public void testOverlappedTokensSausage() throws Exception {
// Two tokens on top of each other (sausage):
final TokenStream ts =
new CannedTokenStream(new Token[] {token("abc", 1, 1), token("xyz", 0, 1)});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton a1 = BasicAutomata.makeString("abc");
final Automaton a2 = BasicAutomata.makeString("xyz");
final Automaton expected = BasicOperations.union(a1, a2);
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
public void testSynHangingOverEnd() throws Exception {
final TokenStream ts =
new CannedTokenStream(
new Token[] {
token("a", 1, 1), token("X", 0, 10),
});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton expected =
BasicOperations.union(BasicAutomata.makeString("a"), BasicAutomata.makeString("X"));
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
/** Test a configuration that behaves a lot like KeepWordFilter */
public void testKeep() throws Exception {
CharacterRunAutomaton keepWords =
new CharacterRunAutomaton(
BasicOperations.complement(
Automaton.union(
Arrays.asList(
BasicAutomata.makeString("foo"), BasicAutomata.makeString("bar")))));
Analyzer a = new MockAnalyzer(MockTokenizer.SIMPLE, true, keepWords, true);
assertAnalyzesTo(
a,
"quick foo brown bar bar fox foo",
new String[] {"foo", "bar", "bar", "foo"},
new int[] {2, 2, 1, 2});
}
private Automaton join(String... strings) {
List<Automaton> as = new ArrayList<Automaton>();
for (String s : strings) {
as.add(BasicAutomata.makeString(s));
as.add(SEP_A);
}
as.remove(as.size() - 1);
return BasicOperations.concatenate(as);
}
public void testSingleToken() throws Exception {
final TokenStream ts =
new CannedTokenStream(
new Token[] {
token("abc", 1, 1),
});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton expected = BasicAutomata.makeString("abc");
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
public void testSynOverHole2() throws Exception {
final TokenStream ts =
new CannedTokenStream(
new Token[] {
token("xyz", 1, 1), token("abc", 0, 3), token("def", 2, 1),
});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton expected =
BasicOperations.union(
join(s2a("xyz"), SEP_A, HOLE_A, SEP_A, s2a("def")), BasicAutomata.makeString("abc"));
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
public void testOverlappedTokensLattice2() throws Exception {
final TokenStream ts =
new CannedTokenStream(
new Token[] {
token("abc", 1, 1), token("xyz", 0, 3), token("def", 1, 1), token("ghi", 1, 1),
});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton a1 = BasicAutomata.makeString("xyz");
final Automaton a2 = join("abc", "def", "ghi");
final Automaton expected = BasicOperations.union(a1, a2);
// toDot(actual);
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
public void testSynOverHole() throws Exception {
final TokenStream ts =
new CannedTokenStream(
new Token[] {
token("a", 1, 1), token("X", 0, 2), token("b", 2, 1),
});
final Automaton actual = (new TokenStreamToAutomaton()).toAutomaton(ts);
final Automaton a1 =
BasicOperations.union(join(s2a("a"), SEP_A, HOLE_A), BasicAutomata.makeString("X"));
final Automaton expected = BasicOperations.concatenate(a1, join(SEP_A, s2a("b")));
// toDot(actual);
assertTrue(BasicOperations.sameLanguage(expected, actual));
}
private Automaton s2a(String s) {
return BasicAutomata.makeString(s);
}
/**
* Extracts all MultiTermQueries for {@code field}, and returns equivalent automata that will
* match terms.
