public FilterConfusingRules(LexicalizedParser parser)
{
BinaryGrammar binaryGrammar = parser.bg;
UnaryGrammar unaryGrammar = parser.ug;
Options op = parser.GetOp();
IIndex <string> stateIndex = parser.stateIndex;
foreach (UnaryRule unaryRule in unaryGrammar)
{
// only make one matrix for each parent state, and only use the
// basic category for that
string childState = stateIndex.Get(unaryRule.child);
string childBasic = op.Langpack().BasicCategory(childState);
unaryRules.Add(childBasic);
}
foreach (BinaryRule binaryRule in binaryGrammar)
{
// only make one matrix for each parent state, and only use the
// basic category for that
string leftState = stateIndex.Get(binaryRule.leftChild);
string leftBasic = op.Langpack().BasicCategory(leftState);
string rightState = stateIndex.Get(binaryRule.rightChild);
string rightBasic = op.Langpack().BasicCategory(rightState);
binaryRules.Add(leftBasic, rightBasic);
}
}
/// <exception cref="System.IO.IOException"/>
/// <exception cref="System.TypeLoadException"/>
public static void Main(string[] args)
{
string taggerFile = null;
string inputFile = null;
string outputFile = null;
double weight = 1.0;
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-tagger"))
{
taggerFile = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-input"))
{
inputFile = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-output"))
{
outputFile = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-weight"))
{
weight = double.ValueOf(args[argIndex + 1]);
argIndex += 2;
}
else
{
throw new ArgumentException("Unknown argument: " + args[argIndex]);
}
}
}
}
}
LexicalizedParser parser = ((LexicalizedParser)LexicalizedParser.LoadModel(inputFile));
MaxentTagger tagger = new MaxentTagger(taggerFile);
parser.reranker = new TaggerReranker(tagger, parser.GetOp());
parser.SaveParserToSerialized(outputFile);
}
public virtual void RunTest(string[] args)
{
// get a parser from file
LexicalizedParser pd = ((LexicalizedParser)LexicalizedParser.LoadModel(args[0]));
op = pd.GetOp();
// in case a serialized options was read in
Treebank testTreebank = op.tlpParams.MemoryTreebank();
int testlow = System.Convert.ToInt32(args[2]);
int testhigh = System.Convert.ToInt32(args[3]);
testTreebank.LoadPath(args[1], new NumberRangeFileFilter(testlow, testhigh, true));
op.SetOptionsOrWarn(args, 4, args.Length);
TestOnTreebank(pd, new EnglishTreebankParserParams(), testTreebank, args[1], pd.stateIndex);
}
public DVParser(LexicalizedParser parser)
{
this.parser = parser;
this.op = parser.GetOp();
if (op.trainOptions.randomSeed == 0)
{
op.trainOptions.randomSeed = Runtime.NanoTime();
log.Info("Random seed not set, using randomly chosen seed of " + op.trainOptions.randomSeed);
}
else
{
log.Info("Random seed set to " + op.trainOptions.randomSeed);
}
log.Info("Word vector file: " + op.lexOptions.wordVectorFile);
log.Info("Size of word vectors: " + op.lexOptions.numHid);
log.Info("Number of hypothesis trees to train against: " + op.trainOptions.dvKBest);
log.Info("Number of trees in one batch: " + op.trainOptions.batchSize);
log.Info("Number of iterations of trees: " + op.trainOptions.trainingIterations);
log.Info("Number of qn iterations per batch: " + op.trainOptions.qnIterationsPerBatch);
log.Info("Learning rate: " + op.trainOptions.learningRate);
log.Info("Delta margin: " + op.trainOptions.deltaMargin);
log.Info("regCost: " + op.trainOptions.regCost);
log.Info("Using unknown word vector for numbers: " + op.trainOptions.unknownNumberVector);
log.Info("Using unknown dashed word vector heuristics: " + op.trainOptions.unknownDashedWordVectors);
log.Info("Using unknown word vector for capitalized words: " + op.trainOptions.unknownCapsVector);
log.Info("Using unknown number vector for Chinese words: " + op.trainOptions.unknownChineseNumberVector);
log.Info("Using unknown year vector for Chinese words: " + op.trainOptions.unknownChineseYearVector);
log.Info("Using unknown percent vector for Chinese words: " + op.trainOptions.unknownChinesePercentVector);
log.Info("Initial matrices scaled by: " + op.trainOptions.scalingForInit);
log.Info("Training will use " + op.trainOptions.trainingThreads + " thread(s)");
log.Info("Context words are " + ((op.trainOptions.useContextWords) ? "on" : "off"));
log.Info("Model will " + ((op.trainOptions.dvSimplifiedModel) ? string.Empty : "not ") + "be simplified");
this.dvModel = new DVModel(op, parser.stateIndex, parser.ug, parser.bg);
if (dvModel.unaryTransform.Count != dvModel.unaryScore.Count)
{
throw new AssertionError("Unary transform and score size not the same");
}
