When working with the app some pictos have undefined or miscellaneous wordtype, and also if we plan to implement this NLG somewhere else we might get sentences where no word has a wordtype assigned.
Objectives:
[ ] add method to assume best wordtype for words with no type provided. This differs from the one below because if other words have wordtypes we might infer them from that context.
[ ] add method to assume most probable wordtype if ALL sentence has no wordtype. We could assign a definite type to those word that only have one possible type, and then use the method designed for the objective above to infer the ones that have many.
Steps:
[ ] Create or find a source for wordtypes
[ ] Add a method that only assigns the most probable wordtype to each word
[ ] Find and implement a way to infer when words have many wordtypes
[ ] Join all in one process that fist goes through definitive types and then tries to infer the ones with more than one possibility, if it can't just assign the most probable.
When working with the app some pictos have undefined or miscellaneous wordtype, and also if we plan to implement this NLG somewhere else we might get sentences where no word has a wordtype assigned.
Objectives:
Steps: