|
| 1 | +/* |
| 2 | + * Licensed to the Apache Software Foundation (ASF) under one or more |
| 3 | + * contributor license agreements. See the NOTICE file distributed with |
| 4 | + * this work for additional information regarding copyright ownership. |
| 5 | + * The ASF licenses this file to You under the Apache License, Version 2.0 |
| 6 | + * (the "License"); you may not use this file except in compliance with |
| 7 | + * the License. You may obtain a copy of the License at |
| 8 | + * |
| 9 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 10 | + * |
| 11 | + * Unless required by applicable law or agreed to in writing, software |
| 12 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 13 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 14 | + * See the License for the specific language governing permissions and |
| 15 | + * limitations under the License. |
| 16 | + */ |
| 17 | + |
| 18 | +package opennlp.tools.postag; |
| 19 | + |
| 20 | +import java.io.IOException; |
| 21 | +import java.util.ArrayList; |
| 22 | +import java.util.HashMap; |
| 23 | +import java.util.List; |
| 24 | +import java.util.Map; |
| 25 | +import java.util.stream.Stream; |
| 26 | + |
| 27 | +import org.junit.jupiter.api.BeforeAll; |
| 28 | +import org.junit.jupiter.params.ParameterizedTest; |
| 29 | +import org.junit.jupiter.params.provider.Arguments; |
| 30 | +import org.junit.jupiter.params.provider.MethodSource; |
| 31 | + |
| 32 | +import opennlp.tools.tokenize.ThreadSafeTokenizerME; |
| 33 | +import opennlp.tools.tokenize.Tokenizer; |
| 34 | + |
| 35 | +import static org.junit.jupiter.api.Assertions.assertEquals; |
| 36 | +import static org.junit.jupiter.api.Assertions.assertNotNull; |
| 37 | +import static org.junit.jupiter.api.Assertions.assertTrue; |
| 38 | + |
| 39 | +public class POSTaggerMEIT { |
| 40 | + |
| 41 | + private static final String CATALAN = "ca"; |
| 42 | + private static final String ENGLISH = "en"; |
| 43 | + private static final String GERMAN = "de"; |
| 44 | + private static final String POLISH = "pl"; |
| 45 | + private static final String PORTUGUESE = "pt"; |
| 46 | + |
| 47 | + private static final Map<String, Tokenizer> TOKENIZERS = new HashMap<>(); |
| 48 | + private static final Map<String, POSTagger> TAGGERS = new HashMap<>(); |
| 49 | + |
| 50 | + private static final boolean debug = false; |
| 51 | + |
| 52 | + @BeforeAll |
| 53 | + public static void initResources() throws IOException { |
| 54 | + List<String> langs = List.of(CATALAN, ENGLISH, GERMAN, POLISH, PORTUGUESE); |
| 55 | + for (String langCode: langs) { |
| 56 | + TOKENIZERS.put(langCode, new ThreadSafeTokenizerME(langCode)); |
| 57 | + TAGGERS.put(langCode, new ThreadSafePOSTaggerME(langCode)); |
| 58 | + } |
| 59 | + } |
| 60 | + |
| 61 | + @ParameterizedTest(name = "Verify \"{0}\" sample") |
| 62 | + @MethodSource(value = "provideData") |
| 63 | + void testPOSTagger(String langCode, int allowedDelta, String input, String[] expectedTags) { |
| 64 | + |
| 65 | + final String[] tokens = TOKENIZERS.get(langCode).tokenize(input); |
| 66 | + assertNotNull(tokens); |
| 67 | + assertEquals(expectedTags.length, tokens.length); |
| 68 | + final String[] tags = TAGGERS.get(langCode).tag(tokens); |
| 69 | + assertNotNull(tags); |
| 70 | + assertEquals(expectedTags.length, tags.length); |
| 71 | + StringBuilder fullyTagged = new StringBuilder(); |
| 72 | + for (int i = 0; i < tags.length; i++) { |
| 73 | + fullyTagged.append(tokens[i]).append("_").append(tags[i]).append(" "); |
| 74 | + } |
| 75 | + if (debug) { |
| 76 | + System.out.println(fullyTagged); |
| 77 | + } |
| 78 | + |
| 79 | + List<Integer> incorrectTagsPositions = new ArrayList<>(); |
| 80 | + for (int i = 0; i < tags.length; i++) { |
| 81 | + StringBuilder sb = new StringBuilder(); |
| 82 | + sb.append(tokens[i]).append("[").append(tags[i]).append("]"); |
| 83 | + if (expectedTags[i].equals(tags[i])) { |
| 84 | + sb.append(" <-- " + "OK"); |
| 85 | + } else { |
| 86 | + sb.append(" <-- " + "NOK" + ", pos=").append(i); |
| 87 | + incorrectTagsPositions.add(i); |
| 88 | + } |
| 89 | + if (debug) { |
| 90 | + System.out.println(sb); |
| 91 | + } |
| 92 | + } |
| 93 | + assertTrue(incorrectTagsPositions.size() <= allowedDelta); |
| 94 | + } |
