public UintSet Occurances(IEnumerable <char> sequence)
{
if (!sequence.Any())
{
return(new UintSet());
}
if (!chars.ContainsKey(sequence.First()))
{
return(new UintSet());
}
UintSet locations = chars[sequence.First()];
int i = 1;
foreach (var c in sequence.Skip(1))
{
if (!chars.ContainsKey(c))
{
return(new UintSet());
}
var next = chars[c];
locations = locations.Intersect(next << i++);
if (locations.Count == 0)
{
break;
}
}
return(locations);
}
public void Immutable64FasterIntersectionThanSetT()
{
var hash_set_a = ImmutableHashSet <uint> .Empty;
var hash_set_b = ImmutableHashSet <uint> .Empty;
var simd_set_a = new UintSet();
var simd_set_b = new UintSet();
for (uint i = 0; i < 1000; i++)
{
if (i % 3 == 0)
{
hash_set_a = hash_set_a.Add(i);
}
if (i % 2 == 0)
{
hash_set_b = hash_set_b.Add(i);
}
}
for (uint i = 0; i < 1000; i++)
{
if (i % 3 == 0)
{
simd_set_a = simd_set_a.Add(i);
}
if (i % 2 == 0)
{
simd_set_b = simd_set_b.Add(i);
}
}
var timer = new Stopwatch();
timer.Start();
for (int i = 0; i < 100_000; i++)
{
_ = hash_set_a.Intersect(hash_set_b);
}
timer.Stop();
double time1 = timer.Elapsed.TotalMilliseconds;
timer.Restart();
for (int i = 0; i < 100_000; i++)
{
_ = simd_set_a.Intersect(simd_set_b);
}
timer.Stop();
double time2 = timer.Elapsed.TotalMilliseconds;
double ratio = time1 / time2;
time2.Should().BeLessThan(time1);
hash_set_a.Intersect(hash_set_b).ToList().Should().BeEquivalentTo(simd_set_a.Intersect(simd_set_b).ToList());
}
public SubstringCollection(IEnumerable <char> value, bool caseSensitive = false)
{
CaseSensative = caseSensitive;
uint i = 0;
foreach (var character in value)
{
char c = character;
if (!CaseSensative && char.IsLetter(c))
{
c = char.ToUpperInvariant(c);
}
if (!chars.ContainsKey(c))
{
chars[c] = new UintSet();
}
chars[c] = chars[c].Add(i++);
}
}
public UintSet Occurances(IEnumerable <char> sequence, UintSet starts)
{
if (!sequence.Any())
{
return(starts);
}
UintSet locations = starts;
int i = 0;
foreach (var c in sequence)
{
if (!chars.ContainsKey(c))
{
return(new UintSet());
}
var next = chars[c];
locations = locations.Intersect(next << i++);
if (locations.Count == 0)
{
break;
}
}
return(locations);
}