if (dvModel.binaryTransform.Size() != dvModel.binaryScore.Size())
{
throw new AssertionError("Binary transform and score size not the same");
}
}
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
IDictionary <string, int> flagsToNumArgs = Generics.NewHashMap();
flagsToNumArgs["-parser"] = int.Parse(3);
flagsToNumArgs["-lex"] = int.Parse(3);
flagsToNumArgs["-test"] = int.Parse(2);
flagsToNumArgs["-out"] = int.Parse(1);
flagsToNumArgs["-lengthPenalty"] = int.Parse(1);
flagsToNumArgs["-penaltyType"] = int.Parse(1);
flagsToNumArgs["-maxLength"] = int.Parse(1);
flagsToNumArgs["-stats"] = int.Parse(2);
IDictionary <string, string[]> argMap = StringUtils.ArgsToMap(args, flagsToNumArgs);
bool eval = argMap.Contains("-eval");
PrintWriter pw = null;
if (argMap.Contains("-out"))
{
pw = new PrintWriter(new OutputStreamWriter(new FileOutputStream((argMap["-out"])[0]), "GB18030"), true);
}
log.Info("ChineseCharacterBasedLexicon called with args:");
ChineseTreebankParserParams ctpp = new ChineseTreebankParserParams();
for (int i = 0; i < args.Length; i++)
{
ctpp.SetOptionFlag(args, i);
log.Info(" " + args[i]);
}
log.Info();
Options op = new Options(ctpp);
if (argMap.Contains("-stats"))
{
string[] statArgs = (argMap["-stats"]);
MemoryTreebank rawTrainTreebank = op.tlpParams.MemoryTreebank();
IFileFilter trainFilt = new NumberRangesFileFilter(statArgs[1], false);
rawTrainTreebank.LoadPath(new File(statArgs[0]), trainFilt);
log.Info("Done reading trees.");
MemoryTreebank trainTreebank;
if (argMap.Contains("-annotate"))
{
trainTreebank = new MemoryTreebank();
TreeAnnotator annotator = new TreeAnnotator(ctpp.HeadFinder(), ctpp, op);
foreach (Tree tree in rawTrainTreebank)
{
trainTreebank.Add(annotator.TransformTree(tree));
}
log.Info("Done annotating trees.");
}
else
{
trainTreebank = rawTrainTreebank;
}
PrintStats(trainTreebank, pw);
System.Environment.Exit(0);
}
int maxLength = 1000000;
// Test.verbose = true;
if (argMap.Contains("-norm"))
{
op.testOptions.lengthNormalization = true;
}
if (argMap.Contains("-maxLength"))
{
maxLength = System.Convert.ToInt32((argMap["-maxLength"])[0]);
}
op.testOptions.maxLength = 120;
bool combo = argMap.Contains("-combo");
if (combo)
{
ctpp.useCharacterBasedLexicon = true;
op.testOptions.maxSpanForTags = 10;
op.doDep = false;
op.dcTags = false;
}
LexicalizedParser lp = null;
ILexicon lex = null;
if (argMap.Contains("-parser"))
{
string[] parserArgs = (argMap["-parser"]);
if (parserArgs.Length > 1)
{
IFileFilter trainFilt = new NumberRangesFileFilter(parserArgs[1], false);
lp = LexicalizedParser.TrainFromTreebank(parserArgs[0], trainFilt, op);
if (parserArgs.Length == 3)
{
string filename = parserArgs[2];
log.Info("Writing parser in serialized format to file " + filename + " ");
System.Console.Error.Flush();
ObjectOutputStream @out = IOUtils.WriteStreamFromString(filename);
@out.WriteObject(lp);
@out.Close();
log.Info("done.");
}
}
else
{
string parserFile = parserArgs[0];
lp = LexicalizedParser.LoadModel(parserFile, op);
}
lex = lp.GetLexicon();
op = lp.GetOp();
ctpp = (ChineseTreebankParserParams)op.tlpParams;
}
if (argMap.Contains("-rad"))
{
ctpp.useUnknownCharacterModel = true;
}
if (argMap.Contains("-lengthPenalty"))
{
ctpp.lengthPenalty = double.Parse((argMap["-lengthPenalty"])[0]);
}
if (argMap.Contains("-penaltyType"))
{
ctpp.penaltyType = System.Convert.ToInt32((argMap["-penaltyType"])[0]);
}
if (argMap.Contains("-lex"))
{
string[] lexArgs = (argMap["-lex"]);
if (lexArgs.Length > 1)
{
IIndex <string> wordIndex = new HashIndex <string>();
IIndex <string> tagIndex = new HashIndex <string>();
lex = ctpp.Lex(op, wordIndex, tagIndex);
MemoryTreebank rawTrainTreebank = op.tlpParams.MemoryTreebank();
IFileFilter trainFilt = new NumberRangesFileFilter(lexArgs[1], false);
rawTrainTreebank.LoadPath(new File(lexArgs[0]), trainFilt);
log.Info("Done reading trees.");
MemoryTreebank trainTreebank;
if (argMap.Contains("-annotate"))
{
trainTreebank = new MemoryTreebank();
TreeAnnotator annotator = new TreeAnnotator(ctpp.HeadFinder(), ctpp, op);
foreach (Tree tree in rawTrainTreebank)
{
tree = annotator.TransformTree(tree);
trainTreebank.Add(tree);
}
log.Info("Done annotating trees.");
}
else
{
trainTreebank = rawTrainTreebank;
}
lex.InitializeTraining(trainTreebank.Count);
lex.Train(trainTreebank);
lex.FinishTraining();
log.Info("Done training lexicon.");
if (lexArgs.Length == 3)
{
string filename = lexArgs.Length == 3 ? lexArgs[2] : "parsers/chineseCharLex.ser.gz";
log.Info("Writing lexicon in serialized format to file " + filename + " ");
System.Console.Error.Flush();
ObjectOutputStream @out = IOUtils.WriteStreamFromString(filename);