| 95 | + |
| 96 | + private static Stream<Arguments> provideData() { |
| 97 | + return Stream.of( |
| 98 | + // see: Dev Manual |
| 99 | + Arguments.of(ENGLISH, 0, |
| 100 | + "Mr. Vinken is chairman of Elsevier N.V. , the Dutch publishing group .", |
| 101 | + new String[]{"PROPN", "PROPN", "AUX", "NOUN", "ADP", "ADJ", "PROPN", "PUNCT", "DET", "PROPN", |
| 102 | + "VERB", "NOUN", "PUNCT"}), |
| 103 | + // see: 'de-ud-train-sample.conllu' |
| 104 | + Arguments.of(GERMAN, 0, |
| 105 | + "Fachlich kompetent, sehr gute Beratung und ein freundliches Team .", |
| 106 | + new String[]{"ADJ", "ADJ", "PUNCT", "ADV", "ADJ", "NOUN", "CCONJ", "DET", "ADJ", "NOUN", "PUNCT"}), |
| 107 | + // see: 'pt-br-ud-sample.conllu' |
| 108 | + Arguments.of(PORTUGUESE, 1, |
| 109 | + "Numa reunião entre representantes da Secretaria da Criança do DF ea juíza da Vara de Execuções de " + |
| 110 | + "Medidas Socioeducativas, Lavínia Tupi Vieira Fonseca, ficou acordado que dos 25 internos, " + |
| 111 | + "12 serão internados na Unidade de Planaltina e os outros 13 devem retornar para a Unidade do " + |
| 112 | + "Recanto das Emas, antigo Ciago .", |
| 113 | + // pos=10 -> NOK |
| 114 | + new String[]{"ADP+DET", "NOUN", "ADP", "NOUN", "ADP+DET", "PROPN", "ADP+DET", "PROPN", "ADP+DET", |
| 115 | + "PROPN", "CCONJ", "NOUN", "ADP+DET", "PROPN", "ADP", "PROPN", "ADP", "PROPN", "PROPN", "PUNCT", |
| 116 | + "PROPN", "PROPN", "PROPN", "PROPN", "PUNCT", "VERB", "ADJ", "CCONJ", "ADP+DET", "NUM", "NOUN", |
| 117 | + "PUNCT", "NUM", "AUX", "VERB", "ADP+DET", "PROPN", "ADP", "PROPN", "CCONJ", "DET", "DET", "NUM", |
| 118 | + "AUX", "VERB", "ADP", "DET", "PROPN", "ADP+DET", "PROPN", "ADP+DET", "PROPN", "PUNCT", "ADJ", |
| 119 | + "PROPN", "PUNCT"}), |
| 120 | + // via @alsmolarczyk, original by Lem, Stanisław (1961/2022): |
| 121 | + // Solaris, Wydawnictwo Literackie, Kraków, S. 81. |
| 122 | + Arguments.of(POLISH, 1, |
| 123 | + "Zerwałem się ze stosu zwiniętych spadochronów i pobiegłem prosto do radiostacji .", |
| 124 | + new String[]{"VERB+AUX", "PART", "ADP", "NOUN", "ADJ", "NOUN", "CCONJ", "VERB+AUX", "ADV", "ADP", |
| 125 | + "NOUN", "PUNCT"}), |
| 126 | + // via @alsmolarczyk, original by Tokarczuk, Olga (2009/2021): |
| 127 | + // Prowadź swój pług przez kości umarłych, Wydawnictwo Literackie, Kraków, S. 43-44. |
| 128 | + Arguments.of(POLISH, 0, |
| 129 | + "Więzienie nie tkwi na zewnątrz, ale jest w środku każdego z nas .", |
| 130 | + new String[]{"NOUN", "PART", "VERB", "ADP", "ADV", "PUNCT", "CCONJ", "VERB", "ADP", "NOUN", |
| 131 | + "DET", "ADP", "PRON", "PUNCT"}), |
| 132 | + // via @alsmolarczyk, original by Zalega, Dariusz (2019): |
| 133 | + // Śląsk zbuntowany, Wydawnictwo Czarne, Wołowiec, S. 96. |
| 134 | + Arguments.of(POLISH, 0, |
| 135 | + "Działacze stosowali też różne formy nacisku na polski konsulat , żeby zaopiekował się " + |
| 136 | + "bezrobotnymi z Polski albo dał im choćby na bezpłatny bilet do kraju .", |
| 137 | + new String[]{"NOUN", "VERB", "PART", "ADJ", "NOUN", "NOUN", "ADP", "ADJ", "NOUN", "PUNCT", "SCONJ", |
| 138 | + "VERB", "PRON", "ADJ", "ADP", "PROPN", "CCONJ", "VERB", "PRON", "PART", "ADP", "ADJ", "NOUN", |
| 139 | + "ADP", "NOUN", "PUNCT"}), |
| 140 | + // via: @kinow |
| 141 | + Arguments.of(CATALAN, 1, |
| 142 | + "Un gran embossament d'aire fred es comença a despenjar cap al centre d'Europa.", |
| 143 | + // OpenNLP, different at: idx pos 2, 3, 5, and 13(+14) -> however, only pos 5 is "wrong" (ref) |
| 144 | + new String[]{"DET", "ADJ", "NOUN", "ADP", "NOUN", "ADJ", "PRON", "VERB", "ADP", "VERB", "NOUN", |
| 145 | + "ADP+DET", "NOUN", "ADP", "PROPN", "PUNCT"}) |
| 146 | + // REFERENCE ("gold"): |
| 147 | + // "DET", "ADJ", "NOUN", "ADP", "NOUN", "ADJ", "PRON", "VERB", "ADP", "VERB", "NOUN", "ADP+DET", |
| 148 | + // "NOUN", "ADP", "PROPN", "PUNCT"}) |
| 149 | + |
| 150 | + // Spacy, wrong tags at: idx pos 2, 3 and 14 |
| 151 | + //"DET", "ADJ", "ADV", "PROPN", "NOUN", "ADJ", "PRON", "VERB", "ADP", "VERB", "NOUN", "ADP" + "DET", |
| 152 | + // "NOUN", "PROPN", "PROPN", "PUNCT" |
| 153 | + // ok! , ok! , ??? , ??? , ok! , ok! , ok! , ok! , ok! , ok! , ok! , ok! + ok! , |
| 154 | + // ok! , ??? , ok! , ok! |
| 155 | + ); |
| 156 | + } |
| 157 | +} |
0 commit comments