/**
* Returns a function that filters a document map based on the given include and exclude rules.
*
* @see #filter(Map, String[], String[]) for details
*/
public static Function<Map<String, ?>, Map<String, Object>> filter(
String[] includes, String[] excludes) {
CharacterRunAutomaton matchAllAutomaton = new CharacterRunAutomaton(Automata.makeAnyString());
CharacterRunAutomaton include;
if (includes == null || includes.length == 0) {
include = matchAllAutomaton;
} else {
Automaton includeA = Regex.simpleMatchToAutomaton(includes);
includeA = makeMatchDotsInFieldNames(includeA);
include = new CharacterRunAutomaton(includeA);
}
Automaton excludeA;
if (excludes == null || excludes.length == 0) {
excludeA = Automata.makeEmpty();
} else {
excludeA = Regex.simpleMatchToAutomaton(excludes);
excludeA = makeMatchDotsInFieldNames(excludeA);
}
CharacterRunAutomaton exclude = new CharacterRunAutomaton(excludeA);
// NOTE: We cannot use Operations.minus because of the special case that
// we want all sub properties to match as soon as an object matches
return (map) -> filter(map, include, 0, exclude, 0, matchAllAutomaton);
}
/**
* Build the {@link CharacterRunAutomaton} that represents the reindex-from-remote whitelist and
* make sure that it doesn't whitelist the world.
*/
static CharacterRunAutomaton buildRemoteWhitelist(List<String> whitelist) {
if (whitelist.isEmpty()) {
return new CharacterRunAutomaton(Automata.makeEmpty());
}
Automaton automaton = Regex.simpleMatchToAutomaton(whitelist.toArray(Strings.EMPTY_ARRAY));
automaton =
MinimizationOperations.minimize(automaton, Operations.DEFAULT_MAX_DETERMINIZED_STATES);
if (Operations.isTotal(automaton)) {
throw new IllegalArgumentException(
"Refusing to start because whitelist "
+ whitelist
+ " accepts all addresses. "
+ "This would allow users to reindex-from-remote any URL they like effectively having Elasticsearch make HTTP GETs "
+ "for them.");
}
return new CharacterRunAutomaton(automaton);
}
/**
* 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()]);
}
