public Tree<String> getBestParseOld(List<String> sentence) {
// TODO: This implements the CKY algorithm
CounterMap<String, String> parseScores = new CounterMap<String, String>();
System.out.println(sentence.toString());
// First deal with the lexicons
int index = 0;
int span = 1; // All spans are 1 at the lexicon level
for (String word : sentence) {
for (String tag : lexicon.getAllTags()) {
double score = lexicon.scoreTagging(word, tag);
if (score >= 0.0) { // This lexicon may generate this word
// We use a counter map in order to store the scores for this sentence parse.
parseScores.setCount(index + " " + (index + span), tag, score);
}
}
index = index + 1;
}
// handle unary rules now
HashMap<String, Triplet<Integer, String, String>> backHash =
new HashMap<
String, Triplet<Integer, String, String>>(); // hashmap to store back propation
// System.out.println("Lexicons found");
Boolean added = true;
while (added) {
added = false;
for (index = 0; index < sentence.size(); index++) {
// For each index+ span pair, get the counter.
Counter<String> count = parseScores.getCounter(index + " " + (index + span));
PriorityQueue<String> countAsPQ = count.asPriorityQueue();
while (countAsPQ.hasNext()) {
String entry = countAsPQ.next();
// System.out.println("I am fine here!!");
List<UnaryRule> unaryRules = grammar.getUnaryRulesByChild(entry);
for (UnaryRule rule : unaryRules) {
// These are the unary rules which might give rise to the above preterminal
double prob =
rule.getScore() * parseScores.getCount(index + " " + (index + span), entry);
if (prob > parseScores.getCount(index + " " + (index + span), rule.parent)) {
parseScores.setCount(index + " " + (index + span), rule.parent, prob);
backHash.put(
index + " " + (index + span) + " " + rule.parent,
new Triplet<Integer, String, String>(-1, entry, null));
added = true;
}
}
}
}
}
// System.out.println("Lexicon unaries dealt with");
// Now work with the grammar to produce higher level probabilities
for (span = 2; span <= sentence.size(); span++) {
for (int begin = 0; begin <= (sentence.size() - span); begin++) {
int end = begin + span;
for (int split = begin + 1; split <= end - 1; split++) {
Counter<String> countLeft = parseScores.getCounter(begin + " " + split);
Counter<String> countRight = parseScores.getCounter(split + " " + end);
// List<BinaryRule> leftRules= new ArrayList<BinaryRule>();
HashMap<Integer, BinaryRule> leftMap = new HashMap<Integer, BinaryRule>();
// List<BinaryRule> rightRules=new ArrayList<BinaryRule>();
HashMap<Integer, BinaryRule> rightMap = new HashMap<Integer, BinaryRule>();
for (String entry : countLeft.keySet()) {
for (BinaryRule rule : grammar.getBinaryRulesByLeftChild(entry)) {
if (!leftMap.containsKey(rule.hashCode())) {
leftMap.put(rule.hashCode(), rule);
}
}
}
for (String entry : countRight.keySet()) {
for (BinaryRule rule : grammar.getBinaryRulesByRightChild(entry)) {
if (!rightMap.containsKey(rule.hashCode())) {
rightMap.put(rule.hashCode(), rule);
}
}
}
// System.out.println("About to enter the rules loops");
for (Integer ruleHash : leftMap.keySet()) {
if (rightMap.containsKey(ruleHash)) {
BinaryRule ruleRight = rightMap.get(ruleHash);
double prob =
ruleRight.getScore()
* parseScores.getCount(begin + " " + split, ruleRight.leftChild)
* parseScores.getCount(split + " " + end, ruleRight.rightChild);
// System.out.println(begin+" "+ end +" "+ ruleRight.parent+ " "+ prob);
if (prob > parseScores.getCount(begin + " " + end, ruleRight.parent)) {
// System.out.println(begin+" "+ end +" "+ ruleRight.parent+ " "+ prob);
// System.out.println("parentrule :"+ ruleRight.getParent());
parseScores.setCount(begin + " " + end, ruleRight.getParent(), prob);
backHash.put(
begin + " " + end + " " + ruleRight.parent,
new Triplet<Integer, String, String>(
split, ruleRight.leftChild, ruleRight.rightChild));
}
}
}
// System.out.println("Exited rules loop");
}
// System.out.println("Grammar found for " + begin + " "+ end);
// Now handle unary rules
added = true;
while (added) {
added = false;
Counter<String> count = parseScores.getCounter(begin + " " + end);
PriorityQueue<String> countAsPriorityQueue = count.asPriorityQueue();
while (countAsPriorityQueue.hasNext()) {
String entry = countAsPriorityQueue.next();
List<UnaryRule> unaryRules = grammar.getUnaryRulesByChild(entry);
for (UnaryRule rule : unaryRules) {
double prob = rule.getScore() * parseScores.getCount(begin + " " + (end), entry);
if (prob > parseScores.getCount(begin + " " + (end), rule.parent)) {
parseScores.setCount(begin + " " + (end), rule.parent, prob);
backHash.put(
begin + " " + (end) + " " + rule.parent,
new Triplet<Integer, String, String>(-1, entry, null));
