/**
* Reads the dataReader into a StringBuilder, creates a list of Myrrix "Items" and uses
* TranslatingRecommender.ingest and TranslatingRecommender.addItemIDs to train the Myrrix
* instance.
*
* @see Recommender.train
* @param dataReader
* @throws IOException
* @throws TasteException
*/
@Override
public void train(Reader dataReader) throws IOException, TasteException {
BufferedReader br = new BufferedReader(dataReader);
ArrayList<String> myrrixItemList = new ArrayList<String>();
String temp;
StringBuilder dataBuilder = new StringBuilder();
while ((temp = br.readLine()) != null) {
if (temp.contains(",")) {
// Assume dataReader cannot be reset. Cache the data as it is read, to be fed into Myrrix
// translatingRecommender later.
dataBuilder.append(temp).append("\n");
String[] chunks = temp.split(",");
String property = "";
for (int i = 1;
i < chunks.length - 1;
i++) { // Extract only the middle portion (the Myrrix "Item", not the "User")
property += chunks[i];
}
myrrixItemList.add(property); // Creating the list of Myrrix "Items"
}
}
translatingRecommender.ingest(new StringReader(dataBuilder.toString()));
translatingRecommender.addItemIDs(myrrixItemList);
}
/**
* Provide recommendations based on the Myrrix Items in input using collaborative filtering.
*
* @see Recommender.recommend
* @param input
* @param howMany
* @return
* @throws TasteException
*/
@Override
public List<TranslatedRecommendedItem> recommend(List<String> input, int howMany)
throws TasteException {
String[] inputArray = input.toArray(new String[1]);
return translatingRecommender.recommendToAnonymous(inputArray, howMany);
}
