public TermInfo collect(String term) throws IOException {
TermInfo info = new TermInfo();
BytesRef luceneTerm = new BytesRef(term.getBytes());
// this gives documents in which the term is found, but no offset information can be retrieved
PostingsEnum postings =
MultiFields.getTermDocsEnum(indexReader, ngramInfoFieldname, luceneTerm);
// now go through each document
int docId = postings.nextDoc();
while (docId != PostingsEnum.NO_MORE_DOCS) {
// get the term vector for that document.
TermsEnum it = indexReader.getTermVector(docId, ngramInfoFieldname).iterator();
// find the term of interest
it.seekExact(luceneTerm);
// get its posting info. this will contain offset info
PostingsEnum postingsInDoc = it.postings(null, PostingsEnum.OFFSETS);
postingsInDoc.nextDoc();
Document doc = indexReader.document(docId);
String id = doc.get(idFieldname);
JATEDocument jd = new JATEDocument(id);
Set<int[]> offsets = new HashSet<>();
int totalFreq = postingsInDoc.freq();
for (int i = 0; i < totalFreq; i++) {
postingsInDoc.nextPosition();
offsets.add(new int[] {postingsInDoc.startOffset(), postingsInDoc.endOffset()});
}
info.getOffsets().put(jd, offsets);
docId = postings.nextDoc();
}
return info;
}
private FixedBitSet createExpectedResult(
String queryValue, boolean from, IndexReader topLevelReader, IndexIterationContext context)
throws IOException {
final Map<String, List<RandomDoc>> randomValueDocs;
final Map<String, List<RandomDoc>> linkValueDocuments;
if (from) {
randomValueDocs = context.randomValueFromDocs;
linkValueDocuments = context.toDocuments;
} else {
randomValueDocs = context.randomValueToDocs;
linkValueDocuments = context.fromDocuments;
}
FixedBitSet expectedResult = new FixedBitSet(topLevelReader.maxDoc());
List<RandomDoc> matchingDocs = randomValueDocs.get(queryValue);
if (matchingDocs == null) {
return new FixedBitSet(topLevelReader.maxDoc());
}
for (RandomDoc matchingDoc : matchingDocs) {
for (String linkValue : matchingDoc.linkValues) {
List<RandomDoc> otherMatchingDocs = linkValueDocuments.get(linkValue);
if (otherMatchingDocs == null) {
continue;
}
for (RandomDoc otherSideDoc : otherMatchingDocs) {
DocsEnum docsEnum =
MultiFields.getTermDocsEnum(
topLevelReader,
MultiFields.getLiveDocs(topLevelReader),
"id",
new BytesRef(otherSideDoc.id),
0);
assert docsEnum != null;
int doc = docsEnum.nextDoc();
expectedResult.set(doc);
}
}
}
return expectedResult;
}
