PageRenderTime 59ms CodeModel.GetById 21ms RepoModel.GetById 0ms app.codeStats 0ms

/lucene/analysis/common/src/test/org/apache/lucene/analysis/hunspell/TestAllDictionaries2.java

https://github.com/mdodsworth/lucene-solr
Java | 221 lines | 182 code | 12 blank | 27 comment | 7 complexity | 73f13f448f16d74e93f375611a4b13e7 MD5 | raw file
Possible License(s): Apache-2.0
  1. package org.apache.lucene.analysis.hunspell;
  2. /*
  3. * Licensed to the Apache Software Foundation (ASF) under one or more
  4. * contributor license agreements. See the NOTICE file distributed with
  5. * this work for additional information regarding copyright ownership.
  6. * The ASF licenses this file to You under the Apache License, Version 2.0
  7. * (the "License"); you may not use this file except in compliance with
  8. * the License. You may obtain a copy of the License at
  9. *
  10. * http://www.apache.org/licenses/LICENSE-2.0
  11. *
  12. * Unless required by applicable law or agreed to in writing, software
  13. * distributed under the License is distributed on an "AS IS" BASIS,
  14. * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  15. * See the License for the specific language governing permissions and
  16. * limitations under the License.
  17. */
  18. import java.io.File;
  19. import java.io.InputStream;
  20. import java.nio.charset.StandardCharsets;
  21. import java.util.zip.ZipEntry;
  22. import java.util.zip.ZipFile;
  23. import org.apache.lucene.analysis.hunspell.Dictionary;
  24. import org.apache.lucene.util.LuceneTestCase;
  25. import org.apache.lucene.util.RamUsageTester;
  26. import org.apache.lucene.util.LuceneTestCase.SuppressSysoutChecks;
  27. import org.junit.Ignore;
  28. /**
  29. * These thunderbird dictionaries can be retrieved via:
  30. * https://addons.mozilla.org/en-US/thunderbird/language-tools/
  31. * You must click and download every file: sorry!
  32. */
  33. @Ignore("enable manually")
  34. @SuppressSysoutChecks(bugUrl = "prints important memory utilization stats per dictionary")
  35. public class TestAllDictionaries2 extends LuceneTestCase {
  36. // set this to the location of where you downloaded all the files
  37. static final File DICTIONARY_HOME =
  38. new File("/data/thunderbirdDicts");
  39. final String tests[] = {
  40. /* zip file */ /* dictionary */ /* affix */
  41. "addon-0.4.5-an+fx+tb+fn+sm.xpi", "dictionaries/ru.dic", "dictionaries/ru.aff",
  42. "addon-0.5.5-fx+tb.xpi", "dictionaries/ko-KR.dic", "dictionaries/ko-KR.aff",
  43. "afrikaans_spell_checker-20110323-fx+tb+fn+sm.xpi", "dictionaries/af-ZA.dic", "dictionaries/af-ZA.aff",
  44. "albanisches_worterbuch-1.6.9-fx+tb+sm+fn.xpi", "dictionaries/sq.dic", "dictionaries/sq.aff",
  45. "amharic_spell_checker-0.4-fx+fn+tb+sm.xpi", "dictionaries/am_ET.dic", "dictionaries/am_ET.aff",
  46. "arabic_spell_checking_dictionary-3.2.20120321-fx+tb.xpi", "dictionaries/ar.dic", "dictionaries/ar.aff",
  47. "armenian_spell_checker_dictionary-0.32-fx+tb+sm.xpi", "dictionaries/hy_AM.dic", "dictionaries/hy_AM.aff",
  48. "azerbaijani_spell_checker-0.3-fx+tb+fn+sm+sb.xpi", "dictionaries/az-Latn-AZ.dic", "dictionaries/az-Latn-AZ.aff",
  49. "belarusian_classic_dictionary-0.1.2-tb+fx+sm.xpi", "dictionaries/be-classic.dic", "dictionaries/be-classic.aff",
  50. "belarusian_dictionary-0.1.2-fx+sm+tb.xpi", "dictionaries/be.dic", "dictionaries/be.aff",
