public void listTokens(int freq) throws IOException {
IndexReader ireader = null;
TermsEnum iter = null;
Terms terms = null;
try {
ireader = DirectoryReader.open(indexDirectory);
int numDocs = ireader.numDocs();
if (numDocs > 0) {
Fields uFields = MultiFields.getFields(ireader); // reader.getTermVectors(0);
terms = uFields.terms(QueryBuilder.DEFS);
}
iter = terms.iterator(null); // init uid iterator
while (iter.term() != null) {
// if (iter.term().field().startsWith("f")) {
if (iter.docFreq() > 16 && iter.term().utf8ToString().length() > freq) {
log.warning(iter.term().utf8ToString());
}
iter.next();
/*} else {
break;
}*/
}
} finally {
if (ireader != null) {
try {
ireader.close();
} catch (IOException e) {
log.log(Level.WARNING, "An error occured while closing index reader", e);
}
}
}
}
private void printSegment(PrintWriter out, SegmentCommitInfo si) throws Exception {
SegmentReader reader =
new SegmentReader(si, DirectoryReader.DEFAULT_TERMS_INDEX_DIVISOR, newIOContext(random()));
for (int i = 0; i < reader.numDocs(); i++) out.println(reader.document(i));
Fields fields = reader.fields();
for (String field : fields) {
Terms terms = fields.terms(field);
assertNotNull(terms);
TermsEnum tis = terms.iterator(null);
while (tis.next() != null) {
out.print(" term=" + field + ":" + tis.term());
out.println(" DF=" + tis.docFreq());
DocsAndPositionsEnum positions = tis.docsAndPositions(reader.getLiveDocs(), null);
while (positions.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
out.print(" doc=" + positions.docID());
out.print(" TF=" + positions.freq());
out.print(" pos=");
out.print(positions.nextPosition());
for (int j = 1; j < positions.freq(); j++) out.print("," + positions.nextPosition());
out.println("");
}
}
}
reader.close();
}
/*
* listTermDictionary displays the term dictionary for a field.
*/
static void listTermDictionary(IndexReader reader, String fieldName) throws IOException {
System.out.println("\nTerm Dictionary: field " + fieldName);
/*
Grant says:
MultiFields.getTerms(IndexReader, fieldName)
*/
Terms terms = MultiFields.getTerms(reader, fieldName);
if ((terms == null) || (terms.size() == -1))
System.out.println(" The term dictionary is empty.");
else {
System.out.println(" Vocabulary size: " + terms.size() + " terms");
TermsEnum ithTerm = terms.iterator(null);
/*
* Iterate over the terms in this document.
* Information about a term's occurrences (tf and
* positions) is accessed via the indexing API, which
* returns inverted lists that describe (only) the
* current document.
*/
while (ithTerm.next() != null) {
System.out.format(
" %-30s %d %d\n",
ithTerm.term().utf8ToString(), ithTerm.docFreq(), ithTerm.totalTermFreq());
}
;
}
;
}
public static void fillQueue(TermsEnum termsEnum, TermStatsQueue tiq, String field)
throws Exception {
BytesRef term;
while ((term = termsEnum.next()) != null) {
BytesRef r = new BytesRef();
r.copyBytes(term);
tiq.insertWithOverflow(new TermStats(field, r, termsEnum.docFreq()));
}
}
@Override
public final int docFreq(Term term) throws IOException {
final Terms terms = terms(term.field());
if (terms == null) {
return 0;
}
final TermsEnum termsEnum = terms.iterator();
if (termsEnum.seekExact(term.bytes())) {
return termsEnum.docFreq();
} else {
return 0;
}
}
@Override
public void process(ResponseBuilder rb) throws IOException {
SolrParams params = rb.req.getParams();
if (!params.getBool(TermsParams.TERMS, false)) return;
String[] fields = params.getParams(TermsParams.TERMS_FIELD);
NamedList<Object> termsResult = new SimpleOrderedMap<>();
rb.rsp.add("terms", termsResult);