@out.WriteObject(lex);
@out.Close();
log.Info("done.");
}
}
else
{
string lexFile = lexArgs.Length == 1 ? lexArgs[0] : "parsers/chineseCharLex.ser.gz";
log.Info("Reading Lexicon from file " + lexFile);
ObjectInputStream @in = IOUtils.ReadStreamFromString(lexFile);
try
{
lex = (ILexicon)@in.ReadObject();
}
catch (TypeLoadException)
{
throw new Exception("Bad serialized file: " + lexFile);
}
@in.Close();
}
}
if (argMap.Contains("-test"))
{
bool segmentWords = ctpp.segment;
bool parse = lp != null;
System.Diagnostics.Debug.Assert((parse || segmentWords));
// WordCatConstituent.collinizeWords = argMap.containsKey("-collinizeWords");
// WordCatConstituent.collinizeTags = argMap.containsKey("-collinizeTags");
IWordSegmenter seg = null;
if (segmentWords)
{
seg = (IWordSegmenter)lex;
}
string[] testArgs = (argMap["-test"]);
MemoryTreebank testTreebank = op.tlpParams.MemoryTreebank();
IFileFilter testFilt = new NumberRangesFileFilter(testArgs[1], false);
testTreebank.LoadPath(new File(testArgs[0]), testFilt);
ITreeTransformer subcategoryStripper = op.tlpParams.SubcategoryStripper();
ITreeTransformer collinizer = ctpp.Collinizer();
WordCatEquivalenceClasser eqclass = new WordCatEquivalenceClasser();
WordCatEqualityChecker eqcheck = new WordCatEqualityChecker();
EquivalenceClassEval basicEval = new EquivalenceClassEval(eqclass, eqcheck, "basic");
EquivalenceClassEval collinsEval = new EquivalenceClassEval(eqclass, eqcheck, "collinized");
IList <string> evalTypes = new List <string>(3);
bool goodPOS = false;
if (segmentWords)
{
evalTypes.Add(WordCatConstituent.wordType);
if (ctpp.segmentMarkov && !parse)
{
evalTypes.Add(WordCatConstituent.tagType);
goodPOS = true;
}
}
if (parse)
{
evalTypes.Add(WordCatConstituent.tagType);
evalTypes.Add(WordCatConstituent.catType);
if (combo)
{
evalTypes.Add(WordCatConstituent.wordType);
goodPOS = true;
}
}
TreeToBracketProcessor proc = new TreeToBracketProcessor(evalTypes);
log.Info("Testing...");
foreach (Tree goldTop in testTreebank)
{
Tree gold = goldTop.FirstChild();
IList <IHasWord> goldSentence = gold.YieldHasWord();
if (goldSentence.Count > maxLength)
{
log.Info("Skipping sentence; too long: " + goldSentence.Count);
continue;
}
else
{
log.Info("Processing sentence; length: " + goldSentence.Count);
}
IList <IHasWord> s;
if (segmentWords)
{
StringBuilder goldCharBuf = new StringBuilder();
foreach (IHasWord aGoldSentence in goldSentence)
{
StringLabel word = (StringLabel)aGoldSentence;
goldCharBuf.Append(word.Value());
}
string goldChars = goldCharBuf.ToString();
s = seg.Segment(goldChars);
}
else
{
s = goldSentence;
}
Tree tree;
if (parse)
{
tree = lp.ParseTree(s);
if (tree == null)
{
throw new Exception("PARSER RETURNED NULL!!!");
}
}
else
{
tree = Edu.Stanford.Nlp.Trees.Trees.ToFlatTree(s);
tree = subcategoryStripper.TransformTree(tree);
}
if (pw != null)
{
if (parse)
{
tree.PennPrint(pw);
}
else
{
IEnumerator sentIter = s.GetEnumerator();
for (; ;)
{
Word word = (Word)sentIter.Current;
pw.Print(word.Word());
if (sentIter.MoveNext())
{
pw.Print(" ");
}
else
{
break;
}
}
}
pw.Println();
}
if (eval)
{
ICollection ourBrackets;
ICollection goldBrackets;
ourBrackets = proc.AllBrackets(tree);
goldBrackets = proc.AllBrackets(gold);
if (goodPOS)
{
Sharpen.Collections.AddAll(ourBrackets, TreeToBracketProcessor.CommonWordTagTypeBrackets(tree, gold));
Sharpen.Collections.AddAll(goldBrackets, TreeToBracketProcessor.CommonWordTagTypeBrackets(gold, tree));
}
basicEval.Eval(ourBrackets, goldBrackets);
System.Console.Out.WriteLine("\nScores:");
basicEval.DisplayLast();
Tree collinsTree = collinizer.TransformTree(tree);
Tree collinsGold = collinizer.TransformTree(gold);
ourBrackets = proc.AllBrackets(collinsTree);
goldBrackets = proc.AllBrackets(collinsGold);
if (goodPOS)
{
Sharpen.Collections.AddAll(ourBrackets, TreeToBracketProcessor.CommonWordTagTypeBrackets(collinsTree, collinsGold));
Sharpen.Collections.AddAll(goldBrackets, TreeToBracketProcessor.CommonWordTagTypeBrackets(collinsGold, collinsTree));
}
collinsEval.Eval(ourBrackets, goldBrackets);
System.Console.Out.WriteLine("\nCollinized scores:");
collinsEval.DisplayLast();
System.Console.Out.WriteLine();
}
}
if (eval)
{
basicEval.Display();
System.Console.Out.WriteLine();
collinsEval.Display();
}
}
}
public DVParser(DVModel model, LexicalizedParser parser)
{
this.parser = parser;
this.op = parser.GetOp();
this.dvModel = model;
}
/// <summary>This example shows a few more ways of providing input to a parser.</summary>
/// <remarks>
/// This example shows a few more ways of providing input to a parser.
