package ivory.bloomir.util;

import java.util.List;
import java.util.SortedMap;
import java.util.Map;

import com.google.common.base.Preconditions;
import com.google.common.collect.Lists;
import com.google.common.collect.Maps;

import ivory.core.data.stat.SpamPercentileScore;

/**
 * @author Nima Asadi
 */
public class DocumentUtility {
  /**
   * Generates new document ids based on spam/quality scores.
   *
   * @param spamScores Spam percentile scores
   * @return Array of integers where the value at index i
   *         is the new document id for document i.
   */
  public static int[] spamSortDocids(SpamPercentileScore spamScores) {
    Preconditions.checkNotNull(spamScores);

    int[] newDocids = new int[spamScores.getDocCount() + 1];
    SortedMap<Integer, List<Integer>> spamSortedDocids = Maps.newTreeMap();

    for (int i = 1; i < newDocids.length; i++) {
      int score = -spamScores.getRawScore(i);
      if(!spamSortedDocids.containsKey(score)) {
        List<Integer> list = Lists.newArrayList();
        spamSortedDocids.put(score, list);
      }
      spamSortedDocids.get(score).add(i);
    }

    int index = 1;
    for(Map.Entry<Integer, List<Integer>> entry : spamSortedDocids.entrySet()) {
      for(int docno: entry.getValue()) {
        newDocids[docno] = index++;
      }
    }

    return newDocids;
  }

  /**
   * Retrieves old docid for spam sorted docids generated by {@link #spamSortDocids},
   * thereby providing a reverse-look up table.
   *
   * @param spamSortedDocidsMap Array of integers generated
   *        by {@link #spamSortDocids}
   * @return An array of integers where the value at index i
   *         is the old document id for document i
   */
  public static int[] reverseLookupSpamSortedDocids(int[] spamSortedDocidsMap) {
    Preconditions.checkNotNull(spamSortedDocidsMap);

    int[] buffer = new int[spamSortedDocidsMap.length];

    for(int i = 0; i < buffer.length; i++) {
      buffer[spamSortedDocidsMap[i]] = i;
    }

    return buffer;
  }
}