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();
}
/*
* Utility function to display a term vector.
*/
static void termVectorDisplay(Terms terms) throws IOException {
if ((terms == null) || (terms.size() == -1)) System.out.println(" The field is not stored.");
else {
/*
* The terms for this field are stored.
*/
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(
" %10d %-20s %d ",
ithTerm.ord(), ithTerm.term().utf8ToString(), ithTerm.totalTermFreq());
DocsAndPositionsEnum currDoc = ithTerm.docsAndPositions(null, null);
currDoc.nextDoc();
for (int jthPosition = 0; jthPosition < ithTerm.totalTermFreq(); jthPosition++)
System.out.print(currDoc.nextPosition() + " ");
System.out.println();
}
;
}
;
}
/**
* Returns {@link DocsAndPositionsEnum} for the specified term. This will return null if the field
* or term does not exist or positions weren't indexed.
*
* @deprecated use {@link #postings(Term, int)} instead
*/
@Deprecated
public final DocsAndPositionsEnum termPositionsEnum(Term term) throws IOException {
assert term.field() != null;
assert term.bytes() != null;
final Terms terms = terms(term.field());
if (terms != null) {
final TermsEnum termsEnum = terms.iterator();
if (termsEnum.seekExact(term.bytes())) {
return termsEnum.docsAndPositions(getLiveDocs(), null);
}
}
return null;
}
/**
* Reconstruct document fields.
*
* @param docNum document number. If this document is deleted, but the index is not optimized yet,
* the reconstruction process may still yield the reconstructed field content even from
* deleted documents.
* @return reconstructed document
* @throws Exception
*/
public Reconstructed reconstruct(int docNum) throws Exception {
if (docNum < 0 || docNum > reader.maxDoc()) {
throw new Exception("Document number outside of valid range.");
}
Reconstructed res = new Reconstructed();
if (deleted != null && deleted.get(docNum)) {
throw new Exception("Document is deleted.");
} else {
Document doc = reader.document(docNum);
for (int i = 0; i < fieldNames.length; i++) {
Field[] fs = doc.getFields(fieldNames[i]);
if (fs != null && fs.length > 0) {
res.getStoredFields().put(fieldNames[i], fs);
}
}
}
// collect values from unstored fields
HashSet<String> fields = new HashSet<String>(Arrays.asList(fieldNames));
// try to use term vectors if available
progress.maxValue = fieldNames.length;
progress.curValue = 0;
progress.minValue = 0;
for (int i = 0; i < fieldNames.length; i++) {
TermFreqVector tvf = reader.getTermFreqVector(docNum, fieldNames[i]);
if (tvf != null && tvf.size() > 0 && (tvf instanceof TermPositionVector)) {
TermPositionVector tpv = (TermPositionVector) tvf;
progress.message = "Reading term vectors ...";
progress.curValue = i;
setChanged();
notifyObservers(progress);
BytesRef[] tv = tpv.getTerms();
for (int k = 0; k < tv.length; k++) {
// do we have positions?
