// TODO: use linked list in place of arraylist. It may get faster.
private void findAnchorWords() {
// intersect unique words in the ground truth and ocr output
anchorsRef = new ArrayList<IndexEntry>(MAX_NUMBER_OF_CANDIDATE_ANCHORS);
anchorsCand = new ArrayList<IndexEntry>(MAX_NUMBER_OF_CANDIDATE_ANCHORS);
ArrayList<IndexEntry> newAnchorsRef =
new ArrayList<IndexEntry>(MAX_NUMBER_OF_CANDIDATE_ANCHORS);
ArrayList<IndexEntry> newAnchorsCand =
new ArrayList<IndexEntry>(MAX_NUMBER_OF_CANDIDATE_ANCHORS);
int i = 0;
int startRef, endRef, startCand, endCand;
boolean start = true;
do {
if (start) { // initial condition: input sequences are considered to be two big segments.
startRef = 0;
endRef = refIndex.getNumOfTokens();
startCand = 0;
endCand = ocrIndex.getNumOfTokens();
i--; // not to skip the first subsegment in the list
start = false;
} else {
// if the first iteration could not find any anchor words, then simply quit.
if (anchorsRef.isEmpty()) {
break;
}
// RECURSIVE STAGE
// INDEX SUB-SEGMENTS FOR MORE ANCHOR WORDS
if (i == 0) { // first segment
startRef = 0;
endRef = anchorsRef.get(i).getPosInt();
startCand = 0;
endCand = anchorsCand.get(i).getPosInt();
} // last segment
else if (i == anchorsRef.size()) {
startRef = anchorsRef.get(i - 1).getPosInt() + 1;
endRef = refIndex.getNumOfTokens();
startCand = anchorsCand.get(i - 1).getPosInt() + 1;
endCand = ocrIndex.getNumOfTokens();
} else { // inbetween segments
startRef = anchorsRef.get(i - 1).getPosInt() + 1;
endRef = anchorsRef.get(i).getPosInt();
startCand = anchorsCand.get(i - 1).getPosInt() + 1;
endCand = anchorsCand.get(i).getPosInt();
}
}
// if one of the segments is still large, go on recursive operations.
if (((long) (endRef - startRef)) > MAX_SEGMENT_LENGTH
|| ((long) (endCand - startCand)) > MAX_SEGMENT_LENGTH) {
newAnchorsRef.clear();
newAnchorsCand.clear();
findCommonUniqueWords(startRef, endRef, startCand, endCand, newAnchorsRef, newAnchorsCand);
// if there are no common unique words, then do not try to chop it down
if (newAnchorsRef.isEmpty()) {
if (i < anchorsRef.size()) {
i++; // skip over the segment
continue;
} else {
break; // if it is the last segment, quit anchor word search
}
}
} else { // if the segment is small enough to use edit distance based alignment, then skip
// over
if (i < anchorsRef.size()) {
i++;
continue;
} else {
break; // if it is the last segment, quit anchor word search
}
}
int LCSindices[] = LCS.findLCS(newAnchorsRef, newAnchorsCand);
int lcsLength = LCSindices.length / 2;
// use achors if and only if they are in the longest common subsequence
ArrayList<IndexEntry> newAnchorsRefRefined = new ArrayList<IndexEntry>();
ArrayList<IndexEntry> newAnchorsCandRefined = new ArrayList<IndexEntry>();
for (int kk = 0; kk < lcsLength; kk++) {
IndexEntry refEntry = newAnchorsRef.get(LCSindices[kk]);
IndexEntry candEntry = newAnchorsCand.get(LCSindices[(kk + lcsLength)]);
newAnchorsRefRefined.add(refEntry);
newAnchorsCandRefined.add(candEntry);
}
newAnchorsRef = newAnchorsRefRefined;
newAnchorsCand = newAnchorsCandRefined;
if (lcsLength > 0) {
if (i == -1) { // for the first iteration
anchorsRef.addAll(newAnchorsRef);
anchorsCand.addAll(newAnchorsCand);
} else {
anchorsRef.addAll(i, newAnchorsRef);
anchorsCand.addAll(i, newAnchorsCand);
i--;
}
}
i++;
// break; // test - stop recursion at the first stage
} while (i <= anchorsRef.size());
}
public static void main(String[] args) {
if (args.length != 2) {
System.err.println("Please pass the names of the matrix and string files as parameters.");
System.exit(1);
}
long start, end;
System.out.println("*** PROBLEM #1: Largest Square Submatrix of 1's");
try {
int[][] matrix = readMatrix(args[0]);
start = System.currentTimeMillis();
LSS.findLSS(matrix);
end = System.currentTimeMillis();
System.out.println("(It took " + (end - start) + " ms.)");
} catch (FileNotFoundException ex) {
System.err.println("The file: '" + args[0] + "' was not found.");
}
System.out.println();
System.out.println("*** PROBLEM #2: Longest Common Subsequence of two strings");
try {
String[] strings = readLCSWords(args[1]);
System.out.println("The lonesg common subsequence between the two words in the file is:");
start = System.currentTimeMillis();
String lcs = LCS.calcLCS(strings[0], strings[1]);
end = System.currentTimeMillis();
System.out.println(lcs);
System.out.println("The LCS is of size " + lcs.length());
System.out.println("(It took " + (end - start) + " ms.)");
} catch (FileNotFoundException ex) {
System.err.println("The file: '" + args[1] + "' was not found.");
} catch (IllegalArgumentException ex) {
System.err.println(ex.getMessage());
}
System.out.println();
System.out.println("*** PROBLEM #3: Optimal Coin Change");
final List<Integer> coins = new ArrayList<>();
coins.add(1);
coins.add(5);
coins.add(10);
coins.add(18);
coins.add(25);
OptimalChangeCalc calc = new OptimalChangeCalc(coins);
start = System.currentTimeMillis();
calc.calcOptimalChange(100);
end = System.currentTimeMillis();
System.out.println("Coins utilized: 1, 5, 10, 18, 25");
System.out.println(
"The following is a table of cents along with the list "
+ " of minimum coins needed to make that change using the coins above.");
calc.printOptimalChange();
System.out.println("(It took " + (end - start) + " ms.)");
}
public static void main(String[] args) {
String a = "ABCBDAB";
String b = "BDCABA";
int lcs = LCS.lcs_length(a, b);
System.out.println(lcs);
}
public static void main(String[] args) {
char[] x = {' ', 'A', 'B', 'C', 'B', 'D', 'A', 'B'};
char[] y = {' ', 'B', 'D', 'C', 'A', 'B', 'A'};
LCS lcs = new LCS();
lcs.printLCS(lcs.lcsLength(x, y), x, x.length - 1, y.length - 1);
}