if (fields == null || fields.length == 0) return;
int limit = params.getInt(TermsParams.TERMS_LIMIT, 10);
if (limit < 0) {
limit = Integer.MAX_VALUE;
}
String lowerStr = params.get(TermsParams.TERMS_LOWER);
String upperStr = params.get(TermsParams.TERMS_UPPER);
boolean upperIncl = params.getBool(TermsParams.TERMS_UPPER_INCLUSIVE, false);
boolean lowerIncl = params.getBool(TermsParams.TERMS_LOWER_INCLUSIVE, true);
boolean sort =
!TermsParams.TERMS_SORT_INDEX.equals(
params.get(TermsParams.TERMS_SORT, TermsParams.TERMS_SORT_COUNT));
int freqmin = params.getInt(TermsParams.TERMS_MINCOUNT, 1);
int freqmax = params.getInt(TermsParams.TERMS_MAXCOUNT, UNLIMITED_MAX_COUNT);
if (freqmax < 0) {
freqmax = Integer.MAX_VALUE;
}
String prefix = params.get(TermsParams.TERMS_PREFIX_STR);
String regexp = params.get(TermsParams.TERMS_REGEXP_STR);
Pattern pattern = regexp != null ? Pattern.compile(regexp, resolveRegexpFlags(params)) : null;
boolean raw = params.getBool(TermsParams.TERMS_RAW, false);
final AtomicReader indexReader = rb.req.getSearcher().getAtomicReader();
Fields lfields = indexReader.fields();
for (String field : fields) {
NamedList<Integer> fieldTerms = new NamedList<>();
termsResult.add(field, fieldTerms);
Terms terms = lfields == null ? null : lfields.terms(field);
if (terms == null) {
// no terms for this field
continue;
}
FieldType ft = raw ? null : rb.req.getSchema().getFieldTypeNoEx(field);
if (ft == null) ft = new StrField();
// prefix must currently be text
BytesRef prefixBytes = prefix == null ? null : new BytesRef(prefix);
BytesRef upperBytes = null;
if (upperStr != null) {
upperBytes = new BytesRef();
ft.readableToIndexed(upperStr, upperBytes);
}
BytesRef lowerBytes;
if (lowerStr == null) {
// If no lower bound was specified, use the prefix
lowerBytes = prefixBytes;
} else {
lowerBytes = new BytesRef();
if (raw) {
// TODO: how to handle binary? perhaps we don't for "raw"... or if the field exists
// perhaps we detect if the FieldType is non-character and expect hex if so?
lowerBytes = new BytesRef(lowerStr);
} else {
lowerBytes = new BytesRef();
ft.readableToIndexed(lowerStr, lowerBytes);
}
}
TermsEnum termsEnum = terms.iterator(null);
BytesRef term = null;
if (lowerBytes != null) {
if (termsEnum.seekCeil(lowerBytes) == TermsEnum.SeekStatus.END) {
termsEnum = null;
} else {
term = termsEnum.term();
// Only advance the enum if we are excluding the lower bound and the lower Term actually
// matches
if (lowerIncl == false && term.equals(lowerBytes)) {
term = termsEnum.next();
}
}
} else {
// position termsEnum on first term
term = termsEnum.next();
}
int i = 0;
BoundedTreeSet<CountPair<BytesRef, Integer>> queue =
(sort ? new BoundedTreeSet<CountPair<BytesRef, Integer>>(limit) : null);
CharsRef external = new CharsRef();
while (term != null && (i < limit || sort)) {
boolean externalized = false; // did we fill in "external" yet for this term?
// stop if the prefix doesn't match
if (prefixBytes != null && !StringHelper.startsWith(term, prefixBytes)) break;
if (pattern != null) {
// indexed text or external text?
// TODO: support "raw" mode?
ft.indexedToReadable(term, external);
externalized = true;
if (!pattern.matcher(external).matches()) {
term = termsEnum.next();
continue;
}
}
if (upperBytes != null) {
int upperCmp = term.compareTo(upperBytes);
// if we are past the upper term, or equal to it (when don't include upper) then stop.
if (upperCmp > 0 || (upperCmp == 0 && !upperIncl)) break;
}
// This is a good term in the range. Check if mincount/maxcount conditions are satisfied.