  51. "bengali_bangladesh_dictionary-0.08-sm+tb+fx.xpi", "dictionaries/bn-BD.dic", "dictionaries/bn-BD.aff",
  52. "brazilian_portuguese_dictionary_former_spelling-28.20140203-tb+sm+fx.xpi", "dictionaries/pt-BR-antigo.dic", "dictionaries/pt-BR-antigo.aff",
  53. "brazilian_portuguese_dictionary_new_spelling-28.20140203-fx+sm+tb.xpi", "dictionaries/pt-BR.dic", "dictionaries/pt-BR.aff",
  54. "british_english_dictionary_updated-1.19.5-sm+fx+tb.xpi", "dictionaries/en-GB.dic", "dictionaries/en-GB.aff",
  55. "bulgarian_dictionary-4.3-fx+tb+sm.xpi", "dictionaries/bg.dic", "dictionaries/bg.aff",
  56. "canadian_english_dictionary-2.0.8-fx+sm+tb.xpi", "dictionaries/en-CA.dic", "dictionaries/en-CA.aff",
  57. "ceske_slovniky_pro_kontrolu_pravopisu-1.0.4-tb+sm+fx.xpi", "dictionaries/cs.dic", "dictionaries/cs.aff",
  58. "chichewa_spell_checker-0.3-fx+tb+fn+sm+sb.xpi", "dictionaries/ny_MW.dic", "dictionaries/ny_MW.aff",
  59. "corrector_de_galego-13.10.0-fn+sm+tb+fx.xpi", "dictionaries/gl_ES.dic", "dictionaries/gl_ES.aff",
  60. //BUG: broken flags "corrector_orthographic_de_interlingua-6.0-fn+sm+tb+fx.xpi", "dictionaries/ia-ia.dic", "dictionaries/ia-ia.aff",
  61. "corrector_ortografico_aragones-0.2-fx+tb+sm.xpi", "dictionaries/an_ES.dic", "dictionaries/an_ES.aff",
  62. "croatian_dictionary_-_hrvatski_rjecnik-1.0.1-firefox+thunderbird+seamonkey.xpi", "dictionaries/hr.dic", "dictionaries/hr.aff",
  63. "croatian_dictionary_hrvatski_rjecnik-1.0.9-an+fx+fn+tb+sm.xpi", "dictionaries/hr-HR.dic", "dictionaries/hr-HR.aff",
  64. "dansk_ordbog_til_stavekontrollen-2.2.1-sm+tb+fx.xpi", "dictionaries/da.dic", "dictionaries/da.aff",
  65. "deutsches_worterbuch_de_de_alte_rechtschreibung-2.1.8-sm.xpi", "dictionaries/de-DE-1901.dic", "dictionaries/de-DE-1901.aff",
  66. "diccionario_de_espanolespana-1.7-sm+tb+fn+fx.xpi", "dictionaries/es-ES.dic", "dictionaries/es-ES.aff",
  67. "diccionario_en_espanol_para_venezuela-1.1.17-sm+an+tb+fn+fx.xpi", "dictionaries/es_VE.dic", "dictionaries/es_VE.aff",
  68. "diccionario_espanol_argentina-2.5.1-tb+fx+sm.xpi", "dictionaries/es_AR.dic", "dictionaries/es_AR.aff",
  69. "diccionario_espanol_mexico-1.1.3-fn+tb+fx+sm.xpi", "dictionaries/es_MX.dic", "dictionaries/es_MX.aff",
  70. "diccionario_ortografico_valenciano-2.2.0-fx+tb+fn+sm.xpi", "dictionaries/roa-ES-val.dic", "dictionaries/roa-ES-val.aff",
  71. "diccionario_papiamentoaruba-0.2-fn+sm+tb+fx.xpi", "dictionaries/Papiamento.dic", "dictionaries/Papiamento.aff",
  72. "dictionnaires_francais-5.0.2-fx+tb+sm.xpi", "dictionaries/fr-classic-reform.dic", "dictionaries/fr-classic-reform.aff",
  73. "dictionnaires_francais-5.0.2-fx+tb+sm.xpi", "dictionaries/fr-classic.dic", "dictionaries/fr-classic.aff",
  74. "dictionnaires_francais-5.0.2-fx+tb+sm.xpi", "dictionaries/fr-modern.dic", "dictionaries/fr-modern.aff",
  75. "dictionnaires_francais-5.0.2-fx+tb+sm.xpi", "dictionaries/fr-reform.dic", "dictionaries/fr-reform.aff",
  76. "difazier_an_drouizig-0.12-tb+sm+fx.xpi", "dictionaries/br.dic", "dictionaries/br.aff",
  77. "dikshonario_papiamentuantia_hulandes-0.5-fx+tb+fn+sb+sm.xpi", "dictionaries/Papiamentu.dic", "dictionaries/Papiamentu.aff",
  78. "dizionari_furlan-3.1-tb+fx+sm.xpi", "dictionaries/fur-IT.dic", "dictionaries/fur-IT.aff",
  79. "dizionario_italiano-3.3.2-fx+sm+tb.xpi", "dictionaries/it_IT.dic", "dictionaries/it_IT.aff",
  80. "eesti_keele_speller-3.2-fx+tb+sm.xpi", "dictionaries/et-EE.dic", "dictionaries/et-EE.aff",
