diff options
author | Kelly Rauchenberger <fefferburbia@gmail.com> | 2016-03-24 23:16:07 -0400 |
---|---|---|
committer | Kelly Rauchenberger <fefferburbia@gmail.com> | 2016-03-24 23:16:07 -0400 |
commit | eef5de613c75661e5d94baa086f6f2ddc26c7ed0 (patch) | |
tree | 180230f6a245c5bca94d894273f5d2b93ded3f04 /lib/adverb.cpp | |
parent | d5ee4e39e5b5b3b8daa85cd972802195ad35e965 (diff) | |
download | verbly-eef5de613c75661e5d94baa086f6f2ddc26c7ed0.tar.gz verbly-eef5de613c75661e5d94baa086f6f2ddc26c7ed0.tar.bz2 verbly-eef5de613c75661e5d94baa086f6f2ddc26c7ed0.zip |
Added verb frames
In addition: - Added prepositions. - Rewrote a lot of the query interface. It now, for a lot of relationships, supports nested AND, OR, and NOT logic. - Rewrote the token class. It is now a union-like class instead of being polymorphic, which means smart pointers are no longer necessary. - Querying with regards to word derivation has been temporarily removed. - Sentinel values are now supported for all word types. - The VerbNet data retrieved from http://verbs.colorado.edu/~mpalmer/projects/verbnet/downloads.html was found to not be perfectly satisfactory in some regards, especially regarding adjective phrases. A patch file is now included in the repository describing the changes made to the VerbNet v3.2 download for the canonical verbly datafile.
Diffstat (limited to 'lib/adverb.cpp')
-rw-r--r-- | lib/adverb.cpp | 443 |
1 files changed, 23 insertions, 420 deletions
diff --git a/lib/adverb.cpp b/lib/adverb.cpp index 8fcddad..442574e 100644 --- a/lib/adverb.cpp +++ b/lib/adverb.cpp | |||
@@ -2,6 +2,11 @@ | |||
2 | 2 | ||
3 | namespace verbly { | 3 | namespace verbly { |
4 | 4 | ||
5 | adverb::adverb() | ||
6 | { | ||
7 | |||
8 | } | ||
9 | |||
5 | adverb::adverb(const data& _data, int _id) : word(_data, _id) | 10 | adverb::adverb(const data& _data, int _id) : word(_data, _id) |
6 | { | 11 | { |
7 | 12 | ||
@@ -9,460 +14,58 @@ namespace verbly { | |||
9 | 14 | ||
10 | std::string adverb::base_form() const | 15 | std::string adverb::base_form() const |
11 | { | 16 | { |
17 | assert(_valid == true); | ||
18 | |||
12 | return _base_form; | 19 | return _base_form; |
13 | } | 20 | } |
14 | 21 | ||
15 | std::string adverb::comparative_form() const | 22 | std::string adverb::comparative_form() const |
16 | { | 23 | { |
24 | assert(_valid == true); | ||
25 | |||
17 | return _comparative_form; | 26 | return _comparative_form; |
18 | } | 27 | } |
19 | 28 | ||
20 | std::string adverb::superlative_form() const | 29 | std::string adverb::superlative_form() const |
21 | { | 30 | { |
31 | assert(_valid == true); | ||
32 | |||
22 | return _superlative_form; | 33 | return _superlative_form; |
23 | } | 34 | } |
24 | 35 | ||
25 | bool adverb::has_comparative_form() const | 36 | bool adverb::has_comparative_form() const |
26 | { | 37 | { |
38 | assert(_valid == true); | ||
39 | |||
27 | return !_comparative_form.empty(); | 40 | return !_comparative_form.empty(); |
28 | } | 41 | } |
29 | 42 | ||
30 | bool adverb::has_superlative_form() const | 43 | bool adverb::has_superlative_form() const |
31 | { | 44 | { |
45 | assert(_valid == true); | ||
46 | |||
32 | return !_superlative_form.empty(); | 47 | return !_superlative_form.empty(); |
33 | } | 48 | } |
34 | 49 | ||
35 | adverb_query adverb::antonyms() const | 50 | adverb_query adverb::antonyms() const |
36 | { | 51 | { |
37 | return _data.adverbs().antonym_of(*this); | 52 | assert(_valid == true); |
38 | } | ||
39 | |||
40 | adverb_query adverb::synonyms() const | ||
41 | { | ||
42 | return _data.adverbs().synonym_of(*this); | ||
43 | } | ||
44 | |||
45 | adjective_query adverb::anti_mannernyms() const | ||
46 | { | ||
47 | return _data.adjectives().anti_mannernym_of(*this); | ||