/// Usage: ParserDemo2 [grammar [textFile]]
/// </remarks>
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
string grammar = args.Length > 0 ? args[0] : "edu/stanford/nlp/models/lexparser/englishPCFG.ser.gz";
string[] options = new string[] { "-maxLength", "80", "-retainTmpSubcategories" };
LexicalizedParser lp = ((LexicalizedParser)LexicalizedParser.LoadModel(grammar, options));
ITreebankLanguagePack tlp = lp.GetOp().Langpack();
IGrammaticalStructureFactory gsf = tlp.GrammaticalStructureFactory();
IEnumerable <IList <IHasWord> > sentences;
if (args.Length > 1)
{
DocumentPreprocessor dp = new DocumentPreprocessor(args[1]);
IList <IList <IHasWord> > tmp = new List <IList <IHasWord> >();
foreach (IList <IHasWord> sentence in dp)
{
tmp.Add(sentence);
}
sentences = tmp;
}
else
{
// Showing tokenization and parsing in code a couple of different ways.
string[] sent = new string[] { "This", "is", "an", "easy", "sentence", "." };
IList <IHasWord> sentence = new List <IHasWord>();
foreach (string word in sent)
{
sentence.Add(new Word(word));
}
string sent2 = ("This is a slightly longer and more complex " + "sentence requiring tokenization.");
// Use the default tokenizer for this TreebankLanguagePack
ITokenizer <IHasWord> toke = tlp.GetTokenizerFactory().GetTokenizer(new StringReader(sent2));
IList <IHasWord> sentence2 = toke.Tokenize();
string[] sent3 = new string[] { "It", "can", "can", "it", "." };
string[] tag3 = new string[] { "PRP", "MD", "VB", "PRP", "." };
// Parser gets second "can" wrong without help
IList <TaggedWord> sentence3 = new List <TaggedWord>();
for (int i = 0; i < sent3.Length; i++)
{
sentence3.Add(new TaggedWord(sent3[i], tag3[i]));
}
Tree parse = lp.Parse(sentence3);
parse.PennPrint();
IList <IList <IHasWord> > tmp = new List <IList <IHasWord> >();
tmp.Add(sentence);
tmp.Add(sentence2);
tmp.Add(sentence3);
sentences = tmp;
}
foreach (IList <IHasWord> sentence_1 in sentences)
{
Tree parse = lp.Parse(sentence_1);
parse.PennPrint();
System.Console.Out.WriteLine();
GrammaticalStructure gs = gsf.NewGrammaticalStructure(parse);
IList <TypedDependency> tdl = gs.TypedDependenciesCCprocessed();
System.Console.Out.WriteLine(tdl);
System.Console.Out.WriteLine();
System.Console.Out.WriteLine("The words of the sentence:");
foreach (ILabel lab in parse.Yield())
{
if (lab is CoreLabel)
{
System.Console.Out.WriteLine(((CoreLabel)lab).ToString(CoreLabel.OutputFormat.ValueMap));
}
else
{
System.Console.Out.WriteLine(lab);
}
}
System.Console.Out.WriteLine();
System.Console.Out.WriteLine(parse.TaggedYield());
System.Console.Out.WriteLine();
}
// This method turns the String into a single sentence using the
// default tokenizer for the TreebankLanguagePack.