*/
static CharacterRunAutomaton[] extractAutomata(Query query, String field) {
List<CharacterRunAutomaton> list = new ArrayList<>();
if (query instanceof BooleanQuery) {
BooleanClause clauses[] = ((BooleanQuery) query).getClauses();
for (BooleanClause clause : clauses) {
if (!clause.isProhibited()) {
list.addAll(Arrays.asList(extractAutomata(clause.getQuery(), field)));
}
}
} else if (query instanceof DisjunctionMaxQuery) {
for (Query sub : ((DisjunctionMaxQuery) query).getDisjuncts()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanOrQuery) {
for (Query sub : ((SpanOrQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanNearQuery) {
for (Query sub : ((SpanNearQuery) query).getClauses()) {
list.addAll(Arrays.asList(extractAutomata(sub, field)));
}
} else if (query instanceof SpanNotQuery) {
list.addAll(Arrays.asList(extractAutomata(((SpanNotQuery) query).getInclude(), field)));
} else if (query instanceof SpanPositionCheckQuery) {
list.addAll(
Arrays.asList(extractAutomata(((SpanPositionCheckQuery) query).getMatch(), field)));
} else if (query instanceof SpanMultiTermQueryWrapper) {
list.addAll(
Arrays.asList(
extractAutomata(((SpanMultiTermQueryWrapper<?>) query).getWrappedQuery(), field)));
} else if (query instanceof AutomatonQuery) {
final AutomatonQuery aq = (AutomatonQuery) query;
if (aq.getField().equals(field)) {
list.add(
new CharacterRunAutomaton(aq.getAutomaton()) {
@Override
public String toString() {
return aq.toString();
}
});
}
} else if (query instanceof PrefixQuery) {
final PrefixQuery pq = (PrefixQuery) query;
Term prefix = pq.getPrefix();
if (prefix.field().equals(field)) {
list.add(
new CharacterRunAutomaton(
BasicOperations.concatenate(
BasicAutomata.makeString(prefix.text()), BasicAutomata.makeAnyString())) {
@Override
public String toString() {
return pq.toString();
}
});
}
} else if (query instanceof FuzzyQuery) {
final FuzzyQuery fq = (FuzzyQuery) query;
if (fq.getField().equals(field)) {
String utf16 = fq.getTerm().text();
int termText[] = new int[utf16.codePointCount(0, utf16.length())];
for (int cp, i = 0, j = 0; i < utf16.length(); i += Character.charCount(cp)) {
termText[j++] = cp = utf16.codePointAt(i);
}
int termLength = termText.length;
int prefixLength = Math.min(fq.getPrefixLength(), termLength);
String suffix =
UnicodeUtil.newString(termText, prefixLength, termText.length - prefixLength);
LevenshteinAutomata builder = new LevenshteinAutomata(suffix, fq.getTranspositions());
Automaton automaton = builder.toAutomaton(fq.getMaxEdits());
if (prefixLength > 0) {
Automaton prefix =
BasicAutomata.makeString(UnicodeUtil.newString(termText, 0, prefixLength));
automaton = BasicOperations.concatenate(prefix, automaton);
}
list.add(
new CharacterRunAutomaton(automaton) {
@Override
public String toString() {
return fq.toString();
}
});
}
} else if (query instanceof TermRangeQuery) {
final TermRangeQuery tq = (TermRangeQuery) query;
if (tq.getField().equals(field)) {
final CharsRef lowerBound;
if (tq.getLowerTerm() == null) {
lowerBound = null;
} else {
lowerBound = new CharsRef(tq.getLowerTerm().utf8ToString());
}
final CharsRef upperBound;
if (tq.getUpperTerm() == null) {
upperBound = null;
} else {
upperBound = new CharsRef(tq.getUpperTerm().utf8ToString());
}
final boolean includeLower = tq.includesLower();
final boolean includeUpper = tq.includesUpper();
final CharsRef scratch = new CharsRef();
final Comparator<CharsRef> comparator = CharsRef.getUTF16SortedAsUTF8Comparator();
// this is *not* an automaton, but its very simple
list.add(
new CharacterRunAutomaton(BasicAutomata.makeEmpty()) {
@Override
public boolean run(char[] s, int offset, int length) {
scratch.chars = s;
scratch.offset = offset;
scratch.length = length;
if (lowerBound != null) {
int cmp = comparator.compare(scratch, lowerBound);
if (cmp < 0 || (!includeLower && cmp == 0)) {
return false;
}
}
if (upperBound != null) {
int cmp = comparator.compare(scratch, upperBound);
if (cmp > 0 || (!includeUpper && cmp == 0)) {
return false;
}
}
return true;
}
@Override
public String toString() {
return tq.toString();
}
});
}
}
return list.toArray(new CharacterRunAutomaton[list.size()]);
}