/** Test a configuration that behaves a lot like KeepWordFilter */
public void testKeep() throws Exception {
CharacterRunAutomaton keepWords =
new CharacterRunAutomaton(
Operations.complement(
Operations.union(
Arrays.asList(Automata.makeString("foo"), Automata.makeString("bar"))),
DEFAULT_MAX_DETERMINIZED_STATES));
Analyzer a = new MockAnalyzer(random(), MockTokenizer.SIMPLE, true, keepWords);
assertAnalyzesTo(
a,
"quick foo brown bar bar fox foo",
new String[] {"foo", "bar", "bar", "foo"},
new int[] {2, 2, 1, 2});
}
@BeforeClass
public static void beforeClass() throws Exception {
Random random = random();
directory = newDirectory();
stopword = "" + randomChar();
CharacterRunAutomaton stopset = new CharacterRunAutomaton(Automata.makeString(stopword));
analyzer = new MockAnalyzer(random, MockTokenizer.WHITESPACE, false, stopset);
RandomIndexWriter iw = new RandomIndexWriter(random, directory, analyzer);
Document doc = new Document();
Field id = new StringField("id", "", Field.Store.NO);
Field field = new TextField("field", "", Field.Store.NO);
doc.add(id);
doc.add(field);
// index some docs
int numDocs = atLeast(1000);
for (int i = 0; i < numDocs; i++) {
id.setStringValue(Integer.toString(i));
field.setStringValue(randomFieldContents());
iw.addDocument(doc);
}
// delete some docs
int numDeletes = numDocs / 20;
for (int i = 0; i < numDeletes; i++) {
Term toDelete = new Term("id", Integer.toString(random.nextInt(numDocs)));
if (random.nextBoolean()) {
iw.deleteDocuments(toDelete);
} else {
iw.deleteDocuments(new TermQuery(toDelete));
}
}
reader = iw.getReader();
s1 = newSearcher(reader);
s2 = newSearcher(reader);
iw.close();
}
/**
* 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) {
for (BooleanClause clause : (BooleanQuery) query) {
if (!clause.isProhibited()) {
list.addAll(Arrays.asList(extractAutomata(clause.getQuery(), field)));
}
}
} else if (query instanceof ConstantScoreQuery) {
list.addAll(Arrays.asList(extractAutomata(((ConstantScoreQuery) query).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(
Operations.concatenate(
Automata.makeString(prefix.text()), Automata.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());
String prefix = UnicodeUtil.newString(termText, 0, prefixLength);
Automaton automaton = builder.toAutomaton(fq.getMaxEdits(), prefix);
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 it's very simple
list.add(
new CharacterRunAutomaton(Automata.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()]);
}