opennlp.tools.formats.Conll02NameSampleStream Maven / Gradle / Ivy
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package opennlp.tools.formats;
import java.io.IOException;
import java.nio.charset.StandardCharsets;
import java.util.ArrayList;
import java.util.List;
import opennlp.tools.commons.Internal;
import opennlp.tools.namefind.NameSample;
import opennlp.tools.util.InputStreamFactory;
import opennlp.tools.util.InvalidFormatException;
import opennlp.tools.util.ObjectStream;
import opennlp.tools.util.PlainTextByLineStream;
import opennlp.tools.util.Span;
import opennlp.tools.util.StringUtil;
/**
* Parser for the Dutch and Spanish ner training files of the CONLL 2002 shared task.
*
* The Dutch data has a {@link #DOCSTART} tag to mark article boundaries,
* adaptive data in the feature generators will be cleared before every article.
* The Spanish data does not contain article boundaries,
* adaptive data will be cleared for every sentence.
*
* The data contains four named entity types: Person, Organization, Location and Misc.
*
* Data can be found on this
* web site.
*
* Note:
* Do not use this class, internal use only!
*/
@Internal
public class Conll02NameSampleStream implements ObjectStream {
public enum LANGUAGE {
NLD,
SPA
}
public static final int GENERATE_PERSON_ENTITIES = 0x01;
public static final int GENERATE_ORGANIZATION_ENTITIES = 0x01 << 1;
public static final int GENERATE_LOCATION_ENTITIES = 0x01 << 2;
public static final int GENERATE_MISC_ENTITIES = 0x01 << 3;
public static final String DOCSTART = "-DOCSTART-";
private final LANGUAGE lang;
private final ObjectStream lineStream;
private final int types;
/**
* Initializes a {@link Conll02NameSampleStream}.
*
* @param lang The language of the CONLL 02 data.
* @param lineStream An {@link ObjectStream} over the lines
* in the CONLL 02 data file.
* @param types The entity types to include in the Name Sample object stream.
*/
public Conll02NameSampleStream(LANGUAGE lang, ObjectStream lineStream, int types) {
this.lang = lang;
this.lineStream = lineStream;
this.types = types;
}
/**
* Initializes a {@link Conll02NameSampleStream}.
*
* @param lang The language of the CONLL 02 data.
* @param in The {@link InputStreamFactory} for the input file.
* @param types The entity types to include in the Name Sample object stream.
*
* @throws IOException Thrown if IO errors occurred.
*/
public Conll02NameSampleStream(LANGUAGE lang, InputStreamFactory in, int types) throws IOException {
/*
* NOTE: KEEP this encoding here! The original CONLL 2002 data is provided as: ISO_8859_1.
*/
this (lang, new PlainTextByLineStream(in, StandardCharsets.ISO_8859_1), types);
/*
* If related files are (incorrectly) interpreted as 'UTF-8' without prior conversion of
* the train/test files, then á, é, ñ,.. will be misinterpreted during processing and in
* resulting outcomes, e.g. produced via TokenNameFinderConverter.
*
* As a consequence, users of related tooling (OpenNLP Doc: CONLL 2002) will thus suffer
* from corrupted intermediate files, as an out-of the box experience.
*
* Details see: https://issues.apache.org/jira/browse/OPENNLP-1512
*/
}
static Span extract(int begin, int end, String beginTag) throws InvalidFormatException {
String type = beginTag.substring(2);
switch (type) {
case "PER":
type = "person";
break;
case "LOC":
type = "location";
break;
case "MISC":
type = "misc";
break;
case "ORG":
type = "organization";
break;
default:
throw new InvalidFormatException("Unknown type: " + type);
}
return new Span(begin, end, type);
}
@Override
public NameSample read() throws IOException {
List sentence = new ArrayList<>();
List tags = new ArrayList<>();
boolean isClearAdaptiveData = false;
// Empty line indicates end of sentence
String line;
while ((line = lineStream.read()) != null && !StringUtil.isEmpty(line)) {
if (LANGUAGE.NLD.equals(lang) && line.startsWith(DOCSTART)) {
isClearAdaptiveData = true;
continue;
}
String[] fields = line.split(" ");
if (fields.length == 3) {
sentence.add(fields[0]);
tags.add(fields[2]);
}
else {
throw new IOException("Expected three fields per line in training data, got " +
fields.length + " for line '" + line + "'!");
}
}
// Always clear adaptive data for spanish
if (LANGUAGE.SPA.equals(lang))
isClearAdaptiveData = true;
if (sentence.size() > 0) {
// convert name tags into spans
List names = new ArrayList<>();
int beginIndex = -1;
int endIndex = -1;
for (int i = 0; i < tags.size(); i++) {
String tag = tags.get(i);
if (tag.endsWith("PER") && (types & GENERATE_PERSON_ENTITIES) == 0)
tag = "O";
if (tag.endsWith("ORG") && (types & GENERATE_ORGANIZATION_ENTITIES) == 0)
tag = "O";
if (tag.endsWith("LOC") && (types & GENERATE_LOCATION_ENTITIES) == 0)
tag = "O";
if (tag.endsWith("MISC") && (types & GENERATE_MISC_ENTITIES) == 0)
tag = "O";
if (tag.startsWith("B-")) {
if (beginIndex != -1) {
names.add(extract(beginIndex, endIndex, tags.get(beginIndex)));
beginIndex = -1;
endIndex = -1;
}
beginIndex = i;
endIndex = i + 1;
}
else if (tag.startsWith("I-")) {
endIndex++;
}
else if (tag.equals("O")) {
if (beginIndex != -1) {
names.add(extract(beginIndex, endIndex, tags.get(beginIndex)));
beginIndex = -1;
endIndex = -1;
}
}
else {
throw new IOException("Invalid tag: " + tag);
}
}
// if one span remains, create it here
if (beginIndex != -1)
names.add(extract(beginIndex, endIndex, tags.get(beginIndex)));
return new NameSample(sentence.toArray(new String[0]),
names.toArray(new Span[0]), isClearAdaptiveData);
}
else if (line != null) {
// Just filter out empty events, if two lines in a row are empty
return read();
}
else {
// source stream is not returning anymore lines
return null;
}
}
@Override
public void reset() throws IOException, UnsupportedOperationException {
lineStream.reset();
}
@Override
public void close() throws IOException {
lineStream.close();
}
}