int[] posArr = tpv.getTermPositions(k);
if (posArr == null) {
// only offsets
TermVectorOffsetInfo[] offsets = tpv.getOffsets(k);
if (offsets.length == 0) {
continue;
}
// convert offsets into positions
posArr = convertOffsets(offsets);
}
GrowableStringArray gsa = res.getReconstructedFields().get(fieldNames[i]);
if (gsa == null) {
gsa = new GrowableStringArray();
res.getReconstructedFields().put(fieldNames[i], gsa);
}
for (int m = 0; m < posArr.length; m++) {
gsa.append(posArr[m], "|", tv[k].utf8ToString());
}
}
fields.remove(fieldNames[i]); // got what we wanted
}
}
// this loop collects data only from left-over fields
// not yet collected through term vectors
progress.maxValue = fields.size();
progress.curValue = 0;
progress.minValue = 0;
for (String fld : fields) {
progress.message = "Collecting terms in " + fld + " ...";
progress.curValue++;
setChanged();
notifyObservers(progress);
Terms terms = MultiFields.getTerms(reader, fld);
if (terms == null) { // no terms in this field
continue;
}
TermsEnum te = terms.iterator();
while (te.next() != null) {
DocsAndPositionsEnum dpe = te.docsAndPositions(deleted, null);
if (dpe == null) { // no position info for this field
break;
}
int num = dpe.advance(docNum);
if (num != docNum) { // either greater than or NO_MORE_DOCS
continue; // no data for this term in this doc
}
String term = te.term().utf8ToString();
GrowableStringArray gsa = (GrowableStringArray) res.getReconstructedFields().get(fld);
if (gsa == null) {
gsa = new GrowableStringArray();
res.getReconstructedFields().put(fld, gsa);
}
for (int k = 0; k < dpe.freq(); k++) {
int pos = dpe.nextPosition();
gsa.append(pos, "|", term);
}
}
}
progress.message = "Done.";
progress.curValue = 100;
setChanged();
notifyObservers(progress);
return res;
}
protected void validateResponse(
TermVectorResponse esResponse, Fields luceneFields, TestConfig testConfig)
throws IOException {
TestDoc testDoc = testConfig.doc;
HashSet<String> selectedFields =
testConfig.selectedFields == null
? null
: new HashSet<String>(Arrays.asList(testConfig.selectedFields));
Fields esTermVectorFields = esResponse.getFields();
for (TestFieldSetting field : testDoc.fieldSettings) {
Terms esTerms = esTermVectorFields.terms(field.name);
if (selectedFields != null && !selectedFields.contains(field.name)) {
assertNull(esTerms);
continue;
}
assertNotNull(esTerms);
Terms luceneTerms = luceneFields.terms(field.name);
TermsEnum esTermEnum = esTerms.iterator(null);
TermsEnum luceneTermEnum = luceneTerms.iterator(null);
while (esTermEnum.next() != null) {
assertNotNull(luceneTermEnum.next());
assertThat(esTermEnum.totalTermFreq(), equalTo(luceneTermEnum.totalTermFreq()));
DocsAndPositionsEnum esDocsPosEnum = esTermEnum.docsAndPositions(null, null, 0);
DocsAndPositionsEnum luceneDocsPosEnum = luceneTermEnum.docsAndPositions(null, null, 0);
if (luceneDocsPosEnum == null) {
// test we expect that...
assertFalse(field.storedOffset);
assertFalse(field.storedPayloads);
assertFalse(field.storedPositions);
continue;
}
String currentTerm = esTermEnum.term().utf8ToString();
assertThat(
"Token mismatch for field: " + field.name,
currentTerm,
equalTo(luceneTermEnum.term().utf8ToString()));
esDocsPosEnum.nextDoc();
luceneDocsPosEnum.nextDoc();
int freq = esDocsPosEnum.freq();
assertThat(freq, equalTo(luceneDocsPosEnum.freq()));
for (int i = 0; i < freq; i++) {
String failDesc = " (field:" + field.name + " term:" + currentTerm + ")";
int lucenePos = luceneDocsPosEnum.nextPosition();
int esPos = esDocsPosEnum.nextPosition();
if (field.storedPositions && testConfig.requestPositions) {
assertThat("Position test failed" + failDesc, lucenePos, equalTo(esPos));
} else {
assertThat("Missing position test failed" + failDesc, esPos, equalTo(-1));
