De-duped pronunciations in generated database hkutil

Identical pronunciations will now share an idea and be re-used by multiple forms. This has a negligible effect on database size, but it's useful for writing queries looking for words with the exact same pronunciations.

This constitutes a minor database update, which we will call d1.2.

1 files changed, 1 insertions, 0 deletions

@@ -139,6 +139,7 @@ namespace verbly {
       std::map<std::string, std::set<word*>> wordsByBaseForm_;
       std::map<std::string, lemma*> lemmaByBaseForm_;
       std::map<std::string, form*> formByText_;
+      std::map<std::string, pronunciation*> pronunciationByPhonemes_;
 
       // Caches