protected boolean isConfigured() {
boolean isConfigured = true;
// check if there are input files and if they exist.
if (inputFiles.size() > 0) {
for (Iterator<File> iterator = inputFiles.iterator(); iterator.hasNext(); ) {
File next = iterator.next();
if (!next.exists()) {
isConfigured = false;
System.err.println("Input file " + next.getPath() + " does not exist.");
}
}
}
return isConfigured;
}
public void run() {
// do it ...
try {
IndexWriter indexWriter =
LuceneUtils.createIndexWriter(
indexPath, overwriteIndex, LuceneUtils.AnalyzerType.WhitespaceAnalyzer);
for (Iterator<File> iterator = inputFiles.iterator(); iterator.hasNext(); ) {
File inputFile = iterator.next();
if (verbose) System.out.println("Processing " + inputFile.getPath() + ".");
if (verbose) System.out.println("Counting images.");
run = 0;
readFile(indexWriter, inputFile);
if (verbose) System.out.printf("%d images found in the data file.\n", docCount);
int numberOfRepresentatives = 1000; // TODO: clever selection.
// select a number of representative "fixed stars" randomly from file
if (numberOfRepresentatives > Math.sqrt(docCount))
numberOfRepresentatives = (int) Math.sqrt(docCount);
if (verbose)
System.out.printf(
"Selecting %d representative images for hashing.\n", numberOfRepresentatives);
representativesID = new HashSet<Integer>(numberOfRepresentatives);
while (representativesID.size() < numberOfRepresentatives) {
representativesID.add((int) Math.floor(Math.random() * (docCount - 1)));
}
representatives = new ArrayList<LireFeature>(numberOfRepresentatives);
docCount = 0;
run = 1;
if (verbose) System.out.println("Now getting representatives from the data file.");
readFile(indexWriter, inputFile);
docCount = 0;
run = 2;
if (verbose) System.out.println("Finally we start the indexing process, please wait ...");
readFile(indexWriter, inputFile);
if (verbose) System.out.println("Indexing finished.");
}
indexWriter.commit();
indexWriter.close();
} catch (Exception e) {
e.printStackTrace();
}
}
private int[] getHashes(LireFeature feature) {
// result = new int[maximumHits];
hashingResultScoreDocs.clear();
maxDistance = 0f;
tmpScore = 0f;
int rep = 0;
LireFeature tmpFeature;
for (Iterator<LireFeature> iterator = representatives.iterator(); iterator.hasNext(); ) {
tmpFeature = iterator.next();
tmpScore = tmpFeature.getDistance(feature);
if (hashingResultScoreDocs.size() < maximumHits) {
hashingResultScoreDocs.add(new SimpleResult(tmpScore, null, rep));
maxDistance = Math.max(maxDistance, tmpScore);
} else if (tmpScore < maxDistance) {
hashingResultScoreDocs.add(new SimpleResult(tmpScore, null, rep));
}
while (hashingResultScoreDocs.size() > maximumHits) {
hashingResultScoreDocs.remove(hashingResultScoreDocs.last());
maxDistance = hashingResultScoreDocs.last().getDistance();
}
rep++;
}
rep = 0;
for (Iterator<SimpleResult> iterator = hashingResultScoreDocs.iterator();
iterator.hasNext(); ) {
SimpleResult next = iterator.next();
result[rep] = next.getIndexNumber();
rep++;
}
return result;
}