}
if (field.storedOffset && testConfig.requestOffsets) {
assertThat(
"Offset test failed" + failDesc,
luceneDocsPosEnum.startOffset(),
equalTo(esDocsPosEnum.startOffset()));
assertThat(
"Offset test failed" + failDesc,
luceneDocsPosEnum.endOffset(),
equalTo(esDocsPosEnum.endOffset()));
} else {
assertThat(
"Missing offset test failed" + failDesc, esDocsPosEnum.startOffset(), equalTo(-1));
assertThat(
"Missing offset test failed" + failDesc, esDocsPosEnum.endOffset(), equalTo(-1));
}
if (field.storedPayloads && testConfig.requestPayloads) {
assertThat(
"Payload test failed" + failDesc,
luceneDocsPosEnum.getPayload(),
equalTo(esDocsPosEnum.getPayload()));
} else {
assertThat(
"Missing payload test failed" + failDesc,
esDocsPosEnum.getPayload(),
equalTo(null));
}
}
}
assertNull("Es returned terms are done but lucene isn't", luceneTermEnum.next());
}
}
public static void verifyEquals(Fields d1, Fields d2) throws IOException {
if (d1 == null) {
assertTrue(d2 == null || d2.size() == 0);
return;
}
assertTrue(d2 != null);
Iterator<String> fieldsEnum2 = d2.iterator();
for (String field1 : d1) {
String field2 = fieldsEnum2.next();
assertEquals(field1, field2);
Terms terms1 = d1.terms(field1);
assertNotNull(terms1);
TermsEnum termsEnum1 = terms1.iterator(null);
Terms terms2 = d2.terms(field2);
assertNotNull(terms2);
TermsEnum termsEnum2 = terms2.iterator(null);
DocsAndPositionsEnum dpEnum1 = null;
DocsAndPositionsEnum dpEnum2 = null;
DocsEnum dEnum1 = null;
DocsEnum dEnum2 = null;
BytesRef term1;
while ((term1 = termsEnum1.next()) != null) {
BytesRef term2 = termsEnum2.next();
assertEquals(term1, term2);
assertEquals(termsEnum1.totalTermFreq(), termsEnum2.totalTermFreq());
dpEnum1 = termsEnum1.docsAndPositions(null, dpEnum1);
dpEnum2 = termsEnum2.docsAndPositions(null, dpEnum2);
if (dpEnum1 != null) {
assertNotNull(dpEnum2);
int docID1 = dpEnum1.nextDoc();
dpEnum2.nextDoc();
// docIDs are not supposed to be equal
// int docID2 = dpEnum2.nextDoc();
// assertEquals(docID1, docID2);
assertTrue(docID1 != DocIdSetIterator.NO_MORE_DOCS);
int freq1 = dpEnum1.freq();
int freq2 = dpEnum2.freq();
assertEquals(freq1, freq2);
OffsetAttribute offsetAtt1 =
dpEnum1.attributes().hasAttribute(OffsetAttribute.class)
? dpEnum1.attributes().getAttribute(OffsetAttribute.class)
: null;
OffsetAttribute offsetAtt2 =
dpEnum2.attributes().hasAttribute(OffsetAttribute.class)
? dpEnum2.attributes().getAttribute(OffsetAttribute.class)
: null;
if (offsetAtt1 != null) {
assertNotNull(offsetAtt2);
} else {
assertNull(offsetAtt2);
}
for (int posUpto = 0; posUpto < freq1; posUpto++) {
int pos1 = dpEnum1.nextPosition();
int pos2 = dpEnum2.nextPosition();
assertEquals(pos1, pos2);
if (offsetAtt1 != null) {
assertEquals(offsetAtt1.startOffset(), offsetAtt2.startOffset());
assertEquals(offsetAtt1.endOffset(), offsetAtt2.endOffset());
}
}
assertEquals(DocIdSetIterator.NO_MORE_DOCS, dpEnum1.nextDoc());
assertEquals(DocIdSetIterator.NO_MORE_DOCS, dpEnum2.nextDoc());
} else {
dEnum1 = TestUtil.docs(random(), termsEnum1, null, dEnum1, DocsEnum.FLAG_FREQS);
dEnum2 = TestUtil.docs(random(), termsEnum2, null, dEnum2, DocsEnum.FLAG_FREQS);
assertNotNull(dEnum1);
assertNotNull(dEnum2);
int docID1 = dEnum1.nextDoc();
dEnum2.nextDoc();
// docIDs are not supposed to be equal
// int docID2 = dEnum2.nextDoc();
// assertEquals(docID1, docID2);
assertTrue(docID1 != DocIdSetIterator.NO_MORE_DOCS);
int freq1 = dEnum1.freq();
int freq2 = dEnum2.freq();
assertEquals(freq1, freq2);
assertEquals(DocIdSetIterator.NO_MORE_DOCS, dEnum1.nextDoc());
assertEquals(DocIdSetIterator.NO_MORE_DOCS, dEnum2.nextDoc());
}
}
assertNull(termsEnum2.next());
}
assertFalse(fieldsEnum2.hasNext());
}
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]);
}
}
}