int docFreq = termsEnum.docFreq();
if (docFreq >= freqmin && docFreq <= freqmax) {
// add the term to the list
if (sort) {
queue.add(new CountPair<>(BytesRef.deepCopyOf(term), docFreq));
} else {
// TODO: handle raw somehow
if (!externalized) {
ft.indexedToReadable(term, external);
}
fieldTerms.add(external.toString(), docFreq);
i++;
}
}
term = termsEnum.next();
}
if (sort) {
for (CountPair<BytesRef, Integer> item : queue) {
if (i >= limit) break;
ft.indexedToReadable(item.key, external);
fieldTerms.add(external.toString(), item.val);
i++;
}
}
}
}
public void verifyEquals(DirectoryReader r1, DirectoryReader r2, String idField)
throws Throwable {
if (VERBOSE) {
System.out.println("\nr1 docs:");
printDocs(r1);
System.out.println("\nr2 docs:");
printDocs(r2);
}
if (r1.numDocs() != r2.numDocs()) {
assert false : "r1.numDocs()=" + r1.numDocs() + " vs r2.numDocs()=" + r2.numDocs();
}
boolean hasDeletes = !(r1.maxDoc() == r2.maxDoc() && r1.numDocs() == r1.maxDoc());
int[] r2r1 = new int[r2.maxDoc()]; // r2 id to r1 id mapping
// create mapping from id2 space to id2 based on idField
final Fields f1 = MultiFields.getFields(r1);
if (f1 == null) {
// make sure r2 is empty
assertNull(MultiFields.getFields(r2));
return;
}
final Terms terms1 = f1.terms(idField);
if (terms1 == null) {
assertTrue(
MultiFields.getFields(r2) == null || MultiFields.getFields(r2).terms(idField) == null);
return;
}
final TermsEnum termsEnum = terms1.iterator(null);
final Bits liveDocs1 = MultiFields.getLiveDocs(r1);
final Bits liveDocs2 = MultiFields.getLiveDocs(r2);
Fields fields = MultiFields.getFields(r2);
if (fields == null) {
// make sure r1 is in fact empty (eg has only all
// deleted docs):
Bits liveDocs = MultiFields.getLiveDocs(r1);
DocsEnum docs = null;
while (termsEnum.next() != null) {
docs = TestUtil.docs(random(), termsEnum, liveDocs, docs, DocsEnum.FLAG_NONE);
while (docs.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
fail("r1 is not empty but r2 is");
}
}
return;
}
Terms terms2 = fields.terms(idField);
TermsEnum termsEnum2 = terms2.iterator(null);
DocsEnum termDocs1 = null;
DocsEnum termDocs2 = null;
while (true) {
BytesRef term = termsEnum.next();
// System.out.println("TEST: match id term=" + term);
if (term == null) {
break;
}
termDocs1 = TestUtil.docs(random(), termsEnum, liveDocs1, termDocs1, DocsEnum.FLAG_NONE);
if (termsEnum2.seekExact(term)) {
termDocs2 = TestUtil.docs(random(), termsEnum2, liveDocs2, termDocs2, DocsEnum.FLAG_NONE);
} else {
termDocs2 = null;
}
if (termDocs1.nextDoc() == DocIdSetIterator.NO_MORE_DOCS) {
// This doc is deleted and wasn't replaced
assertTrue(termDocs2 == null || termDocs2.nextDoc() == DocIdSetIterator.NO_MORE_DOCS);
continue;
}
int id1 = termDocs1.docID();
assertEquals(DocIdSetIterator.NO_MORE_DOCS, termDocs1.nextDoc());
assertTrue(termDocs2.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int id2 = termDocs2.docID();
assertEquals(DocIdSetIterator.NO_MORE_DOCS, termDocs2.nextDoc());
r2r1[id2] = id1;
// verify stored fields are equivalent
try {
verifyEquals(r1.document(id1), r2.document(id2));
} catch (Throwable t) {
System.out.println("FAILED id=" + term + " id1=" + id1 + " id2=" + id2 + " term=" + term);
System.out.println(" d1=" + r1.document(id1));
System.out.println(" d2=" + r2.document(id2));
throw t;
}
try {
// verify term vectors are equivalent
verifyEquals(r1.getTermVectors(id1), r2.getTermVectors(id2));
} catch (Throwable e) {
System.out.println("FAILED id=" + term + " id1=" + id1 + " id2=" + id2);
Fields tv1 = r1.getTermVectors(id1);
System.out.println(" d1=" + tv1);
if (tv1 != null) {
DocsAndPositionsEnum dpEnum = null;
DocsEnum dEnum = null;
for (String field : tv1) {
System.out.println(" " + field + ":");
Terms terms3 = tv1.terms(field);
assertNotNull(terms3);
TermsEnum termsEnum3 = terms3.iterator(null);
BytesRef term2;
while ((term2 = termsEnum3.next()) != null) {
System.out.println(
" " + term2.utf8ToString() + ": freq=" + termsEnum3.totalTermFreq());
dpEnum = termsEnum3.docsAndPositions(null, dpEnum);
if (dpEnum != null) {
assertTrue(dpEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
final int freq = dpEnum.freq();
System.out.println(" doc=" + dpEnum.docID() + " freq=" + freq);