48 | } | ||
49 | |||
50 | adverb_query::adverb_query(const data& _data) : _data(_data) | ||
51 | { | ||
52 | |||
53 | } | ||
54 | |||
55 | adverb_query& adverb_query::limit(int _limit) | ||
56 | { | ||
57 | if ((_limit > 0) || (_limit == unlimited)) | ||
58 | { | ||
59 | this->_limit = _limit; | ||
60 | } | ||
61 | |||
62 | return *this; | ||
63 | } | ||
64 | |||
65 | adverb_query& adverb_query::random(bool _random) | ||
66 | { | ||
67 | this->_random = _random; | ||
68 | |||
69 | return *this; | ||
70 | } | ||
71 | |||
72 | adverb_query& adverb_query::except(const adverb& _word) | ||
73 | { | ||
74 | _except.push_back(_word); | ||
75 | 53 | ||
76 | return *this; | 54 | return _data->adverbs().antonym_of(*this); |
77 | } | 55 | } |
78 | 56 | ||
79 | adverb_query& adverb_query::rhymes_with(const word& _word) | 57 | adverb_query adverb::synonyms() const |
80 | { | ||
81 | for (auto rhyme : _word.rhyme_phonemes()) | ||
82 | { | ||
83 | _rhymes.push_back(rhyme); | ||
84 | } | ||
85 | |||
86 | if (dynamic_cast<const adverb*>(&_word) != nullptr) | ||
87 | { | ||
88 | _except.push_back(dynamic_cast<const adverb&>(_word)); | ||
89 | } | ||
90 | |||
91 | return *this; | ||
92 | } | ||
93 | |||
94 | adverb_query& adverb_query::has_pronunciation(bool _has_prn) | ||
95 | { | ||
96 | this->_has_prn = _has_prn; | ||
97 | |||
98 | return *this; | ||
99 | } | ||
100 | |||
101 | adverb_query& adverb_query::requires_comparative_form(bool _arg) | ||
102 | { | ||
103 | _requires_comparative_form = _arg; | ||
104 | |||
105 | return *this; | ||
106 | } | ||
107 | |||
108 | adverb_query& adverb_query::requires_superlative_form(bool _arg) | ||
109 | { | ||
110 | _requires_superlative_form = _arg; | ||
111 | |||
112 | return *this; | ||
113 | } | ||
114 | |||
115 | adverb_query& adverb_query::has_antonyms(bool _arg) | ||
116 | { | ||
117 | _has_antonyms = _arg; | ||
118 | |||
119 | return *this; | ||
120 | } | ||
121 | |||
122 | adverb_query& adverb_query::antonym_of(const adverb& _adv) | ||
123 | { | ||
124 | _antonym_of.push_back(_adv); | ||
125 | |||
126 | return *this; | ||
127 | } | ||
128 | |||
129 | adverb_query& adverb_query::not_antonym_of(const adverb& _adv) | ||
130 | { | ||
131 | _not_antonym_of.push_back(_adv); | ||
132 | |||
133 | return *this; | ||
134 | } | ||
135 | |||
136 | adverb_query& adverb_query::has_synonyms(bool _arg) | ||
137 | { | ||
138 | _has_synonyms = _arg; | ||
139 | |||
140 | return *this; | ||
141 | } | ||
142 | |||
143 | adverb_query& adverb_query::synonym_of(const adverb& _adv) | ||
144 | { | ||
145 | _synonym_of.push_back(_adv); | ||
146 | |||
147 | return *this; | ||
148 | } | ||
149 | |||
150 | adverb_query& adverb_query::not_synonym_of(const adverb& _adv) | ||
151 | { | ||
152 | _not_synonym_of.push_back(_adv); | ||
153 | |||
154 | return *this; | ||
155 | } | ||
156 | |||
157 | adverb_query& adverb_query::is_mannernymic(bool _arg) | ||
158 | { | ||
159 | _is_mannernymic = _arg; | ||
160 | |||
161 | return *this; | ||
162 | } | ||
163 | |||
164 | adverb_query& adverb_query::mannernym_of(const adjective& _adj) | ||
165 | { | ||
166 | _mannernym_of.push_back(_adj); | ||
167 | |||
168 | return *this; | ||
169 | } | ||
170 | |||
171 | adverb_query& adverb_query::derived_from(const word& _w) | ||
172 | { | ||
173 | if (dynamic_cast<const adjective*>(&_w) != nullptr) | ||
174 | { | ||
175 | _derived_from_adjective.push_back(dynamic_cast<const adjective&>(_w)); | ||
176 | } else if (dynamic_cast<const adverb*>(&_w) != nullptr) | ||
177 | { | ||
178 | _derived_from_adverb.push_back(dynamic_cast<const adverb&>(_w)); | ||
179 | } else if (dynamic_cast<const noun*>(&_w) != nullptr) | ||
180 | { | ||
181 | _derived_from_noun.push_back(dynamic_cast<const noun&>(_w)); | ||
