private static void AssertAutomaton(Automaton a)
{
Automaton clone = (Automaton)a.Clone();
// complement(complement(a)) = a
Automaton equivalent = BasicOperations.Complement(BasicOperations.Complement(a));
Assert.IsTrue(BasicOperations.SameLanguage(a, equivalent));
// a union a = a
equivalent = BasicOperations.Union(a, clone);
Assert.IsTrue(BasicOperations.SameLanguage(a, equivalent));
// a intersect a = a
equivalent = BasicOperations.Intersection(a, clone);
Assert.IsTrue(BasicOperations.SameLanguage(a, equivalent));
// a minus a = empty
Automaton empty = BasicOperations.Minus(a, clone);
Assert.IsTrue(BasicOperations.IsEmpty(empty));
// as long as don't accept the empty string
// then optional(a) - empty = a
if (!BasicOperations.Run(a, ""))
{
//System.out.println("test " + a);
Automaton optional = BasicOperations.Optional(a);
//System.out.println("optional " + optional);
equivalent = BasicOperations.Minus(optional, BasicAutomata.MakeEmptyString());
//System.out.println("equiv " + equivalent);
Assert.IsTrue(BasicOperations.SameLanguage(a, equivalent));
}
}
public virtual void TestEmptyLanguageConcatenate()
{
Automaton a = BasicAutomata.MakeString("a");
Automaton concat = BasicOperations.Concatenate(a, BasicAutomata.MakeEmpty());
Assert.IsTrue(BasicOperations.IsEmpty(concat));
}
/// <summary>
/// Returns an automaton that accepts the union of the languages of the given
/// automata.
/// <para/>
/// Complexity: linear in number of states.
/// </summary>
public static Automaton Union(ICollection <Automaton> l)
{
JCG.HashSet <int> ids = new JCG.HashSet <int>();
foreach (Automaton a in l)
{
ids.Add(a.GetHashCode());
}
bool has_aliases = ids.Count != l.Count;
State s = new State();
foreach (Automaton b in l)
{
if (BasicOperations.IsEmpty(b))
{
continue;
}
Automaton bb = b;
if (has_aliases)
{
bb = bb.CloneExpanded();
}
else
{
bb = bb.CloneExpandedIfRequired();
}
s.AddEpsilon(bb.initial);
}
Automaton a_ = new Automaton
{
initial = s,
deterministic = false
};
//a.clearHashCode();
a_.ClearNumberedStates();
a_.CheckMinimizeAlways();
return(a_);
}
/// <summary>
/// Returns a (deterministic) automaton that accepts the intersection of the
/// language of <paramref name="a1"/> and the complement of the language of
/// <paramref name="a2"/>. As a side-effect, the automata may be determinized, if not
/// already deterministic.
/// <para/>
/// Complexity: quadratic in number of states (if already deterministic).
/// </summary>
public static Automaton Minus(Automaton a1, Automaton a2)
{
if (BasicOperations.IsEmpty(a1) || a1 == a2)
{
return(BasicAutomata.MakeEmpty());
}
if (BasicOperations.IsEmpty(a2))
{
return(a1.CloneIfRequired());
}
if (a1.IsSingleton)
{
if (BasicOperations.Run(a2, a1.singleton))
{
return(BasicAutomata.MakeEmpty());
}
else
{
return(a1.CloneIfRequired());
}
}
return(Intersection(a1, a2.Complement()));
}
public virtual void TestGetRandomAcceptedString()
{
int ITER1 = AtLeast(100);
int ITER2 = AtLeast(100);
for (int i = 0; i < ITER1; i++)
{
RegExp re = new RegExp(AutomatonTestUtil.RandomRegexp(Random), RegExpSyntax.NONE);
Automaton a = re.ToAutomaton();
Assert.IsFalse(BasicOperations.IsEmpty(a));
RandomAcceptedStrings rx = new RandomAcceptedStrings(a);
for (int j = 0; j < ITER2; j++)
{
int[] acc = null;
try
{
acc = rx.GetRandomAcceptedString(Random);
string s = UnicodeUtil.NewString(acc, 0, acc.Length);
Assert.IsTrue(BasicOperations.Run(a, s));
}
catch (Exception /*t*/)
{
Console.WriteLine("regexp: " + re);
if (acc != null)
{
Console.WriteLine("fail acc re=" + re + " count=" + acc.Length);
for (int k = 0; k < acc.Length; k++)
{
Console.WriteLine(" " + acc[k].ToString("x"));
}
}
throw; // LUCENENET: CA2200: Rethrow to preserve stack details (https://docs.microsoft.com/en-us/visualstudio/code-quality/ca2200-rethrow-to-preserve-stack-details)
}
}
}
}
public virtual void TestGetRandomAcceptedString()
{
int ITER1 = AtLeast(100);
int ITER2 = AtLeast(100);
for (int i = 0; i < ITER1; i++)
{
RegExp re = new RegExp(AutomatonTestUtil.RandomRegexp(Random()), RegExp.NONE);
Automaton a = re.ToAutomaton();
Assert.IsFalse(BasicOperations.IsEmpty(a));
