public void TrieConstructorTest()
{
Trie <TrieElement> trie = new Trie <TrieElement>();
Stream stream = Assembly.GetExecutingAssembly().GetManifestResourceStream("VLUnitTests.Resources.TrieTest.txt");
string text = stream.ReadAll(); // read all data from testing file
// add "references" that will be searched
trie.Add("id.aliquam.lorem");
trie.Add("a.b.c.d");
trie.Add("x.y");
trie.Add("eeee.eeee");
// complete building the trie
trie.CreatePredecessorsAndShortcuts();
TrieElement e = trie.Root;
List <string> foundWords = new List <string>();
// create list of expeced results
List <string> expectedWords = new List <string>();
expectedWords.Add("id.aliquam.lorem");
expectedWords.Add("id.aliquam.lorem");
expectedWords.Add("eeee.eeee");
expectedWords.Add("id.aliquam.lorem");
expectedWords.Add("eeee.eeee");
expectedWords.Add("eeee.eeee");
expectedWords.Add("eeee.eeee");
expectedWords.Add("a.b.c.d");
expectedWords.Add("a.b.c.d");
expectedWords.Add("a.b.c.d");
expectedWords.Add("a.b.c.d");
expectedWords.Add("a.b.c.d");
// run the algorithm
foreach (char c in text)
{
e = trie.Step(e, c);
if (e.IsTerminal)
{
foundWords.Add(e.Word);
}
}
// compare with expected
if (foundWords.Count == expectedWords.Count)
{
bool ok = true;
for (int i = 0; i < foundWords.Count; i++)
{
ok = ok && foundWords[i] == expectedWords[i];
}
Assert.IsTrue(ok);
}
else
{
Assert.Fail("Found and expected words count don't match.");
}
}
private void Add(IList <T> key, IEnumerable <T> fullkey)
{
if (!key.Any())
{
_children[new T[0]] = new TrieElement(fullkey, _comp);
return;
}
bool included;
var match = _children.FirstOrDefault(a => a.Key.StartsWith(key, _tokencomp), out included);
if (included)
{
if (match.Key.CompareCount(key) != 0)
{
//we need to split an existing key
_children.Remove(match.Key);
var newchild = new Trie <T>(_comp, _tokencomp)
{
{ new T[0], fullkey }
};
newchild._children[match.Key.Slice(key.Count)] = match.Value;
_children[key] = newchild;
return;
}
// we add epsilon value
var trie = match.Value as Trie <T>;
trie?.Add(new T[0], fullkey);
return;
}
match = _children.FirstOrDefault(a => a.Key.Any() && key.StartsWith(a.Key, _tokencomp), out included);
if (included)
{
//we need to split this key
var trie = match.Value as Trie <T>;
if (trie == null)
{
_children.Remove(match);
trie = new Trie <T>(_comp, _tokencomp)
{
{ new T[0], match.Value.First() }
};
_children[match.Key] = trie;
}
trie.Add(key.Slice(match.Key.Count), fullkey);
return;
}
var submatch = _children.Attach(a => a.Key.LongestCommonPrefix(key, _tokencomp)).FirstOrDefault(a => a.Item2.Any(), out included);
if (included)
{
//we need to split both
match = submatch.Item1;
var prefix = submatch.Item2.Select(a => a.Item1).ToArray();
_children.Remove(match.Key);
var newchild = new Trie <T>(_comp, _tokencomp)
{
{ key.Slice(prefix.Length), fullkey }
};
newchild._children[match.Key.Slice(prefix.Length)] = match.Value;
_children[prefix] = newchild;
return;
}
//we have no definitions for anything like this key
this._children[key] = new TrieElement(fullkey, _comp);
}