182 | } | ||
183 | |||
184 | return *this; | ||
185 | } | ||
186 | |||
187 | adverb_query& adverb_query::not_derived_from(const word& _w) | ||
188 | { | 58 | { |
189 | if (dynamic_cast<const adjective*>(&_w) != nullptr) | 59 | assert(_valid == true); |
190 | { | ||
191 | _not_derived_from_adjective.push_back(dynamic_cast<const adjective&>(_w)); | ||
192 | } else if (dynamic_cast<const adverb*>(&_w) != nullptr) | ||
193 | { | ||
194 | _not_derived_from_adverb.push_back(dynamic_cast<const adverb&>(_w)); | ||
195 | } else if (dynamic_cast<const noun*>(&_w) != nullptr) | ||
196 | { | ||
197 | _not_derived_from_noun.push_back(dynamic_cast<const noun&>(_w)); | ||
198 | } | ||
199 | 60 | ||
200 | return *this; | 61 | return _data->adverbs().synonym_of(*this); |
201 | } | 62 | } |
202 | 63 | ||
203 | std::list<adverb> adverb_query::run() const | 64 | adjective_query adverb::anti_mannernyms() const |
204 | { | 65 | { |
205 | std::stringstream construct; | 66 | assert(_valid == true); |
206 | construct << "SELECT adverb_id, base_form, comparative, superlative FROM adverbs"; | ||
207 | std::list<std::string> conditions; | ||
208 | |||
209 | if (_has_prn) | ||
210 | { | ||
211 | conditions.push_back("adverb_id IN (SELECT adverb_id FROM adverb_pronunciations)"); | ||
212 | } | ||
213 | |||
214 | if (!_rhymes.empty()) | ||
215 | { | ||
216 | std::list<std::string> clauses(_rhymes.size(), "pronunciation LIKE @RHMPRN"); | ||
217 | std::string cond = "adverb_id IN (SELECT adverb_id FROM adverb_pronunciations WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
218 | conditions.push_back(cond); | ||
219 | } | ||
220 | |||
221 | for (auto except : _except) | ||
222 | { | ||
223 | conditions.push_back("adverb_id != @EXCID"); | ||
224 | } | ||
225 | |||
226 | if (_requires_comparative_form) | ||
227 | { | ||
228 | conditions.push_back("comparative IS NOT NULL"); | ||
229 | } | ||
230 | |||
231 | if (_requires_superlative_form) | ||
232 | { | ||
233 | conditions.push_back("superlative IS NOT NULL"); | ||
234 | } | ||
235 | |||
236 | if (_has_antonyms) | ||
237 | { | ||
238 | conditions.push_back("adverb_id IN (SELECT adverb_2_id FROM adverb_antonymy)"); | ||
239 | } | ||
240 | |||
241 | if (!_antonym_of.empty()) | ||
242 | { | ||
243 | std::list<std::string> clauses(_antonym_of.size(), "adverb_1_id = @ANTID"); | ||
244 | std::string cond = "adverb_id IN (SELECT adverb_2_id FROM adverb_antonymy WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
245 | conditions.push_back(cond); | ||
246 | } | ||
247 | |||
248 | if (!_not_antonym_of.empty()) | ||
249 | { | ||
250 | std::list<std::string> clauses(_not_antonym_of.size(), "adverb_1_id = @NANTID"); | ||
251 | std::string cond = "adverb_id NOT IN (SELECT adverb_2_id FROM adverb_antonymy WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
252 | conditions.push_back(cond); | ||
253 | } | ||
254 | |||
255 | if (_has_synonyms) | ||
256 | { | ||
257 | conditions.push_back("adverb_id IN (SELECT adverb_2_id FROM adverb_synonymy)"); | ||
258 | } | ||
259 | 67 | ||
260 | if (!_synonym_of.empty()) | 68 | return _data->adjectives().anti_mannernym_of(*this); |
261 | { | ||
262 | std::list<std::string> clauses(_synonym_of.size(), "adverb_1_id = @SYNID"); | ||
263 | std::string cond = "adverb_id IN (SELECT adverb_2_id FROM adverb_synonymy WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
264 | conditions.push_back(cond); | ||
265 | } | ||
266 | |||
267 | if (!_not_synonym_of.empty()) | ||
268 | { | ||
269 | std::list<std::string> clauses(_not_synonym_of.size(), "adverb_1_id = @NSYNID"); | ||
270 | std::string cond = "adverb_id NOT IN (SELECT adverb_2_id FROM adverb_synonymy WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
