public WordIndex()
{
this.Words = new String8Set();
this.Counts = new List <int>();
this.Index = new Dictionary <String8, int>();
this.Indexed = true;
}
public int[] Optimize()
{
if (!this.Indexed)
{
Reindex();
}
int[] remapping = new int[Count];
int[] indices = new int[Count];
for (int i = 0; i < indices.Length; ++i)
{
indices[i] = i;
}
// First, sort words in descending frequency order
Array.Sort(indices, new CountDescendingComparer(this));
// Next, within each set with the same byte length, sort by ordinal
IComparer <int> wordComparer = new WordComparer(this);
int countDone = 0;
int countForLength = 1 << WordCompressor.BitsPerByte;
do
{
int countToDo = Math.Min(countForLength, Count - countDone);
Array.Sort(indices, countDone, countToDo, wordComparer);
countDone += countToDo;
countForLength = countForLength << WordCompressor.BitsPerByte;
} while (countDone < Count);
// Look up the old index for each word to map to the new index
for (int i = 0; i < Count; ++i)
{
remapping[indices[i]] = i;
}
// Rebuild the word set
String8Set newSet = new String8Set(Count, Words.LengthBytes);
for (int i = 0; i < Count; ++i)
{
newSet.Add(Words[indices[i]]);
}
Words = newSet;
// Clear the index (rebuild if needed later)
Index.Clear();
Indexed = false;
return(remapping);
}
public String8SetEnumerator(String8Set set)
{
_set = set;
_index = -1;
}