static Query parseQueryString(
ExtendedCommonTermsQuery query,
Object queryString,
String field,
Analyzer analyzer,
String lowFreqMinimumShouldMatch,
String highFreqMinimumShouldMatch)
throws IOException {
// Logic similar to QueryParser#getFieldQuery
int count = 0;
try (TokenStream source = analyzer.tokenStream(field, queryString.toString())) {
source.reset();
CharTermAttribute termAtt = source.addAttribute(CharTermAttribute.class);
BytesRefBuilder builder = new BytesRefBuilder();
while (source.incrementToken()) {
// UTF-8
builder.copyChars(termAtt);
query.add(new Term(field, builder.toBytesRef()));
count++;
}
}
if (count == 0) {
return null;
}
query.setLowFreqMinimumNumberShouldMatch(lowFreqMinimumShouldMatch);
query.setHighFreqMinimumNumberShouldMatch(highFreqMinimumShouldMatch);
return query;
}
/** Unmarshals a string-based field value. */
protected static Object unmarshalStringSortValue(Object value) {
if (null == value) {
return null;
}
BytesRefBuilder spare = new BytesRefBuilder();
String stringVal = (String) value;
spare.copyChars(stringVal);
return spare.get();
}
// NOTE: while it's tempting to make this public, since
// caller's parser likely knows the
// numInput/numOutputWords, sneaky exceptions, much later
// on, will result if these values are wrong; so we always
// recompute ourselves to be safe:
private void add(
CharsRef input,
int numInputWords,
CharsRef output,
int numOutputWords,
boolean includeOrig) {
// first convert to UTF-8
if (numInputWords <= 0) {
throw new IllegalArgumentException("numInputWords must be > 0 (got " + numInputWords + ")");
}
if (input.length <= 0) {
throw new IllegalArgumentException("input.length must be > 0 (got " + input.length + ")");
}
if (numOutputWords <= 0) {
throw new IllegalArgumentException(
"numOutputWords must be > 0 (got " + numOutputWords + ")");
}
if (output.length <= 0) {
throw new IllegalArgumentException("output.length must be > 0 (got " + output.length + ")");
}
assert !hasHoles(input) : "input has holes: " + input;
assert !hasHoles(output) : "output has holes: " + output;
// System.out.println("fmap.add input=" + input + " numInputWords=" + numInputWords + "
// output=" + output + " numOutputWords=" + numOutputWords);
utf8Scratch.copyChars(output.chars, output.offset, output.length);
// lookup in hash
int ord = words.add(utf8Scratch.get());
if (ord < 0) {
// already exists in our hash
ord = (-ord) - 1;
// System.out.println(" output=" + output + " old ord=" + ord);
} else {
// System.out.println(" output=" + output + " new ord=" + ord);
}
MapEntry e = workingSet.get(input);
if (e == null) {
e = new MapEntry();
workingSet.put(
CharsRef.deepCopyOf(input), e); // make a copy, since we will keep around in our map
}
e.ords.add(ord);
e.includeOrig |= includeOrig;
maxHorizontalContext = Math.max(maxHorizontalContext, numInputWords);
maxHorizontalContext = Math.max(maxHorizontalContext, numOutputWords);
}
/** Given the readable value, return the term value that will match it. */
public void readableToIndexed(CharSequence val, BytesRefBuilder result) {
final String internal = readableToIndexed(val.toString());
result.copyChars(internal);
}
public static NamedList<Integer> getCounts(
SolrIndexSearcher searcher,
DocSet docs,
String fieldName,
int offset,
int limit,
int mincount,
boolean missing,
String sort,
String prefix,
String contains,
boolean ignoreCase)
throws IOException {
SchemaField schemaField = searcher.getSchema().getField(fieldName);
FieldType ft = schemaField.getType();
NamedList<Integer> res = new NamedList<>();
// TODO: remove multiValuedFieldCache(), check dv type / uninversion type?
final boolean multiValued = schemaField.multiValued() || ft.multiValuedFieldCache();
final SortedSetDocValues si; // for term lookups only
OrdinalMap ordinalMap = null; // for mapping per-segment ords to global ones
if (multiValued) {
si = searcher.getLeafReader().getSortedSetDocValues(fieldName);
if (si instanceof MultiSortedSetDocValues) {
ordinalMap = ((MultiSortedSetDocValues) si).mapping;
}
} else {
SortedDocValues single = searcher.getLeafReader().getSortedDocValues(fieldName);
si = single == null ? null : DocValues.singleton(single);
if (single instanceof MultiSortedDocValues) {
ordinalMap = ((MultiSortedDocValues) single).mapping;
}
}
if (si == null) {
return finalize(res, searcher, schemaField, docs, -1, missing);
}
if (si.getValueCount() >= Integer.MAX_VALUE) {
throw new UnsupportedOperationException(
"Currently this faceting method is limited to " + Integer.MAX_VALUE + " unique terms");
}
final BytesRefBuilder prefixRef;
if (prefix == null) {
prefixRef = null;
} else if (prefix.length() == 0) {
prefix = null;
prefixRef = null;
} else {
prefixRef = new BytesRefBuilder();
prefixRef.copyChars(prefix);
}
int startTermIndex, endTermIndex;
if (prefix != null) {
startTermIndex = (int) si.lookupTerm(prefixRef.get());
if (startTermIndex < 0) startTermIndex = -startTermIndex - 1;
prefixRef.append(UnicodeUtil.BIG_TERM);
endTermIndex = (int) si.lookupTerm(prefixRef.get());
assert endTermIndex < 0;
endTermIndex = -endTermIndex - 1;
} else {
startTermIndex = -1;
endTermIndex = (int) si.getValueCount();
}
final int nTerms = endTermIndex - startTermIndex;
int missingCount = -1;
final CharsRefBuilder charsRef = new CharsRefBuilder();
if (nTerms > 0 && docs.size() >= mincount) {
// count collection array only needs to be as big as the number of terms we are
// going to collect counts for.