added = true;
}
}
}
}
// System.out.println("Unaries dealt for " + begin + " "+ end);
}
}
// Create and return the parse tree
Tree<String> parseTree = new Tree<String>("null");
// System.out.println(parseScores.getCounter(0+" "+sentence.size()).toString());
String parent = parseScores.getCounter(0 + " " + sentence.size()).argMax();
if (parent == null) {
System.out.println(parseScores.getCounter(0 + " " + sentence.size()).toString());
System.out.println("THIS IS WEIRD");
}
parent = "ROOT";
parseTree = getParseTreeOld(sentence, backHash, 0, sentence.size(), parent);
// System.out.println("PARSE SCORES");
// System.out.println(parseScores.toString());
// System.out.println("BACK HASH");
// System.out.println(backHash.toString());
// parseTree = addRoot(parseTree);
// System.out.println(parseTree.toString());
// return parseTree;
return TreeAnnotations.unAnnotateTree(parseTree);
}
public void setTransitionCount(S start, S end, double count) {
forwardTransitions.setCount(start, end, count);
backwardTransitions.setCount(end, start, count);
}
private CounterMap<String,String> trainEM(int maxIterations) {
Set<String> englishVocab = new HashSet<String>();
Set<String> frenchVocab = new HashSet<String>();
CounterMap<String,String> translations = new CounterMap<String,String>();
englishVocab.add(NULL);
int iteration = 0;
final double thresholdProb = 0.0001;
for (SentencePair sentencePair : trainingSentencePairs) {
List<String> frenchWords = sentencePair.getFrenchWords();
List<String> englishWords = sentencePair.getEnglishWords();
// add words from list to vocabulary sets
englishVocab.addAll(englishWords);
frenchVocab.addAll(frenchWords);
}
System.out.println("Ready");
// We need to initialize translations.getCount(f,e) uniformly
// t(f|e) summed over all e in {E + NULL} = 1
final double initialCount = 1.0 / englishVocab.size();
while(iteration < maxIterations) {
CounterMap<String,String> counts = new CounterMap<String,String>(); // set count(f|e) to 0 for all e,f
Counter<String> totalEnglish = new Counter<String>(); // set total(e) to 0 for all e
// E-step: loop over all sentences and update counts
for (SentencePair sentencePair : trainingSentencePairs) {
List<String> frenchWords = sentencePair.getFrenchWords();
List<String> englishWords = sentencePair.getEnglishWords();
int numFrenchWords = frenchWords.size();
int numEnglishWords = englishWords.size();
Counter<String> sTotalF = new Counter<String>();
// compute normalization constant sTotalF
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
String f = frenchWords.get(frenchPosition);
// initialize and compute for English = NULL
if (!translations.containsKey(f) && initialize)
translations.setCount(f, NULL, initialCount);
else if (!translations.containsKey(f))
translations.setCount(f, NULL, thresholdProb);
sTotalF.incrementCount(f, translations.getCount(f, NULL));
for (int englishPosition = 0; englishPosition < numEnglishWords; englishPosition++) {
String e = englishWords.get(englishPosition);
if (!(translations.getCounter(f)).containsKey(e) && initialize)
translations.setCount(f, e, initialCount);
else if (!(translations.getCounter(f)).containsKey(e))
translations.setCount(f, e, thresholdProb);
sTotalF.incrementCount(f, translations.getCount(f, e));
}
}
// collect counts in counts and totalEnglish
for (int frenchPosition = 0; frenchPosition < numFrenchWords; frenchPosition++) {
String f = frenchWords.get(frenchPosition);
// collect counts for English = NULL
double count = translations.getCount(f, NULL) / sTotalF.getCount(f);
counts.incrementCount(NULL, f, count);
totalEnglish.incrementCount(NULL, count);
for (int englishPosition = 0; englishPosition < numEnglishWords; englishPosition++) {
String e = englishWords.get(englishPosition);
count = translations.getCount(f, e) / sTotalF.getCount(f);
counts.incrementCount(e, f, count);
totalEnglish.incrementCount(e, count);
}
}
} // end of E-step
System.out.println("Completed E-step");
// M-step: update probabilities with counts from E-step and check for convergence
iteration++;
for (String e : counts.keySet()) {//englishVocab) {
double normalizer = totalEnglish.getCount(e);
for (String f : (counts.getCounter(e)).keySet()) {//frenchVocab) {
// To speed implementation, we want to update translations only when count / normalizer > threshold
double prob = counts.getCount(e, f) / normalizer;
if (!initialize) {
if (prob > thresholdProb)
translations.setCount(f, e, prob);
else
(translations.getCounter(f)).removeKey(e);
}
else {
translations.setCount(f, e, prob);
}
}
}
System.out.println("Completed iteration " + iteration);
} // end of M-step
System.out.println("Trained!");
return translations;
}