private static Map<Integer, Alignment> readAlignments(String fileName) {
Map<Integer,Alignment> alignments = new HashMap<Integer, Alignment>();
try {
BufferedReader in = new BufferedReader(new FileReader(fileName));
while (in.ready()) {
String line = in.readLine();
String[] words = line.split("\\s+");
if (words.length != 4)
throw new RuntimeException("Bad alignment file "+fileName+", bad line was "+line);
Integer sentenceID = Integer.parseInt(words[0]);
Integer englishPosition = Integer.parseInt(words[1])-1;
Integer frenchPosition = Integer.parseInt(words[2])-1;
String type = words[3];
Alignment alignment = alignments.get(sentenceID);
if (alignment == null) {
alignment = new Alignment();
alignments.put(sentenceID, alignment);
}
alignment.addAlignment(englishPosition, frenchPosition, type.equals("S"));
}
} catch (IOException e) {
throw new RuntimeException(e);
}
return alignments;
}
static Alignment getFeatureContaining(List<Alignment> features, int right) {
int leftBounds = 0;
int rightBounds = features.size() - 1;
int idx = features.size() / 2;
int lastIdx = -1;
while (idx != lastIdx) {
lastIdx = idx;
Alignment f = features.get(idx);
if (f.contains(right)) {
return f;
}
if (f.getStart() > right) {
rightBounds = idx;
idx = (leftBounds + idx) / 2;
} else {
leftBounds = idx;
idx = (rightBounds + idx) / 2;
}
}
// Check the extremes
if (features.get(0).contains(right)) {
return features.get(0);
}
if (features.get(rightBounds).contains(right)) {
return features.get(rightBounds);
}
return null;
}
public Alignment alignSentencePair(SentencePair sentencePair) {
Alignment alignment = new Alignment();
List<String> frenchWords = sentencePair.getFrenchWords();
List<String> englishWords = sentencePair.getEnglishWords();
int numFrenchWords = frenchWords.size();
int numEnglishWords = englishWords.size();
// Model 1 assumes all alignments are equally likely
// So we can just take the argMax of t(f|e) to get the englishMaxPosition
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
String f = frenchWords.get(frenchPosition);
int englishMaxPosition = -1;
double maxTranslationProb = translationProbs.getCount(f, NULL);
for (int englishPosition = 0; englishPosition < numEnglishWords; englishPosition++) {
String e = englishWords.get(englishPosition);
double translationProb = translationProbs.getCount(f, e);
if (translationProb > maxTranslationProb) {
maxTranslationProb = translationProb;
englishMaxPosition = englishPosition;
}
}
alignment.addAlignment(englishMaxPosition, frenchPosition, true);
}
return alignment;
}
public Alignment alignSentencePair(SentencePair sentencePair) {
Alignment alignment = new Alignment();
List<String> frenchWords = sentencePair.getFrenchWords();
List<String> englishWords = sentencePair.getEnglishWords();
int numFrenchWords = frenchWords.size();
int numEnglishWords = englishWords.size();
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
String f = frenchWords.get(frenchPosition);
int englishMaxPosition = frenchPosition;
if (englishMaxPosition >= numEnglishWords)
englishMaxPosition = -1; // map French word to BASELINE if c(f,e) = 0 for all English words
double maxDice = 0;
for (int englishPosition = 0; englishPosition < numEnglishWords; englishPosition++) {
String e = englishWords.get(englishPosition);
double dice = getDiceCoefficient(f,e);
if (dice > maxDice) {
maxDice = dice;
englishMaxPosition = englishPosition;
}
}
alignment.addAlignment(englishMaxPosition, frenchPosition, true);
}
return alignment;
}
public Alignment alignSentencePair(SentencePair sentencePair) {
Alignment alignment = new Alignment();
List<String> frenchWords = sentencePair.getFrenchWords();
List<String> englishWords = sentencePair.getEnglishWords();
int numFrenchWords = frenchWords.size();
int numEnglishWords = englishWords.size();
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
String f = frenchWords.get(frenchPosition);
int englishMaxPosition = frenchPosition;
if (englishMaxPosition >= numEnglishWords)
englishMaxPosition = -1; // map French word to BASELINE if c(f,e) = 0 for all English words
double maxConditionalProb = 0;
for (int englishPosition = 0; englishPosition < numEnglishWords; englishPosition++) {
String e = englishWords.get(englishPosition);
double conditionalGivenEnglish = collocationCounts.getCount(f, e) / (eCounts.getCount(e));
if (conditionalGivenEnglish > maxConditionalProb) {
maxConditionalProb = conditionalGivenEnglish;
englishMaxPosition = englishPosition;
}
}
alignment.addAlignment(englishMaxPosition, frenchPosition, true);
}
return alignment;
}
public static String render(Alignment reference, Alignment proposed, SentencePair sentencePair) {
StringBuilder sb = new StringBuilder();
for (int frenchPosition = 0; frenchPosition < sentencePair.getFrenchWords().size(); frenchPosition++) {
for (int englishPosition = 0; englishPosition < sentencePair.getEnglishWords().size(); englishPosition++) {
boolean sure = reference.containsSureAlignment(englishPosition, frenchPosition);