final int[] counts = new int[nTerms];
Filter filter = docs.getTopFilter();
List<LeafReaderContext> leaves = searcher.getTopReaderContext().leaves();
for (int subIndex = 0; subIndex < leaves.size(); subIndex++) {
LeafReaderContext leaf = leaves.get(subIndex);
DocIdSet dis =
filter.getDocIdSet(leaf, null); // solr docsets already exclude any deleted docs
DocIdSetIterator disi = null;
if (dis != null) {
disi = dis.iterator();
}
if (disi != null) {
if (multiValued) {
SortedSetDocValues sub = leaf.reader().getSortedSetDocValues(fieldName);
if (sub == null) {
sub = DocValues.emptySortedSet();
}
final SortedDocValues singleton = DocValues.unwrapSingleton(sub);
if (singleton != null) {
// some codecs may optimize SORTED_SET storage for single-valued fields
accumSingle(counts, startTermIndex, singleton, disi, subIndex, ordinalMap);
} else {
accumMulti(counts, startTermIndex, sub, disi, subIndex, ordinalMap);
}
} else {
SortedDocValues sub = leaf.reader().getSortedDocValues(fieldName);
if (sub == null) {
sub = DocValues.emptySorted();
}
accumSingle(counts, startTermIndex, sub, disi, subIndex, ordinalMap);
}
}
}
if (startTermIndex == -1) {
missingCount = counts[0];
}
// IDEA: we could also maintain a count of "other"... everything that fell outside
// of the top 'N'
int off = offset;
int lim = limit >= 0 ? limit : Integer.MAX_VALUE;
if (sort.equals(FacetParams.FACET_SORT_COUNT)
|| sort.equals(FacetParams.FACET_SORT_COUNT_LEGACY)) {
int maxsize = limit > 0 ? offset + limit : Integer.MAX_VALUE - 1;
maxsize = Math.min(maxsize, nTerms);
LongPriorityQueue queue =
new LongPriorityQueue(Math.min(maxsize, 1000), maxsize, Long.MIN_VALUE);
int min = mincount - 1; // the smallest value in the top 'N' values
for (int i = (startTermIndex == -1) ? 1 : 0; i < nTerms; i++) {
int c = counts[i];
if (contains != null) {
final BytesRef term = si.lookupOrd(startTermIndex + i);
if (!SimpleFacets.contains(term.utf8ToString(), contains, ignoreCase)) {
continue;
}
}
if (c > min) {
// NOTE: we use c>min rather than c>=min as an optimization because we are going in
// index order, so we already know that the keys are ordered. This can be very
// important if a lot of the counts are repeated (like zero counts would be).
// smaller term numbers sort higher, so subtract the term number instead
long pair = (((long) c) << 32) + (Integer.MAX_VALUE - i);
boolean displaced = queue.insert(pair);
if (displaced) min = (int) (queue.top() >>> 32);
}
}
// if we are deep paging, we don't have to order the highest "offset" counts.
int collectCount = Math.max(0, queue.size() - off);
assert collectCount <= lim;
// the start and end indexes of our list "sorted" (starting with the highest value)
int sortedIdxStart = queue.size() - (collectCount - 1);
int sortedIdxEnd = queue.size() + 1;
final long[] sorted = queue.sort(collectCount);
for (int i = sortedIdxStart; i < sortedIdxEnd; i++) {
long pair = sorted[i];
int c = (int) (pair >>> 32);
int tnum = Integer.MAX_VALUE - (int) pair;
final BytesRef term = si.lookupOrd(startTermIndex + tnum);
ft.indexedToReadable(term, charsRef);
res.add(charsRef.toString(), c);
}
} else {
// add results in index order
int i = (startTermIndex == -1) ? 1 : 0;
if (mincount <= 0 && contains == null) {
// if mincount<=0 and we're not examining the values for contains, then
// we won't discard any terms and we know exactly where to start.
i += off;
off = 0;
}
for (; i < nTerms; i++) {
int c = counts[i];
if (c < mincount) continue;
BytesRef term = null;
if (contains != null) {
term = si.lookupOrd(startTermIndex + i);
if (!SimpleFacets.contains(term.utf8ToString(), contains, ignoreCase)) {
continue;
}
}
if (--off >= 0) continue;
if (--lim < 0) break;
if (term == null) {
term = si.lookupOrd(startTermIndex + i);
}
ft.indexedToReadable(term, charsRef);
res.add(charsRef.toString(), c);
}
}
}
return finalize(res, searcher, schemaField, docs, missingCount, missing);
}