Package banner.eval.dataset

Class BC2GMDataset

java.lang.Object

banner.eval.dataset.Dataset

banner.eval.dataset.BC2GMDataset

public class BC2GMDataset
extends Dataset

Author:

bob

Nested Class Summary

Nested classes/interfaces inherited from class banner.eval.dataset.Dataset

Dataset.Tag

Field Summary

Fields inherited from class banner.eval.dataset.Dataset

sentences, tokenizer

Constructor Summary

Constructors

Constructor	Description
BC2GMDataset()
BC2GMDataset(Tokenizer tokenizer)

Method Summary

Modifier and Type	Method	Description
protected void	addAlternateMentions(Sentence sentence, HashMap<String,LinkedList<Dataset.Tag>> tags)
protected HashMap<String,LinkedList<Dataset.Tag>>	getAlternateTags(BufferedReader tagFile, String geneLabel)
protected Sentence	getSentence(String id, String sentenceText, Tokenizer tokenizer, HashMap<String,LinkedList<Dataset.Tag>> tags)
protected HashMap<String,LinkedList<Dataset.Tag>>	getTags(BufferedReader tagFile, String geneLabel)
protected static int	getTokenIndex(List<Token> tokens, int index)
void	load(String sentenceFilename, String mentionsFilename, String alternateMentionsFilename, String geneLabel)
void	load(org.apache.commons.configuration.HierarchicalConfiguration config)
List<Dataset>	split(int n)

Methods inherited from class banner.eval.dataset.Dataset

getSentences, getTokenCountMention, getTokenCountTotal, getTypeCounts, setTokenizer

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

BC2GMDataset

public BC2GMDataset(Tokenizer tokenizer)

BC2GMDataset

public BC2GMDataset()

Method Detail

load

public void load(org.apache.commons.configuration.HierarchicalConfiguration config)

Specified by:

load in class Dataset

load

public void load(String sentenceFilename,
                 String mentionsFilename,
                 String alternateMentionsFilename,
                 String geneLabel)

getTags

protected HashMap<String,LinkedList<Dataset.Tag>> getTags(BufferedReader tagFile,
                                                                String geneLabel)
                                                         throws IOException

Throws:

IOException

getAlternateTags

protected HashMap<String,LinkedList<Dataset.Tag>> getAlternateTags(BufferedReader tagFile,
                                                                         String geneLabel)
                                                                  throws IOException

Throws:

IOException

getSentence

protected Sentence getSentence(String id,
                               String sentenceText,
                               Tokenizer tokenizer,
                               HashMap<String,LinkedList<Dataset.Tag>> tags)

addAlternateMentions

protected void addAlternateMentions(Sentence sentence,
                                    HashMap<String,LinkedList<Dataset.Tag>> tags)

getTokenIndex

protected static int getTokenIndex(List<Token> tokens,
                                   int index)

split

public List<Dataset> split(int n)

Specified by:

split in class Dataset