string sent3_1 = "This is one last test!";
lp.Parse(sent3_1).PennPrint();
}
/// <exception cref="System.IO.IOException"/>
/// <exception cref="System.TypeLoadException"/>
public static void Main(string[] args)
{
string dvmodelFile = null;
string lexparserFile = null;
string testTreebankPath = null;
IFileFilter testTreebankFilter = null;
IList <string> unusedArgs = new List <string>();
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-lexparser"))
{
lexparserFile = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-testTreebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-testTreebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
testTreebankPath = treebankDescription.First();
testTreebankFilter = treebankDescription.Second();
}
else
{
unusedArgs.Add(args[argIndex++]);
}
}
}
log.Info("Loading lexparser from: " + lexparserFile);
string[] newArgs = Sharpen.Collections.ToArray(unusedArgs, new string[unusedArgs.Count]);
LexicalizedParser lexparser = ((LexicalizedParser)LexicalizedParser.LoadModel(lexparserFile, newArgs));
log.Info("... done");
Treebank testTreebank = null;
if (testTreebankPath != null)
{
log.Info("Reading in trees from " + testTreebankPath);
if (testTreebankFilter != null)
{
log.Info("Filtering on " + testTreebankFilter);
}
testTreebank = lexparser.GetOp().tlpParams.MemoryTreebank();
testTreebank.LoadPath(testTreebankPath, testTreebankFilter);
log.Info("Read in " + testTreebank.Count + " trees for testing");
}
double[] labelResults = new double[weights.Length];
double[] tagResults = new double[weights.Length];
for (int i = 0; i < weights.Length; ++i)
{
lexparser.GetOp().baseParserWeight = weights[i];
EvaluateTreebank evaluator = new EvaluateTreebank(lexparser);
evaluator.TestOnTreebank(testTreebank);
labelResults[i] = evaluator.GetLBScore();
tagResults[i] = evaluator.GetTagScore();
}
for (int i_1 = 0; i_1 < weights.Length; ++i_1)
{
log.Info("LexicalizedParser weight " + weights[i_1] + ": labeled " + labelResults[i_1] + " tag " + tagResults[i_1]);
}
}
/// <exception cref="System.IO.IOException"/>
/// <exception cref="System.TypeLoadException"/>
public static void Main(string[] args)
{
string modelPath = null;
IList <string> baseModelPaths = null;
string testTreebankPath = null;
IFileFilter testTreebankFilter = null;
IList <string> unusedArgs = new List <string>();
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-model"))
{
modelPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-testTreebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-testTreebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
testTreebankPath = treebankDescription.First();
testTreebankFilter = treebankDescription.Second();
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-baseModels"))
{
argIndex++;
baseModelPaths = new List <string>();
while (argIndex < args.Length && args[argIndex][0] != '-')
{
baseModelPaths.Add(args[argIndex++]);
}
if (baseModelPaths.Count == 0)
{
throw new ArgumentException("Found an argument -baseModels with no actual models named");
}
}
else
{
unusedArgs.Add(args[argIndex++]);
}
}
}
}
string[] newArgs = Sharpen.Collections.ToArray(unusedArgs, new string[unusedArgs.Count]);
LexicalizedParser underlyingParser = null;
Options options = null;
LexicalizedParser combinedParser = null;
if (baseModelPaths != null)
{
IList <DVModel> dvparsers = new List <DVModel>();
foreach (string baseModelPath in baseModelPaths)
{
log.Info("Loading serialized DVParser from " + baseModelPath);
LexicalizedParser dvparser = ((LexicalizedParser)LexicalizedParser.LoadModel(baseModelPath));
IReranker reranker = dvparser.reranker;
if (!(reranker is DVModelReranker))
{
throw new ArgumentException("Expected parsers with DVModel embedded");
}
dvparsers.Add(((DVModelReranker)reranker).GetModel());
if (underlyingParser == null)
{
underlyingParser = dvparser;
options = underlyingParser.GetOp();
// TODO: other parser's options?
options.SetOptions(newArgs);
}
log.Info("... done");
}
combinedParser = LexicalizedParser.CopyLexicalizedParser(underlyingParser);
CombinedDVModelReranker reranker_1 = new CombinedDVModelReranker(options, dvparsers);
combinedParser.reranker = reranker_1;
combinedParser.SaveParserToSerialized(modelPath);
}
else
{
throw new ArgumentException("Need to specify -model to load an already prepared CombinedParser");
}
Treebank testTreebank = null;
if (testTreebankPath != null)
{
log.Info("Reading in trees from " + testTreebankPath);
if (testTreebankFilter != null)
{
log.Info("Filtering on " + testTreebankFilter);
}
testTreebank = combinedParser.GetOp().tlpParams.MemoryTreebank();
testTreebank.LoadPath(testTreebankPath, testTreebankFilter);
log.Info("Read in " + testTreebank.Count + " trees for testing");
EvaluateTreebank evaluator = new EvaluateTreebank(combinedParser.GetOp(), null, combinedParser);
evaluator.TestOnTreebank(testTreebank);
}
}
/// <summary>
