So I've made an app that makes sentences, much like a keyboard of words and pictures, what I'd like to achieve is a kind of predictive text where if a user chooses X they are shown Y and Z, user doesnt want either of these and chooses A.
The next time the user picks X they are shown A, Y and Z, so it learns like your keyboard would, my words are stored in a database and right now i have a crude solution where if the user chooses X they are shown Y and Z, and I also track how many times each card is clicked and show them too but it has no context of full sentences or previously made sentences, i hope this is making sense as I've no idea where to start I've been looking at maybe a hashmap full of essentially TAGS as key value pairs and adding to it but I need to research this, I've also looked at candidate view but that would mean declaring my app as an Input Method I believe and I need the normal keyboard for other things, but I'm not sure how that works or if it's remotely viable so I'm hoping someone can educate me here or point to a better solution.
Below I'll post excerpts of what I'm doing thus far
//Check what card was pressed and update predictedCardActivityDB example
//to carry on like this would mean a lot of typing
//and not a lot of smart results
if (cardsChoice.predictive == true) {
String item = cardWriter.getCardSpeech();
switch(item){
case " I":
String[] I_String = {"LIKE","LOVE","WANT"};
predictedCardActivityDB.prepareCardData
(I_String,getActivity(),prefString);
break;
case " I'm":
String[] Im_String =
{"HAPPY","SAD","ANGRY","HUNGRY","FEELING"};
predictedCardActivityDB.prepareCardData
(Im_String,getActivity(),prefString);
break;
And this is the prepareCardData method excerpt nothing to see here just updates the lists based on the string array
public static void prepareCardData
(String[] predictionArray, Context context ,String prefString){
//boring database stuff
DaoMaster.DevOpenHelper helper =
new DaoMaster.DevOpenHelper(context, "ADD_NEW_CARD", null);
SQLiteDatabase db = helper.getWritableDatabase();
DaoMaster daoMaster = new DaoMaster(db);
DaoSession daoSession = daoMaster.newSession();
addNewCardDao leaseDao = daoSession.getAddNewCardDao();
QueryBuilder qb = leaseDao.queryBuilder()
.orderDesc(addNewCardDao.Properties.Clicked);
QueryBuilder qb2 = leaseDao.queryBuilder()
.orderDesc(addNewCardDao.Properties.Clicked);
predictsList.clear();
String[] strings1 = {"ORIGINAL","SIMPLE","PHOTOS", "USER"};
switch (prefString){
case "PHOTOS":
if (predictionArray != null){
//qb gets any card clicked >2
//qb gets all favourites tagged by a boolean
qb.where(addNewCardDao.Properties.CardIconType.in(strings1),
qb.or(addNewCardDao.Properties.Clicked.ge(2),
addNewCardDao.Properties.Fav.ge(true)));
predictsList = qb.list();
//qb2 gets any card that matches a word from the passed in
//string array predictionArray
qb2 = leaseDao.queryBuilder();
qb2.where(addNewCardDao.Properties.CardIconType.in(strings1),
qb2.or(addNewCardDao.Properties.CardName.in(predictionArray),
addNewCardDao.Properties.CardTitle.in(predictionArray)));
temptList = qb2.list();
db.close();
temptList.addAll(predictsList);
predicts_card_adapter.notifyItemInserted(predictedCardActivityDB.temptList.size());
predicts_card_adapter.notifyDataSetChanged();
predicts_card_adapter = new predictsCardAdapter(temptList,itemTouchListener);
predictsrecyclerView.setAdapter(predicts_card_adapter);
}else{
qb.where(addNewCardDao.Properties.CardIconType.in(strings1),
qb.or(addNewCardDao.Properties.Clicked.ge(2),
addNewCardDao.Properties.Fav.ge(true)));
predictsList = qb.list();
//this should work
predicts_card_adapter.notifyDataSetChanged();
predicts_card_adapter = new predictsCardAdapter(predictsList,itemTouchListener);
predictsrecyclerView.setAdapter(predicts_card_adapter);
db.close();
System.out.println("else, predictsList size " + predictsList.size());
}
break;
}
//should be able t delete this and use the two commented out piecess of code above
//the result being the predicted cards coming first in the list
/*
predicts_card_adapter.notifyItemInserted(predictedCardActivityDB.temptList.size());
predicts_card_adapter.notifyDataSetChanged();
predicts_card_adapter = new predictsCardAdapter(temptList,itemTouchListener);
predictsrecyclerView.setAdapter(predicts_card_adapter);
*/
}
