@Override
public void index(Document document) {
if (log.isTraceEnabled()) {
log.trace(
"Indexing "
+ sql
+ " id="
+ JdbcIndexDefinition.this.identifier
+ ", key = "
+ JdbcIndexDefinition.this.key);
}
String id = getIdentifier();
if (id != null) {
document.add(
new Field(
"builder",
"VIRTUAL BUILDER",
Field.Store.YES,
Field.Index.NOT_ANALYZED)); // keyword
document.add(
new Field(
"number", getIdentifier(), Field.Store.YES, Field.Index.NOT_ANALYZED)); // keyword
}
try {
for (int i = 1; i <= meta.getColumnCount(); i++) {
String value = org.mmbase.util.Casting.toString(results.getString(i));
if (log.isTraceEnabled()) {
log.trace(
"Indexing " + value + " for " + meta.getColumnName(i) + " on " + getIdentifier());
}
String fieldName = meta.getColumnName(i);
if (keyWords.contains(fieldName)) {
Indexer.addField(
document,
new Field(fieldName, value, Field.Store.YES, Field.Index.NOT_ANALYZED),
nonDefaultMultiples.get(fieldName)); // keyword
} else {
Field field = new Field(fieldName, value, Field.Store.YES, Field.Index.ANALYZED);
Float boost = boosts.get(fieldName);
if (boost != null) {
field.setBoost(boost);
}
Indexer.addField(document, field, nonDefaultMultiples.get(fieldName));
Field fullText = new Field("fulltext", value, Field.Store.YES, Field.Index.ANALYZED);
if (boost != null) {
fullText.setBoost(boost);
}
document.add(fullText);
}
}
} catch (SQLException sqe) {
log.error(sqe.getMessage(), sqe);
}
}
public Map<String, Integer> getLexiconCounts(Indexer indexer, boolean originalContentOnly) {
Map<String, Integer> map = new LinkedHashMap<String, Integer>();
String[] captions =
captionToExpandedQuery.keySet().toArray(new String[captionToExpandedQuery.size()]);
for (String caption : captions) {
String query = captionToExpandedQuery.get(caption);
if (query == null) {
log.warn("Skipping unknown caption '" + caption + "'");
continue;
}
Integer cnt = 0;
try {
if (originalContentOnly)
cnt = indexer.getTotalHits(query, false, Indexer.QueryType.ORIGINAL);
else cnt = indexer.getTotalHits(query, false, Indexer.QueryType.FULL);
} catch (Exception e) {
Util.print_exception("Exception while collecting lexicon counts", e, log);
}
map.put(caption, cnt);
}
return map;
}
/**
* main entry point: returns a category -> docs map for each (non-zero) category in the current
* captionToQueryMap.
*
* @indexer must already have run
* @docs results are restrictes to these docs. assumes all docs if docs is null or empty.
* @captions (null/none = all)
* <p>vihari This is a weird name for a method that returns documents with emotions instead
* of emotions.
*/
public Map<String, Collection<Document>> getEmotions(
Indexer indexer,
Collection<Document> docs,
boolean originalContentOnly,
String... captions) {
Map<String, Collection<Document>> result = new LinkedHashMap<String, Collection<Document>>();
Set<Document> docs_set = Util.castOrCloneAsSet(docs);
// for (String[] emotion: emotionsData)
String[] selected_captions =
captions.length > 0 ? captions : captionToExpandedQuery.keySet().toArray(new String[0]);
for (String caption : selected_captions) {
String query = captionToExpandedQuery.get(caption);
if (query == null) {
log.warn("Skipping unknown caption '" + caption + "'");
continue;
}
// query is simply word1|word2|word3 etc for that sentiment
// the -1 indicates that we want all docs in the indexer that match the query
int threshold = 1;
Indexer.QueryOptions options = new Indexer.QueryOptions();
options.setThreshold(threshold);
options.setQueryType(Indexer.QueryType.ORIGINAL);
Collection<Document> docsForCaption = indexer.docsForQuery(query, options);
/*
log.info (docsForCaption.size() + " before");
threshold = 2;
docsForCaption = indexer.docsForQuery(query, -1, threshold);
log.info (docsForCaption.size() + " after");
*/
// Set<Document> docs = indexer.docsWithPhraseThreshold(query, -1, 2); // in future, we
// may have a higher threshold for sentiment matching
// if @param docs is present, retain only those docs that match, otherwise retain all
if (!Util.nullOrEmpty(docs_set))
// docsForCaption.retainAll(docs_set);
docsForCaption = Util.listIntersection(docsForCaption, docs_set);
// put it in the result only if at least 1 doc matches
if (docsForCaption.size() > 0) result.put(caption, docsForCaption);
}
return result;
}