AutomatonTestUtil.RandomAcceptedStrings rx = new AutomatonTestUtil.RandomAcceptedStrings(a);
for (int j = 0; j < ITER2; j++)
{
int[] acc = null;
try
{
acc = rx.GetRandomAcceptedString(Random());
string s = UnicodeUtil.NewString(acc, 0, acc.Length);
Assert.IsTrue(BasicOperations.Run(a, s));
}
catch (Exception t)
{
Console.WriteLine("regexp: " + re);
if (acc != null)
{
Console.WriteLine("fail acc re=" + re + " count=" + acc.Length);
for (int k = 0; k < acc.Length; k++)
{
Console.WriteLine(" " + acc[k].ToString("x"));
}
}
throw t;
}
}
}
}
public CompiledAutomaton(Automaton automaton, bool?finite, bool simplify)
{
if (simplify)
{
// Test whether the automaton is a "simple" form and
// if so, don't create a runAutomaton. Note that on a
// large automaton these tests could be costly:
if (BasicOperations.IsEmpty(automaton))
{
// matches nothing
Type = AUTOMATON_TYPE.NONE;
Term = null;
CommonSuffixRef = null;
RunAutomaton = null;
sortedTransitions = null;
this.Finite = null;
return;
}
else if (BasicOperations.IsTotal(automaton))
{
// matches all possible strings
Type = AUTOMATON_TYPE.ALL;
Term = null;
CommonSuffixRef = null;
RunAutomaton = null;
sortedTransitions = null;
this.Finite = null;
return;
}
else
{
string commonPrefix;
string singleton;
if (automaton.Singleton == null)
{
commonPrefix = SpecialOperations.GetCommonPrefix(automaton);
if (commonPrefix.Length > 0 && BasicOperations.SameLanguage(automaton, BasicAutomata.MakeString(commonPrefix)))
{
singleton = commonPrefix;
}
else
{
singleton = null;
}
}
else
{
commonPrefix = null;
singleton = automaton.Singleton;
}
if (singleton != null)
{
// matches a fixed string in singleton or expanded
// representation
Type = AUTOMATON_TYPE.SINGLE;
Term = new BytesRef(singleton);
CommonSuffixRef = null;
RunAutomaton = null;
sortedTransitions = null;
this.Finite = null;
return;
}
else if (BasicOperations.SameLanguage(automaton, BasicOperations.Concatenate(BasicAutomata.MakeString(commonPrefix), BasicAutomata.MakeAnyString())))
{
// matches a constant prefix
Type = AUTOMATON_TYPE.PREFIX;
Term = new BytesRef(commonPrefix);
CommonSuffixRef = null;
RunAutomaton = null;
sortedTransitions = null;
this.Finite = null;
return;
}
}
}
Type = AUTOMATON_TYPE.NORMAL;
Term = null;
if (finite == null)
{
this.Finite = SpecialOperations.IsFinite(automaton);
}
else
{
this.Finite = finite;
}
Automaton utf8 = (new UTF32ToUTF8()).Convert(automaton);
if (this.Finite == true)
{
CommonSuffixRef = null;
}
else
{
CommonSuffixRef = SpecialOperations.GetCommonSuffixBytesRef(utf8);
}
RunAutomaton = new ByteRunAutomaton(utf8, true);
sortedTransitions = utf8.GetSortedTransitions();
}
/// <summary>
/// Returns an automaton that accepts the concatenation of the languages of the
/// given automata.
/// <para/>
/// Complexity: linear in total number of states.
/// </summary>
public static Automaton Concatenate(IList <Automaton> l)
{
if (l.Count == 0)
{
return(BasicAutomata.MakeEmptyString());
}
bool all_singleton = true;
foreach (Automaton a in l)
{
if (!a.IsSingleton)
{
all_singleton = false;
break;
}
}
if (all_singleton)
{
StringBuilder b = new StringBuilder();
foreach (Automaton a in l)
{
b.Append(a.singleton);
}
return(BasicAutomata.MakeString(b.ToString()));
}
else
{
foreach (Automaton a in l)
{
if (BasicOperations.IsEmpty(a))
{
return(BasicAutomata.MakeEmpty());
}
}
HashSet <int> ids = new HashSet <int>();
foreach (Automaton a in l)
{
ids.Add(a.GetHashCode());
}
bool has_aliases = ids.Count != l.Count;
Automaton b = l[0];
if (has_aliases)
{
b = b.CloneExpanded();
}
else
{
b = b.CloneExpandedIfRequired();
}
ISet <State> ac = b.GetAcceptStates();
bool first = true;
foreach (Automaton a in l)
{
if (first)
{
first = false;
}
else
{
if (a.IsEmptyString)
{
continue;
}
Automaton aa = a;
if (has_aliases)
{
aa = aa.CloneExpanded();
}
else
{
aa = aa.CloneExpandedIfRequired();
}
ISet <State> ns = aa.GetAcceptStates();
foreach (State s in ac)
{
s.accept = false;
s.AddEpsilon(aa.initial);
if (s.accept)
{
ns.Add(s);
}
}
ac = ns;
}
}
b.deterministic = false;
//b.clearHashCode();
b.ClearNumberedStates();
b.CheckMinimizeAlways();
return(b);
}
}