/// An example of a command line is
/// <br />
/// java -mx1g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model /scr/horatio/dvparser/wsjPCFG.nocompact.simple.ser.gz -output cached9.simple.ser.gz -treebank /afs/ir/data/linguistic-data/Treebank/3/parsed/mrg/wsj 200-202
/// <br />
/// java -mx4g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model ~/scr/dvparser/wsjPCFG.nocompact.simple.ser.gz -output cached.train.simple.ser.gz -treebank /afs/ir/data/linguistic-data/Treebank/3/parsed/mrg/wsj 200-2199 -numThreads 6
/// <br />
/// java -mx4g edu.stanford.nlp.parser.dvparser.CacheParseHypotheses -model ~/scr/dvparser/chinese/xinhuaPCFG.ser.gz -output cached.xinhua.train.ser.gz -treebank /afs/ir/data/linguistic-data/Chinese-Treebank/6/data/utf8/bracketed 026-270,301-499,600-999
/// </summary>
/// <exception cref="System.IO.IOException"/>
public static void Main(string[] args)
{
string parserModel = null;
string output = null;
IList <Pair <string, IFileFilter> > treebanks = Generics.NewArrayList();
int dvKBest = 200;
int numThreads = 1;
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-dvKBest"))
{
dvKBest = System.Convert.ToInt32(args[argIndex + 1]);
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-parser") || args[argIndex].Equals("-model"))
{
parserModel = args[argIndex + 1];
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-output"))
{
output = args[argIndex + 1];
argIndex += 2;
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-treebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-treebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
treebanks.Add(treebankDescription);
continue;
}
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-numThreads"))
{
numThreads = System.Convert.ToInt32(args[argIndex + 1]);
argIndex += 2;
continue;
}
throw new ArgumentException("Unknown argument " + args[argIndex]);
}
if (parserModel == null)
{
throw new ArgumentException("Need to supply a parser model with -model");
}
if (output == null)
{
throw new ArgumentException("Need to supply an output filename with -output");
}
if (treebanks.IsEmpty())
{
throw new ArgumentException("Need to supply a treebank with -treebank");
}
log.Info("Writing output to " + output);
log.Info("Loading parser model " + parserModel);
log.Info("Writing " + dvKBest + " hypothesis trees for each tree");
LexicalizedParser parser = ((LexicalizedParser)LexicalizedParser.LoadModel(parserModel, "-dvKBest", int.ToString(dvKBest)));
CacheParseHypotheses cacher = new CacheParseHypotheses(parser);
ITreeTransformer transformer = DVParser.BuildTrainTransformer(parser.GetOp());
IList <Tree> sentences = new List <Tree>();
foreach (Pair <string, IFileFilter> description in treebanks)
{
log.Info("Reading trees from " + description.first);
Treebank treebank = parser.GetOp().tlpParams.MemoryTreebank();
treebank.LoadPath(description.first, description.second);
treebank = treebank.Transform(transformer);
Sharpen.Collections.AddAll(sentences, treebank);
}
log.Info("Processing " + sentences.Count + " trees");
IList <Pair <Tree, byte[]> > cache = Generics.NewArrayList();
transformer = new SynchronizedTreeTransformer(transformer);
MulticoreWrapper <Tree, Pair <Tree, byte[]> > wrapper = new MulticoreWrapper <Tree, Pair <Tree, byte[]> >(numThreads, new CacheParseHypotheses.CacheProcessor(cacher, parser, dvKBest, transformer));
foreach (Tree tree in sentences)
{
wrapper.Put(tree);
while (wrapper.Peek())
{
cache.Add(wrapper.Poll());
if (cache.Count % 10 == 0)
{
System.Console.Out.WriteLine("Processed " + cache.Count + " trees");
}
}
}
wrapper.Join();
while (wrapper.Peek())
{
cache.Add(wrapper.Poll());
if (cache.Count % 10 == 0)
{
System.Console.Out.WriteLine("Processed " + cache.Count + " trees");
}
}
System.Console.Out.WriteLine("Finished processing " + cache.Count + " trees");
IOUtils.WriteObjectToFile(cache, output);
}
public EvaluateTreebank(LexicalizedParser parser)
: this(parser.GetOp(), parser.lex, parser)
{
}
/// <summary>
/// Command line arguments for this program:
/// <br />
/// -output: the model file to output
/// -input: a list of model files to input
/// </summary>
public static void Main(string[] args)
{
string outputModelFilename = null;
IList <string> inputModelFilenames = Generics.NewArrayList();
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-output"))
{
outputModelFilename = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-input"))
{
for (++argIndex; argIndex < args.Length && !args[argIndex].StartsWith("-"); ++argIndex)
{
Sharpen.Collections.AddAll(inputModelFilenames, Arrays.AsList(args[argIndex].Split(",")));
}
}
else
{
throw new Exception("Unknown argument " + args[argIndex]);
}
}
}
if (outputModelFilename == null)
{
log.Info("Need to specify output model name with -output");
System.Environment.Exit(2);
}
if (inputModelFilenames.Count == 0)
{
