
1510 | db_error(ppdb, query); | ||
1511 | } | ||
1512 | |||
1513 | sqlite3_finalize(ppstmt); | ||
1514 | } | ||
1515 | } | ||
1516 | |||
1517 | sqlite3_close_v2(ppdb); | ||
1518 | |||
1519 | std::cout << "Done." << std::endl; | ||
1520 | } | ||