271 | conditions.push_back(cond); | ||
272 | } | ||
273 | |||
274 | if (_is_mannernymic) | ||
275 | { | ||
276 | conditions.push_back("adverb_id IN (SELECT mannernym_id FROM mannernymy)"); | ||
277 | } | ||
278 | |||
279 | if (!_mannernym_of.empty()) | ||
280 | { | ||
281 | std::list<std::string> clauses(_mannernym_of.size(), "adjective_id = @AMANID"); | ||
282 | std::string cond = "adverb_id IN (SELECT mannernym_id FROM mannernymy WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
283 | conditions.push_back(cond); | ||
284 | } | ||
285 | |||
286 | if (!_derived_from_adjective.empty()) | ||
287 | { | ||
288 | std::list<std::string> clauses(_derived_from_adjective.size(), "adjective_id = @DERADJ"); | ||
289 | std::string cond = "adverb_id IN (SELECT adverb_id FROM adjective_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
290 | conditions.push_back(cond); | ||
291 | } | ||
292 | |||
293 | if (!_not_derived_from_adjective.empty()) | ||
294 | { | ||
295 | std::list<std::string> clauses(_not_derived_from_adjective.size(), "adjective_id = @NDERADJ"); | ||
296 | std::string cond = "adverb_id NOT IN (SELECT adverb_id FROM adjective_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
297 | conditions.push_back(cond); | ||
298 | } | ||
299 | |||
300 | if (!_derived_from_adverb.empty()) | ||
301 | { | ||
302 | std::list<std::string> clauses(_derived_from_adverb.size(), "adverb_2_id = @DERADV"); | ||
303 | std::string cond = "adverb_id IN (SELECT adverb_1_id FROM adverb_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
304 | conditions.push_back(cond); | ||
305 | } | ||
306 | |||
307 | if (!_not_derived_from_adverb.empty()) | ||
308 | { | ||
309 | std::list<std::string> clauses(_not_derived_from_adverb.size(), "adverb_2_id = @NDERADV"); | ||
310 | std::string cond = "adverb_id NOT IN (SELECT adverb_1_id FROM adverb_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
311 | conditions.push_back(cond); | ||
312 | } | ||
313 | |||
314 | if (!_derived_from_noun.empty()) | ||
315 | { | ||
316 | std::list<std::string> clauses(_derived_from_noun.size(), "noun_id = @DERN"); | ||
317 | std::string cond = "adverb_id IN (SELECT adverb_id FROM noun_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
318 | conditions.push_back(cond); | ||
319 | } | ||
320 | |||
321 | if (!_not_derived_from_noun.empty()) | ||
322 | { | ||
323 | std::list<std::string> clauses(_not_derived_from_noun.size(), "noun_id = @NDERN"); | ||
324 | std::string cond = "adverb_id NOT IN (SELECT adverb_id FROM noun_adverb_derivation WHERE " + verbly::implode(std::begin(clauses), std::end(clauses), " OR ") + ")"; | ||
325 | conditions.push_back(cond); | ||
326 | } | ||
327 | |||
328 | if (!conditions.empty()) | ||
329 | { | ||
330 | construct << " WHERE "; | ||
331 | construct << verbly::implode(std::begin(conditions), std::end(conditions), " AND "); | ||
332 | } | ||
333 | |||
334 | if (_random) | ||
335 | { | ||
336 | construct << " ORDER BY RANDOM()"; | ||
337 | } | ||
338 | |||
339 | if (_limit != unlimited) | ||
340 | { | ||
341 | construct << " LIMIT " << _limit; | ||
342 | } | ||
343 | |||
344 | sqlite3_stmt* ppstmt; | ||
345 | std::string query = construct.str(); | ||
346 | if (sqlite3_prepare_v2(_data.ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
347 | { | ||
348 | throw std::runtime_error(sqlite3_errmsg(_data.ppdb)); | ||
349 | } | ||
350 | |||
351 | if (!_rhymes.empty()) | ||
352 | { | ||
353 | int i = 0; | ||
354 | for (auto rhyme : _rhymes) | ||
355 | { | ||
356 | std::string rhymer = "%" + rhyme; | ||
357 | sqlite3_bind_text(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@RHMPRN"), rhymer.c_str(), rhymer.length(), SQLITE_STATIC); | ||
358 | |||
359 | i++; | ||
360 | } | ||
