diff options
Diffstat (limited to 'generator/generator.h')
-rw-r--r-- | generator/generator.h | 151 |
1 files changed, 151 insertions, 0 deletions
diff --git a/generator/generator.h b/generator/generator.h new file mode 100644 index 0000000..e2a7404 --- /dev/null +++ b/generator/generator.h | |||
@@ -0,0 +1,151 @@ | |||
1 | #ifndef GENERATOR_H_5B61CBC5 | ||
2 | #define GENERATOR_H_5B61CBC5 | ||
3 | |||
4 | #include <string> | ||
5 | #include <map> | ||
6 | #include <list> | ||
7 | #include <set> | ||
8 | #include <libxml/parser.h> | ||
9 | #include "database.h" | ||
10 | #include "notion.h" | ||
11 | #include "word.h" | ||
12 | #include "lemma.h" | ||
13 | #include "form.h" | ||
14 | #include "pronunciation.h" | ||
15 | #include "group.h" | ||
16 | #include "frame.h" | ||
17 | |||
18 | namespace verbly { | ||
19 | namespace generator { | ||
20 | |||
21 | enum class part_of_speech; | ||
22 | class selrestr; | ||
23 | |||
24 | class generator { | ||
25 | public: | ||
26 | |||
27 | // Constructor | ||
28 | |||
29 | generator( | ||
30 | std::string verbNetPath, | ||
31 | std::string agidPath, | ||
32 | std::string wordNetPath, | ||
33 | std::string cmudictPath, | ||
34 | std::string imageNetPath, | ||
35 | std::string outputPath); | ||
36 | |||
37 | // Action | ||
38 | |||
39 | void run(); | ||
40 | |||
41 | private: | ||
42 | |||
43 | // Subroutines | ||
44 | |||
45 | void readWordNetSynsets(); | ||
46 | |||
47 | void readAdjectivePositioning(); | ||
48 | |||
49 | void readImageNetUrls(); | ||
50 | |||
51 | void readWordNetSenseKeys(); | ||
52 | |||
53 | void readVerbNet(); | ||
54 | |||
55 | void readAgidInflections(); | ||
56 | |||
57 | void readPrepositions(); | ||
58 | |||
59 | void readCmudictPronunciations(); | ||
60 | |||
61 | void writeSchema(); | ||
62 | |||
63 | void dumpObjects(); | ||
64 | |||
65 | void readWordNetAntonymy(); | ||
66 | |||
67 | void readWordNetVariation(); | ||
68 | |||
69 | void readWordNetClasses(); | ||
70 | |||
71 | void readWordNetCausality(); | ||
72 | |||
73 | void readWordNetEntailment(); | ||
74 | |||
75 | void readWordNetHypernymy(); | ||
76 | |||
77 | void readWordNetInstantiation(); | ||
78 | |||
79 | void readWordNetMemberMeronymy(); | ||
80 | |||
81 | void readWordNetPartMeronymy(); | ||
82 | |||
83 | void readWordNetSubstanceMeronymy(); | ||
84 | |||
85 | void readWordNetPertainymy(); | ||
86 | |||
87 | void readWordNetSpecification(); | ||
88 | |||
89 | void readWordNetSimilarity(); | ||
90 | |||
91 | // Helpers | ||
92 | |||
93 | std::list<std::string> readFile(std::string path); | ||
94 | |||
95 | inline part_of_speech partOfSpeechByWnid(int wnid); | ||
96 | |||
97 | notion& createNotion(part_of_speech partOfSpeech); | ||
98 | |||
99 | notion& lookupOrCreateNotion(int wnid); | ||
100 | |||
101 | lemma& lookupOrCreateLemma(std::string base_form); | ||
102 | |||
103 | form& lookupOrCreateForm(std::string text); | ||
104 | |||
105 | template <typename... Args> word& createWord(Args&&... args); | ||
106 | |||
107 | group& createGroup(xmlNodePtr top); | ||
108 | |||
109 | selrestr parseSelrestr(xmlNodePtr top); | ||
110 | |||
111 | // Input | ||
112 | |||
113 | std::string verbNetPath_; | ||
114 | std::string agidPath_; | ||
115 | std::string wordNetPath_; | ||
116 | std::string cmudictPath_; | ||
117 | std::string imageNetPath_; | ||
118 | |||
119 | // Output | ||
120 | |||
121 | database db_; | ||
122 | |||
123 | // Data | ||
124 | |||
125 | std::list<notion> notions_; | ||
126 | std::list<word> words_; | ||
127 | std::list<lemma> lemmas_; | ||
128 | std::list<form> forms_; | ||
129 | std::list<pronunciation> pronunciations_; | ||
130 | std::list<frame> frames_; | ||
131 | std::list<group> groups_; | ||
132 | |||
133 | // Indexes | ||
134 | |||
135 | std::map<int, notion*> notionByWnid_; | ||
136 | std::map<int, std::set<word*>> wordsByWnid_; | ||
137 | std::map<std::pair<int, int>, word*> wordByWnidAndWnum_; | ||
138 | std::map<std::string, std::set<word*>> wordsByBaseForm_; | ||
139 | std::map<std::string, lemma*> lemmaByBaseForm_; | ||
140 | std::map<std::string, form*> formByText_; | ||
141 | |||
142 | // Caches | ||
143 | |||
144 | std::map<std::string, word*> wnSenseKeys_; | ||
145 | |||
146 | }; | ||
147 | |||
148 | }; | ||
149 | }; | ||
150 | |||
151 | #endif /* end of include guard: GENERATOR_H_5B61CBC5 */ | ||