public List<Integer> getResultWithAND(String query) throws Exception {
String[] terms = query.split(" ");
List<Integer> results = new LinkedList<Integer>();
for (String term : terms) {
String searchTerm = getStemmedQuery(term);
int termid = getTermId(searchTerm);
List<Integer> docs = new LinkedList<Integer>();
if (termid > 0) {
TreeMap<Integer, List<Integer>> docpos = null;
if (!termdocposindex.containsKey(termid)) {
getTermDocPosIndex(termid);
docpos = termdocposindex.get(termid);
} else {
docpos = termdocposindex.get(termid);
}
Iterator docIter = docpos.keySet().iterator();
Object doc = null;
while (docIter.hasNext()) {
doc = docIter.next();
List<Integer> lpos1 = docpos.get(doc);
// list of docs
docs.add((Integer) doc);
}
}
if (results.isEmpty()) results.addAll(docs);
else results.retainAll(docs);
}
return results;
}
public TreeMap<Double, Integer> calculateTopTenQuery() throws IOException {
// calculating relevance and irrelevance docs
ZipInputStream zis;
FileInputStream fis = null;
ZipFile zipFile = null;
ZipEntry ze;
fis = new FileInputStream(indexFile);
zipFile = new ZipFile(indexFile);
zis = new ZipInputStream(fis);
TreeMap<Double, Integer> results = new TreeMap<Double, Integer>();
TreeMap<Integer, List<Integer>> docTermIndex = new TreeMap<Integer, List<Integer>>();
List<Integer> termids = new LinkedList<Integer>();
if (relFlag) {
String[] reldocs = relevance.split(",");
for (String reldoc : reldocs) {
int docid = getDocId(reldoc);
// Editing starts here
while ((ze = zis.getNextEntry()) != null) {
String indexData = "";
String[] splittedString = null;
InputStream is = zipFile.getInputStream(ze);
InputStreamReader r = new InputStreamReader(is);
BufferedReader br = new BufferedReader(r);
indexData = br.readLine();
splittedString = indexData.split(";");
for (int i = 0; i < splittedString.length; i++) {
if (splittedString[i].contains(":" + docid + ":")) {
String[] subToken = splittedString[i].toString().split(":");
int termid = Integer.parseInt(subToken[0]);
int sDocId = Integer.parseInt(subToken[1]);
if (!termids.contains(termid)) {
termids.add(termid);
// docTermIndex.put(sDocId, termids);
// calculating the terms' vector scores here
}
}
}
}
for (int termid : termids) {
System.out.println(termid);
getTermDocPosIndex(termid);
}
// Editing ends here
// Last editing here!!
Iterator iterator = termdocposindex.keySet().iterator();
Object obj;
List<Integer> lterms = new LinkedList<Integer>();
TreeMap<Integer, List<Integer>> tdocs = new TreeMap<Integer, List<Integer>>();
while (iterator.hasNext()) {
obj = iterator.next();
tdocs = termdocposindex.get(obj);
Iterator iterator2 = tdocs.keySet().iterator();
Object obj2;
while (iterator2.hasNext()) {
obj2 = iterator2.next();
if ((Integer) obj2 == docid) {
lterms.add((Integer) obj2);
System.out.println(obj.toString());
}
}
}
// System.out.println(docid);
}
} else if (irrelFlag) {
String[] irreldocs = irrelevance.split(",");
for (String irreldoc : irreldocs) {
int docid = getDocId(irreldoc);
// System.out.println(docid);
}
}
return results;
}
public TreeMap<Double, List<Integer>> getCosineScore(String query) throws IOException {
TreeMap<Double, List<Integer>> rankedResults = new TreeMap<Double, List<Integer>>();
String[] terms = null;
if (query.contains("-")) {
terms = query.split(" ");
String refinedQuery = "";
for (String term : terms) {
if (term.contains("-")) {
String[] splittedTerms = term.split("-");
term = "";
for (String splittedTerm : splittedTerms) {
term += splittedTerm + " ";
}
}
refinedQuery += term + " ";
}
query = refinedQuery;
terms = query.split(" ");
} else {
terms = query.split(" ");
}
double start = System.currentTimeMillis();
List<Integer> results = new LinkedList<Integer>();
HashMap<Integer, Double> docScores = new HashMap<Integer, Double>();
TreeMap<Integer, List<Integer>> docpos = null;
TreeMap<Integer, Double> listOfDocScores = new TreeMap<Integer, Double>();
double tfCorpus = 0;
for (String term : terms) {
String searchTerm = getStemmedQuery(term.trim());
int termid = getTermId(searchTerm);
List<Integer> docs = new LinkedList<Integer>();
if (termid > 0) {
if (!termdocposindex.containsKey(termid)) {
getTermDocPosIndex(termid);
docpos = termdocposindex.get(termid);
} else {
docpos = termdocposindex.get(termid);
}
Iterator docIter = docpos.keySet().iterator();
Object doc = null;
while (docIter.hasNext()) {
doc = docIter.next();
docs.add((Integer) doc);
}
// select only documents contain the query
if (results.isEmpty()) results.addAll(docs);
else {
results.retainAll(docs);
}
}
// calculating tf in document
for (int result : results) {
double tfDocument = 0;
List<Integer> lPositions = docpos.get(result);
tfDocument = lPositions.size();
tfCorpus += tfDocument;
listOfDocScores.put(result, tfDocument);
}
}
// calculating tf in corpus
Iterator iteratorKey = listOfDocScores.keySet().iterator();
Object docKey;
double tfDocument = 0;
while (iteratorKey.hasNext()) {
docKey = iteratorKey.next();
tfDocument = listOfDocScores.get(docKey);
if (!docScores.containsKey(docKey)) {
docScores.put((Integer) docKey, tfDocument / tfCorpus);
}
}
// rank the results ascending
for (Integer docid : docScores.keySet()) {
if (rankedResults.containsKey(docScores.get(docid))) {
List<Integer> l = rankedResults.get(docScores.get(docid));
if (l.contains(docid)) {
rankedResults.put(docScores.get(docid), l);
}
} else {
List<Integer> l = new LinkedList<Integer>();
l.add(docid);
rankedResults.put(docScores.get(docid), l);
}
}
double end = (System.currentTimeMillis() - start) / 1000;
System.out.println("Match documents found in " + end + " seconds.");
return rankedResults;
}