361 | } | ||
362 | |||
363 | for (auto except : _except) | ||
364 | { | ||
365 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@EXCID"), except._id); | ||
366 | } | ||
367 | |||
368 | for (auto antonym : _antonym_of) | ||
369 | { | ||
370 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@ANTID"), antonym._id); | ||
371 | } | ||
372 | |||
373 | for (auto antonym : _not_antonym_of) | ||
374 | { | ||
375 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@NANTID"), antonym._id); | ||
376 | } | ||
377 | |||
378 | for (auto synonym : _synonym_of) | ||
379 | { | ||
380 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@SYNID"), synonym._id); | ||
381 | } | ||
382 | |||
383 | for (auto synonym : _not_synonym_of) | ||
384 | { | ||
385 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@NSYNID"), synonym._id); | ||
386 | } | ||
387 | |||
388 | for (auto adj : _mannernym_of) | ||
389 | { | ||
390 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@AMANID"), adj._id); | ||
391 | } | ||
392 | |||
393 | for (auto adj : _derived_from_adjective) | ||
394 | { | ||
395 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@DERADJ"), adj._id); | ||
396 | } | ||
397 | |||
398 | for (auto adj : _not_derived_from_adjective) | ||
399 | { | ||
400 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@NDERADJ"), adj._id); | ||
401 | } | ||
402 | |||
403 | for (auto adv : _derived_from_adverb) | ||
404 | { | ||
405 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@DERADV"), adv._id); | ||
406 | } | ||
407 | |||
408 | for (auto adv : _not_derived_from_adverb) | ||
409 | { | ||
410 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@NDERADV"), adv._id); | ||
411 | } | ||
412 | |||
413 | for (auto n : _derived_from_noun) | ||
414 | { | ||
415 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@DERN"), n._id); | ||
416 | } | ||
417 | |||
418 | for (auto n : _not_derived_from_noun) | ||
419 | { | ||
420 | sqlite3_bind_int(ppstmt, sqlite3_bind_parameter_index(ppstmt, "@NDERN"), n._id); | ||
421 | } | ||
422 | |||
423 | std::list<adverb> output; | ||
424 | while (sqlite3_step(ppstmt) == SQLITE_ROW) | ||
425 | { | ||
426 | adverb tnc {_data, sqlite3_column_int(ppstmt, 0)}; | ||
427 | tnc._base_form = std::string(reinterpret_cast<const char*>(sqlite3_column_text(ppstmt, 1))); | ||
428 | |||
429 | if (sqlite3_column_type(ppstmt, 2) != SQLITE_NULL) | ||
430 | { | ||
431 | tnc._comparative_form = std::string(reinterpret_cast<const char*>(sqlite3_column_text(ppstmt, 2))); | ||
432 | } | ||
433 | |||
434 | if (sqlite3_column_type(ppstmt, 3) != SQLITE_NULL) | ||
435 | { | ||
436 | tnc._superlative_form = std::string(reinterpret_cast<const char*>(sqlite3_column_text(ppstmt, 3))); | ||
437 | } | ||
438 | |||
439 | output.push_back(tnc); | ||
440 | } | ||
441 | |||
442 | sqlite3_finalize(ppstmt); | ||
443 | |||
444 | for (auto& adverb : output) | ||
445 | { | ||
446 | query = "SELECT pronunciation FROM adverb_pronunciations WHERE adverb_id = ?"; | ||
447 | if (sqlite3_prepare_v2(_data.ppdb, query.c_str(), query.length(), &ppstmt, NULL) != SQLITE_OK) | ||
448 | { | ||
449 | throw std::runtime_error(sqlite3_errmsg(_data.ppdb)); | ||
450 | } | ||
451 | |||
452 | sqlite3_bind_int(ppstmt, 1, adverb._id); | ||
453 | |||
454 | while (sqlite3_step(ppstmt) == SQLITE_ROW) | ||
455 | { | ||
456 | std::string pronunciation(reinterpret_cast<const char*>(sqlite3_column_text(ppstmt, 0))); | ||
457 | auto phonemes = verbly::split<std::list<std::string>>(pronunciation, " "); | ||
458 | |||
459 | adverb.pronunciations.push_back(phonemes); | ||
460 | } | ||
461 | |||
462 | sqlite3_finalize(ppstmt); | ||
463 | } | ||
464 | |||
465 | return output; | ||
466 | } | 69 | } |
467 | 70 | ||
468 | }; | 71 | }; |