log.Info("Need to specify input model names with -input");
System.Environment.Exit(2);
}
log.Info("Averaging " + inputModelFilenames);
log.Info("Outputting result to " + outputModelFilename);
LexicalizedParser lexparser = null;
IList <DVModel> models = Generics.NewArrayList();
foreach (string filename in inputModelFilenames)
{
LexicalizedParser parser = ((LexicalizedParser)LexicalizedParser.LoadModel(filename));
if (lexparser == null)
{
lexparser = parser;
}
models.Add(DVParser.GetModelFromLexicalizedParser(parser));
}
IList <TwoDimensionalMap <string, string, SimpleMatrix> > binaryTransformMaps = CollectionUtils.TransformAsList(models, null);
IList <TwoDimensionalMap <string, string, SimpleMatrix> > binaryScoreMaps = CollectionUtils.TransformAsList(models, null);
IList <IDictionary <string, SimpleMatrix> > unaryTransformMaps = CollectionUtils.TransformAsList(models, null);
IList <IDictionary <string, SimpleMatrix> > unaryScoreMaps = CollectionUtils.TransformAsList(models, null);
IList <IDictionary <string, SimpleMatrix> > wordMaps = CollectionUtils.TransformAsList(models, null);
TwoDimensionalMap <string, string, SimpleMatrix> binaryTransformAverages = AverageBinaryMatrices(binaryTransformMaps);
TwoDimensionalMap <string, string, SimpleMatrix> binaryScoreAverages = AverageBinaryMatrices(binaryScoreMaps);
IDictionary <string, SimpleMatrix> unaryTransformAverages = AverageUnaryMatrices(unaryTransformMaps);
IDictionary <string, SimpleMatrix> unaryScoreAverages = AverageUnaryMatrices(unaryScoreMaps);
IDictionary <string, SimpleMatrix> wordAverages = AverageUnaryMatrices(wordMaps);
DVModel newModel = new DVModel(binaryTransformAverages, unaryTransformAverages, binaryScoreAverages, unaryScoreAverages, wordAverages, lexparser.GetOp());
DVParser newParser = new DVParser(newModel, lexparser);
newParser.SaveModel(outputModelFilename);
}
internal LexicalizedParserQuery(LexicalizedParser parser)
{
this.op = parser.GetOp();
BinaryGrammar bg = parser.bg;
UnaryGrammar ug = parser.ug;
ILexicon lex = parser.lex;
IDependencyGrammar dg = parser.dg;
IIndex <string> stateIndex = parser.stateIndex;
IIndex <string> wordIndex = new DeltaIndex <string>(parser.wordIndex);
IIndex <string> tagIndex = parser.tagIndex;
this.debinarizer = new Debinarizer(op.forceCNF);
this.boundaryRemover = new BoundaryRemover();
if (op.doPCFG)
{
if (op.testOptions.iterativeCKY)
{
pparser = new IterativeCKYPCFGParser(bg, ug, lex, op, stateIndex, wordIndex, tagIndex);
}
else
{
pparser = new ExhaustivePCFGParser(bg, ug, lex, op, stateIndex, wordIndex, tagIndex);
}
}
else
{
pparser = null;
}
if (op.doDep)
{
dg.SetLexicon(lex);
if (!op.testOptions.useFastFactored)
{
dparser = new ExhaustiveDependencyParser(dg, lex, op, wordIndex, tagIndex);
}
else
{
dparser = null;
}
}
else
{
dparser = null;
}
if (op.doDep && op.doPCFG)
{
if (op.testOptions.useFastFactored)
{
MLEDependencyGrammar mledg = (MLEDependencyGrammar)dg;
int numToFind = 1;
if (op.testOptions.printFactoredKGood > 0)
{
numToFind = op.testOptions.printFactoredKGood;
}
bparser = new FastFactoredParser(pparser, mledg, op, numToFind, wordIndex, tagIndex);
}
else
{
IScorer scorer = new TwinScorer(pparser, dparser);
//Scorer scorer = parser;
if (op.testOptions.useN5)
{
bparser = new BiLexPCFGParser.N5BiLexPCFGParser(scorer, pparser, dparser, bg, ug, dg, lex, op, stateIndex, wordIndex, tagIndex);
}
else
{
bparser = new BiLexPCFGParser(scorer, pparser, dparser, bg, ug, dg, lex, op, stateIndex, wordIndex, tagIndex);
}
}
}
else
{
bparser = null;
}
subcategoryStripper = op.tlpParams.SubcategoryStripper();
}
/// <exception cref="System.Exception"/>
public static void Main(string[] args)
{
string modelPath = null;
string outputPath = null;
string testTreebankPath = null;
IFileFilter testTreebankFilter = null;
IList <string> unusedArgs = new List <string>();
for (int argIndex = 0; argIndex < args.Length;)
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-model"))
{
modelPath = args[argIndex + 1];
argIndex += 2;
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-testTreebank"))
{
Pair <string, IFileFilter> treebankDescription = ArgUtils.GetTreebankDescription(args, argIndex, "-testTreebank");
argIndex = argIndex + ArgUtils.NumSubArgs(args, argIndex) + 1;
testTreebankPath = treebankDescription.First();
testTreebankFilter = treebankDescription.Second();
}
else
{
if (Sharpen.Runtime.EqualsIgnoreCase(args[argIndex], "-output"))
{
outputPath = args[argIndex + 1];
argIndex += 2;
}
else
{
unusedArgs.Add(args[argIndex++]);
}
}
}
}
if (modelPath == null)
{
throw new ArgumentException("Need to specify -model");
}
if (testTreebankPath == null)
{
throw new ArgumentException("Need to specify -testTreebank");
}
if (outputPath == null)
{
throw new ArgumentException("Need to specify -output");
}
string[] newArgs = Sharpen.Collections.ToArray(unusedArgs, new string[unusedArgs.Count]);
