| public Build ( MetricDB original, Index refs, int k ) : void | ||
| original | MetricDB | |
| refs | Index | |
| k | int | |
| return | void | |
public void Build(MetricDB db, Index refs, int K=7, int maxcand=1024, SequenceBuilder seq_builder=null)
{
var knrfp = new KnrFP ();
knrfp.Build (db, refs, K);
this.Build (db, knrfp, maxcand, seq_builder);
}
public void Build(MetricDB db, int symbolsPerHash, int neighborhoodExpansion, Random rand)
{
this.DB = db;
int n = db.Count;
this.SymbolsPerHash = symbolsPerHash; // very small values, i.e., 2, 3, 4
this.NeighborhoodExpansion = neighborhoodExpansion; // neighborhoodExpansion >= symbolsPerHash
var numrefs = (int)(Math.Pow(n, 1.0 / this.SymbolsPerHash));
this.CountSymbolBits = countBits(numrefs);
var refs = new SampleSpace("", db, numrefs, rand);
var sat = new SAT_Distal ();
sat.Build (refs, new Random ());
this.R = sat;
var G = new KnrFP ();
G.Build (db, this.R, this.SymbolsPerHash);
var knrcopy = new int[this.SymbolsPerHash];
this.hashTable = new Dictionary<long, List<int>> ();
for (int objID = 0; objID < n; ++objID) {
var knr = G [objID] as int[];
knr.CopyTo (knrcopy, 0);
var hash = this.EncodeKnr (knrcopy); // EncodeKnr destroys the reference order
List<int> L;
if (!this.hashTable.TryGetValue(hash, out L)) {
L = new List<int>();
this.hashTable.Add(hash, L);
}
L.Add (objID);
}
double avg_len = 0;
double avg_len_sq = 0;
foreach (var list in this.hashTable.Values) {
avg_len += list.Count;
avg_len_sq += list.Count * list.Count;
}
avg_len /= this.hashTable.Count;
avg_len_sq /= this.hashTable.Count;
Console.WriteLine ("=== created hash table with {0} keys, items: {1}, popcount mean: {2}, popcount stddev: {3}",
this.hashTable.Count, n, avg_len, Math.Sqrt(avg_len_sq - avg_len * avg_len));
}