for (int posUpto = 0; posUpto < freq; posUpto++) {
System.out.println(" pos=" + dpEnum.nextPosition());
}
} else {
dEnum = TestUtil.docs(random(), termsEnum3, null, dEnum, DocsEnum.FLAG_FREQS);
assertNotNull(dEnum);
assertTrue(dEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
final int freq = dEnum.freq();
System.out.println(" doc=" + dEnum.docID() + " freq=" + freq);
}
}
}
}
Fields tv2 = r2.getTermVectors(id2);
System.out.println(" d2=" + tv2);
if (tv2 != null) {
DocsAndPositionsEnum dpEnum = null;
DocsEnum dEnum = null;
for (String field : tv2) {
System.out.println(" " + field + ":");
Terms terms3 = tv2.terms(field);
assertNotNull(terms3);
TermsEnum termsEnum3 = terms3.iterator(null);
BytesRef term2;
while ((term2 = termsEnum3.next()) != null) {
System.out.println(
" " + term2.utf8ToString() + ": freq=" + termsEnum3.totalTermFreq());
dpEnum = termsEnum3.docsAndPositions(null, dpEnum);
if (dpEnum != null) {
assertTrue(dpEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
final int freq = dpEnum.freq();
System.out.println(" doc=" + dpEnum.docID() + " freq=" + freq);
for (int posUpto = 0; posUpto < freq; posUpto++) {
System.out.println(" pos=" + dpEnum.nextPosition());
}
} else {
dEnum = TestUtil.docs(random(), termsEnum3, null, dEnum, DocsEnum.FLAG_FREQS);
assertNotNull(dEnum);
assertTrue(dEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
final int freq = dEnum.freq();
System.out.println(" doc=" + dEnum.docID() + " freq=" + freq);
}
}
}
}
throw e;
}
}
// System.out.println("TEST: done match id");
// Verify postings
// System.out.println("TEST: create te1");
final Fields fields1 = MultiFields.getFields(r1);
final Iterator<String> fields1Enum = fields1.iterator();
final Fields fields2 = MultiFields.getFields(r2);
final Iterator<String> fields2Enum = fields2.iterator();
String field1 = null, field2 = null;
TermsEnum termsEnum1 = null;
termsEnum2 = null;
DocsEnum docs1 = null, docs2 = null;
// pack both doc and freq into single element for easy sorting
long[] info1 = new long[r1.numDocs()];
long[] info2 = new long[r2.numDocs()];
for (; ; ) {
BytesRef term1 = null, term2 = null;
// iterate until we get some docs
int len1;
for (; ; ) {
len1 = 0;
if (termsEnum1 == null) {
if (!fields1Enum.hasNext()) {
break;
}
field1 = fields1Enum.next();
Terms terms = fields1.terms(field1);
if (terms == null) {
continue;
}
termsEnum1 = terms.iterator(null);
}
term1 = termsEnum1.next();
if (term1 == null) {
// no more terms in this field
termsEnum1 = null;
continue;
}
// System.out.println("TEST: term1=" + term1);
docs1 = TestUtil.docs(random(), termsEnum1, liveDocs1, docs1, DocsEnum.FLAG_FREQS);
while (docs1.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
int d = docs1.docID();
int f = docs1.freq();
info1[len1] = (((long) d) << 32) | f;
len1++;
}
if (len1 > 0) break;
}
// iterate until we get some docs
int len2;
for (; ; ) {
len2 = 0;
if (termsEnum2 == null) {
if (!fields2Enum.hasNext()) {
break;
}
field2 = fields2Enum.next();
Terms terms = fields2.terms(field2);
if (terms == null) {
continue;
}
termsEnum2 = terms.iterator(null);
}
term2 = termsEnum2.next();
if (term2 == null) {
// no more terms in this field
termsEnum2 = null;
continue;
}
// System.out.println("TEST: term1=" + term1);
docs2 = TestUtil.docs(random(), termsEnum2, liveDocs2, docs2, DocsEnum.FLAG_FREQS);
while (docs2.nextDoc() != DocIdSetIterator.NO_MORE_DOCS) {
int d = r2r1[docs2.docID()];
int f = docs2.freq();
info2[len2] = (((long) d) << 32) | f;
len2++;
}
if (len2 > 0) break;
}
assertEquals(len1, len2);
if (len1 == 0) break; // no more terms
assertEquals(field1, field2);
assertTrue(term1.bytesEquals(term2));
if (!hasDeletes) assertEquals(termsEnum1.docFreq(), termsEnum2.docFreq());
assertEquals("len1=" + len1 + " len2=" + len2 + " deletes?=" + hasDeletes, term1, term2);
// sort info2 to get it into ascending docid
Arrays.sort(info2, 0, len2);
// now compare
for (int i = 0; i < len1; i++) {
assertEquals(
"i="
+ i
+ " len="
+ len1
+ " d1="
+ (info1[i] >>> 32)
+ " f1="
+ (info1[i] & Integer.MAX_VALUE)
+ " d2="
+ (info2[i] >>> 32)
+ " f2="
+ (info2[i] & Integer.MAX_VALUE)
+ " field="
+ field1
+ " term="
+ term1.utf8ToString(),
info1[i],
info2[i]);
}
}
}