LexicalizedParser lexparser = ((LexicalizedParser)LexicalizedParser.LoadModel(modelPath, newArgs));
Treebank testTreebank = null;
if (testTreebankPath != null)
{
log.Info("Reading in trees from " + testTreebankPath);
if (testTreebankFilter != null)
{
log.Info("Filtering on " + testTreebankFilter);
}
testTreebank = lexparser.GetOp().tlpParams.MemoryTreebank();
testTreebank.LoadPath(testTreebankPath, testTreebankFilter);
log.Info("Read in " + testTreebank.Count + " trees for testing");
}
FileWriter @out = new FileWriter(outputPath);
BufferedWriter bout = new BufferedWriter(@out);
log.Info("Parsing " + testTreebank.Count + " trees");
int count = 0;
IList <FindNearestNeighbors.ParseRecord> records = Generics.NewArrayList();
foreach (Tree goldTree in testTreebank)
{
IList <Word> tokens = goldTree.YieldWords();
IParserQuery parserQuery = lexparser.ParserQuery();
if (!parserQuery.Parse(tokens))
{
throw new AssertionError("Could not parse: " + tokens);
}
if (!(parserQuery is RerankingParserQuery))
{
throw new ArgumentException("Expected a LexicalizedParser with a Reranker attached");
}
RerankingParserQuery rpq = (RerankingParserQuery)parserQuery;
if (!(rpq.RerankerQuery() is DVModelReranker.Query))
{
throw new ArgumentException("Expected a LexicalizedParser with a DVModel attached");
}
DeepTree tree = ((DVModelReranker.Query)rpq.RerankerQuery()).GetDeepTrees()[0];
SimpleMatrix rootVector = null;
foreach (KeyValuePair <Tree, SimpleMatrix> entry in tree.GetVectors())
{
if (entry.Key.Label().Value().Equals("ROOT"))
{
rootVector = entry.Value;
break;
}
}
if (rootVector == null)
{
throw new AssertionError("Could not find root nodevector");
}
@out.Write(tokens + "\n");
@out.Write(tree.GetTree() + "\n");
for (int i = 0; i < rootVector.GetNumElements(); ++i)
{
@out.Write(" " + rootVector.Get(i));
}
@out.Write("\n\n\n");
count++;
if (count % 10 == 0)
{
log.Info(" " + count);
}
records.Add(new FindNearestNeighbors.ParseRecord(tokens, goldTree, tree.GetTree(), rootVector, tree.GetVectors()));
}
log.Info(" done parsing");
IList <Pair <Tree, SimpleMatrix> > subtrees = Generics.NewArrayList();
foreach (FindNearestNeighbors.ParseRecord record in records)
{
foreach (KeyValuePair <Tree, SimpleMatrix> entry in record.nodeVectors)
{
if (entry.Key.GetLeaves().Count <= maxLength)
{
subtrees.Add(Pair.MakePair(entry.Key, entry.Value));
}
}
}
log.Info("There are " + subtrees.Count + " subtrees in the set of trees");
PriorityQueue <ScoredObject <Pair <Tree, Tree> > > bestmatches = new PriorityQueue <ScoredObject <Pair <Tree, Tree> > >(101, ScoredComparator.DescendingComparator);
for (int i_1 = 0; i_1 < subtrees.Count; ++i_1)
{
log.Info(subtrees[i_1].First().YieldWords());
log.Info(subtrees[i_1].First());
for (int j = 0; j < subtrees.Count; ++j)
{
if (i_1 == j)
{
continue;
}
// TODO: look at basic category?
double normF = subtrees[i_1].Second().Minus(subtrees[j].Second()).NormF();
bestmatches.Add(new ScoredObject <Pair <Tree, Tree> >(Pair.MakePair(subtrees[i_1].First(), subtrees[j].First()), normF));
if (bestmatches.Count > 100)
{
bestmatches.Poll();
}
}
IList <ScoredObject <Pair <Tree, Tree> > > ordered = Generics.NewArrayList();
while (bestmatches.Count > 0)
{
ordered.Add(bestmatches.Poll());
}
Java.Util.Collections.Reverse(ordered);
foreach (ScoredObject <Pair <Tree, Tree> > pair in ordered)
{
log.Info(" MATCHED " + pair.Object().second.YieldWords() + " ... " + pair.Object().Second() + " with a score of " + pair.Score());
}
log.Info();
log.Info();
bestmatches.Clear();
}
/*
* for (int i = 0; i < records.size(); ++i) {
* if (i % 10 == 0) {
* log.info(" " + i);
* }
* List<ScoredObject<ParseRecord>> scored = Generics.newArrayList();
* for (int j = 0; j < records.size(); ++j) {
* if (i == j) continue;
*
* double score = 0.0;
* int matches = 0;
* for (Map.Entry<Tree, SimpleMatrix> first : records.get(i).nodeVectors.entrySet()) {
* for (Map.Entry<Tree, SimpleMatrix> second : records.get(j).nodeVectors.entrySet()) {
* String firstBasic = dvparser.dvModel.basicCategory(first.getKey().label().value());
* String secondBasic = dvparser.dvModel.basicCategory(second.getKey().label().value());
* if (firstBasic.equals(secondBasic)) {
++matches;
* double normF = first.getValue().minus(second.getValue()).normF();
* score += normF * normF;
* }
* }
* }
* if (matches == 0) {
* score = Double.POSITIVE_INFINITY;
* } else {
* score = score / matches;
* }
* //double score = records.get(i).vector.minus(records.get(j).vector).normF();
* scored.add(new ScoredObject<ParseRecord>(records.get(j), score));
* }
* Collections.sort(scored, ScoredComparator.ASCENDING_COMPARATOR);
*
* out.write(records.get(i).sentence.toString() + "\n");
* for (int j = 0; j < numNeighbors; ++j) {
* out.write(" " + scored.get(j).score() + ": " + scored.get(j).object().sentence + "\n");
* }
* out.write("\n\n");
* }
* log.info();
*/
bout.Flush();
@out.Flush();
@out.Close();
}