boolean possible = reference.containsPossibleAlignment(englishPosition, frenchPosition);
char proposedChar = ' ';
if (proposed.containsSureAlignment(englishPosition, frenchPosition))
proposedChar = '#';
if (sure) {
sb.append('[');
sb.append(proposedChar);
sb.append(']');
} else {
if (possible) {
sb.append('(');
sb.append(proposedChar);
sb.append(')');
} else {
sb.append(' ');
sb.append(proposedChar);
sb.append(' ');
}
}
}
sb.append("| ");
sb.append(sentencePair.getFrenchWords().get(frenchPosition));
sb.append('\n');
}
for (int englishPosition = 0; englishPosition < sentencePair.getEnglishWords().size(); englishPosition++) {
sb.append("---");
}
sb.append("'\n");
boolean printed = true;
int index = 0;
while (printed) {
printed = false;
StringBuilder lineSB = new StringBuilder();
for (int englishPosition = 0; englishPosition < sentencePair.getEnglishWords().size(); englishPosition++) {
String englishWord = sentencePair.getEnglishWords().get(englishPosition);
if (englishWord.length() > index) {
printed = true;
lineSB.append(' ');
lineSB.append(englishWord.charAt(index));
lineSB.append(' ');
} else {
lineSB.append(" ");
}
}
index += 1;
if (printed) {
sb.append(lineSB);
sb.append('\n');
}
}
return sb.toString();
}
public Alignment alignSentencePair(SentencePair sentencePair) {
Alignment alignment = new Alignment();
int numFrenchWords = sentencePair.getFrenchWords().size();
int numEnglishWords = sentencePair.getEnglishWords().size();
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
int englishPosition = frenchPosition;
if (englishPosition >= numEnglishWords)
englishPosition = -1;
alignment.addAlignment(englishPosition, frenchPosition, true);
}
return alignment;
}
private static void predict(WordAligner wordAligner, List<SentencePair> testSentencePairs, String path) throws IOException {
BufferedWriter writer = new BufferedWriter(new FileWriter(path));
for (SentencePair sentencePair : testSentencePairs) {
Alignment proposedAlignment = wordAligner.alignSentencePair(sentencePair);
for (int frenchPosition = 0; frenchPosition < sentencePair.getFrenchWords().size(); frenchPosition++) {
for (int englishPosition = 0; englishPosition < sentencePair.getEnglishWords().size(); englishPosition++) {
if (proposedAlignment.containsSureAlignment(englishPosition, frenchPosition)) {
writer.write(frenchPosition + "-" + englishPosition + " ");
}
}
}
writer.write("\n");
}
writer.close();
}
/**
* Creates a MultipleAlignment from the given Alignment object. Alignment is roughly analogous to
* MultipleAlignment, but it's an interface and it corresponds to a multiple alignment that has
* been read from a data store somewhere (presumably a database, but perhaps also a file or other
* source). Creating a MultipleAlignment object "around" it basically makes it useable --
* MultipleAlignment will read the Alignment object and its corresponding AlignBlock pieces, and
* builds data structures so that the underlying multiple alignment can be accessed easily.
*
* @param a
*/
public MultipleAlignment(Alignment a) {
this();
Collection<AlignBlock> blocks = a.getAlignBlocks();
for (AlignBlock block : blocks) {
GappedAlignmentString gaps = new GappedAlignmentString(block.getBitString());
Genome g = block.getGenome();
String chrom = block.getChrom();
String name = String.format("%s_%s", g.getVersion(), chrom);
addGappedAlignment(name, gaps);
}
}
private static void test(WordAligner wordAligner, List<SentencePair> testSentencePairs, Map<Integer, Alignment> testAlignments, boolean verbose) {
int proposedSureCount = 0;
int proposedPossibleCount = 0;
int sureCount = 0;
int proposedCount = 0;
for (SentencePair sentencePair : testSentencePairs) {
Alignment proposedAlignment = wordAligner.alignSentencePair(sentencePair);
Alignment referenceAlignment = testAlignments.get(sentencePair.getSentenceID());
if (referenceAlignment == null)
throw new RuntimeException("No reference alignment found for sentenceID "+sentencePair.getSentenceID());
if (verbose) System.out.println("Alignment:\n"+Alignment.render(referenceAlignment,proposedAlignment,sentencePair));
for (int frenchPosition = 0; frenchPosition < sentencePair.getFrenchWords().size(); frenchPosition++) {
for (int englishPosition = 0; englishPosition < sentencePair.getEnglishWords().size(); englishPosition++) {
boolean proposed = proposedAlignment.containsSureAlignment(englishPosition, frenchPosition);
boolean sure = referenceAlignment.containsSureAlignment(englishPosition, frenchPosition);
boolean possible = referenceAlignment.containsPossibleAlignment(englishPosition, frenchPosition);
if (proposed && sure) proposedSureCount += 1;
if (proposed && possible) proposedPossibleCount += 1;
if (proposed) proposedCount += 1;
if (sure) sureCount += 1;
}
}
}
System.out.println("Precision: "+proposedPossibleCount/(double)proposedCount);
System.out.println("Recall: "+proposedSureCount/(double)sureCount);
System.out.println("AER: "+(1.0-(proposedSureCount+proposedPossibleCount)/(double)(sureCount+proposedCount)));
}