public override void SetUp()
{
base.SetUp();
// build an automaton matching this jvm's letter definition
State initial = new State();
State accept = new State();
accept.Accept = true;
for (int i = 0; i <= 0x10FFFF; i++)
{
if (Character.IsLetter(i))
{
initial.AddTransition(new Transition(i, i, accept));
}
}
Automaton single = new Automaton(initial);
single.Reduce();
Automaton repeat = BasicOperations.Repeat(single);
jvmLetter = new CharacterRunAutomaton(repeat);
}
/// <summary>
/// Compute a DFA that accepts all strings within an edit distance of <code>n</code>.
/// <p>
/// All automata have the following properties:
/// <ul>
/// <li>They are deterministic (DFA).
/// <li>There are no transitions to dead states.
/// <li>They are not minimal (some transitions could be combined).
/// </ul>
/// </p>
/// </summary>
public virtual Automaton ToAutomaton(int n)
{
if (n == 0)
{
return BasicAutomata.MakeString(Word, 0, Word.Length);
}
if (n >= Descriptions.Length)
{
return null;
}
int range = 2 * n + 1;
ParametricDescription description = Descriptions[n];
// the number of states is based on the length of the word and n
State[] states = new State[description.Size()];
// create all states, and mark as accept states if appropriate
for (int i = 0; i < states.Length; i++)
{
states[i] = new State();
states[i].number = i;
states[i].Accept = description.IsAccept(i);
}
// create transitions from state to state
for (int k = 0; k < states.Length; k++)
{
int xpos = description.GetPosition(k);
if (xpos < 0)
{
continue;
}
int end = xpos + Math.Min(Word.Length - xpos, range);
for (int x = 0; x < Alphabet.Length; x++)
{
int ch = Alphabet[x];
// get the characteristic vector at this position wrt ch
int cvec = GetVector(ch, xpos, end);
int dest = description.Transition(k, xpos, cvec);
if (dest >= 0)
{
states[k].AddTransition(new Transition(ch, states[dest]));
}
}
// add transitions for all other chars in unicode
// by definition, their characteristic vectors are always 0,
// because they do not exist in the input string.
int dest_ = description.Transition(k, xpos, 0); // by definition
if (dest_ >= 0)
{
for (int r = 0; r < NumRanges; r++)
{
states[k].AddTransition(new Transition(RangeLower[r], RangeUpper[r], states[dest_]));
}
}
}
Automaton a = new Automaton(states[0]);
a.Deterministic = true;
// we create some useless unconnected states, and its a net-win overall to remove these,
// as well as to combine any adjacent transitions (it makes later algorithms more efficient).
// so, while we could set our numberedStates here, its actually best not to, and instead to
// force a traversal in reduce, pruning the unconnected states while we combine adjacent transitions.
//a.setNumberedStates(states);
a.Reduce();
// we need not trim transitions to dead states, as they are not created.
//a.restoreInvariant();
return a;
}
/// <summary>
/// Compute a DFA that accepts all strings within an edit distance of <paramref name="n"/>.
/// <para>
/// All automata have the following properties:
/// <list type="bullet">
/// <item><description>They are deterministic (DFA).</description></item>
/// <item><description>There are no transitions to dead states.</description></item>
/// <item><description>They are not minimal (some transitions could be combined).</description></item>
/// </list>
/// </para>
/// </summary>
public virtual Automaton ToAutomaton(int n)
{
if (n == 0)
{
return(BasicAutomata.MakeString(word, 0, word.Length));
}
if (n >= descriptions.Length)
{
return(null);
}
int range = 2 * n + 1;
ParametricDescription description = descriptions[n];
// the number of states is based on the length of the word and n
State[] states = new State[description.Count];
// create all states, and mark as accept states if appropriate
for (int i = 0; i < states.Length; i++)
{
states[i] = new State();
states[i].number = i;
states[i].Accept = description.IsAccept(i);
}
// create transitions from state to state
for (int k = 0; k < states.Length; k++)
{
int xpos = description.GetPosition(k);
if (xpos < 0)
{
continue;
}
int end = xpos + Math.Min(word.Length - xpos, range);
for (int x = 0; x < alphabet.Length; x++)
{
int ch = alphabet[x];
// get the characteristic vector at this position wrt ch
int cvec = GetVector(ch, xpos, end);
int dest = description.Transition(k, xpos, cvec);
if (dest >= 0)
{
states[k].AddTransition(new Transition(ch, states[dest]));
}
}
// add transitions for all other chars in unicode
// by definition, their characteristic vectors are always 0,
// because they do not exist in the input string.
int dest_ = description.Transition(k, xpos, 0); // by definition
if (dest_ >= 0)
{
for (int r = 0; r < numRanges; r++)
{
states[k].AddTransition(new Transition(rangeLower[r], rangeUpper[r], states[dest_]));
}
}
}
Automaton a = new Automaton(states[0]);
a.IsDeterministic = true;
// we create some useless unconnected states, and its a net-win overall to remove these,
// as well as to combine any adjacent transitions (it makes later algorithms more efficient).
// so, while we could set our numberedStates here, its actually best not to, and instead to
// force a traversal in reduce, pruning the unconnected states while we combine adjacent transitions.
//a.setNumberedStates(states);
a.Reduce();
// we need not trim transitions to dead states, as they are not created.
//a.restoreInvariant();
return(a);
}