  81. "english_australian_dictionary-2.1.2-tb+fx+sm.xpi", "dictionaries/en-AU.dic", "dictionaries/en-AU.aff",
  82. "esperanta_vortaro-1.0.2-fx+tb+sm.xpi", "dictionaries/eo-EO.dic", "dictionaries/eo-EO.aff",
  83. "european_portuguese_spellchecker-14.1.1.1-tb+fx.xpi", "dictionaries/pt-PT.dic", "dictionaries/pt-PT.aff",
  84. "faroese_spell_checker_faroe_islands-2.0-tb+sm+fx+fn.xpi", "dictionaries/fo_FO.dic", "dictionaries/fo_FO.aff",
  85. "frysk_wurdboek-2.1.1-fn+sm+fx+an+tb.xpi", "dictionaries/fy.dic", "dictionaries/fy.aff",
  86. "geiriadur_cymraeg-1.08-tb+sm+fx.xpi", "dictionaries/cy_GB.dic", "dictionaries/cy_GB.aff",
  87. "general_catalan_dictionary-2.5.0-tb+sm+fn+fx.xpi", "dictionaries/ca.dic", "dictionaries/ca.aff",
  88. "german_dictionary-2.0.3-fn+fx+sm+tb.xpi", "dictionaries/de-DE.dic", "dictionaries/de-DE.aff",
  89. "german_dictionary_de_at_new_orthography-20130905-tb+fn+an+fx+sm.xpi", "dictionaries/de-AT.dic", "dictionaries/de-AT.aff",
  90. "german_dictionary_de_ch_new_orthography-20130905-fx+tb+fn+sm+an.xpi", "dictionaries/de-CH.dic", "dictionaries/de-CH.aff",
  91. "german_dictionary_de_de_new_orthography-20130905-tb+sm+an+fn+fx.xpi", "dictionaries/de-DE.dic", "dictionaries/de-DE.aff",
  92. "german_dictionary_extended_for_austria-2.0.3-fx+fn+sm+tb.xpi", "dictionaries/de-AT.dic", "dictionaries/de-AT.aff",
  93. "german_dictionary_switzerland-2.0.3-sm+fx+tb+fn.xpi", "dictionaries/de-CH.dic", "dictionaries/de-CH.aff",
  94. "greek_spelling_dictionary-0.8.5-fx+tb+sm.xpi", "dictionaries/el-GR.dic", "dictionaries/el-GR.aff",
  95. "gujarati_spell_checker-0.3-fx+tb+fn+sm+sb.xpi", "dictionaries/gu_IN.dic", "dictionaries/gu_IN.aff",
  96. "haitian_creole_spell_checker-0.08-tb+sm+fx.xpi", "dictionaries/ht-HT.dic", "dictionaries/ht-HT.aff",
  97. "hausa_spelling_dictionary-0.2-tb+fx.xpi", "dictionaries/ha-GH.dic", "dictionaries/ha-GH.aff",
  98. "hebrew_spell_checking_dictionary_from_hspell-1.2.0.1-fx+sm+tb.xpi", "dictionaries/he.dic", "dictionaries/he.aff",
  99. "hindi_spell_checker-0.4-fx+tb+sm+sb+fn.xpi", "dictionaries/hi_IN.dic", "dictionaries/hi_IN.aff",
  100. "hungarian_dictionary-1.6.1.1-fx+tb+sm+fn.xpi", "dictionaries/hu.dic", "dictionaries/hu.aff",
  101. //BUG: has no encoding declaration "icelandic_dictionary-1.3-fx+tb+sm.xpi", "dictionaries/is.dic", "dictionaries/is.aff",
  102. "kamus_pengecek_ejaan_bahasa_indonesia-1.1-fx+tb.xpi", "dictionaries/id.dic", "dictionaries/id.aff",
  103. "kannada_spell_checker-2.0.1-tb+sm+fn+an+fx.xpi", "dictionaries/kn.dic", "dictionaries/kn.aff",
  104. "kashubian_spell_checker_poland-0.9-sm+tb+fx.xpi", "dictionaries/Kaszebsczi.dic", "dictionaries/Kaszebsczi.aff",
  105. "kiswahili_spell_checker-0.3-sb+tb+fn+fx+sm.xpi", "dictionaries/sw_TZ.dic", "dictionaries/sw_TZ.aff",
  106. "kurdish_spell_checker-0.96-fx+tb+sm.xpi", "dictionaries/ku-TR.dic", "dictionaries/ku-TR.aff",
  107. "lao_spellchecking_dictionary-0-fx+tb+sm+fn+an.xpi", "dictionaries/lo_LA.dic", "dictionaries/lo_LA.aff",
  108. "latviesu_valodas_pareizrakstibas_parbaudes_vardnica-1.0.0-fn+fx+tb+sm.xpi", "dictionaries/lv_LV.dic", "dictionaries/lv_LV.aff",
  109. "lithuanian_spelling_check_dictionary-1.3-fx+tb+sm+fn.xpi", "dictionaries/lt.dic", "dictionaries/lt.aff",
  110. "litreoir_gaelspell_do_mhozilla-4.7-tb+fx+sm+fn.xpi", "dictionaries/ga.dic", "dictionaries/ga.aff",
  111. "litreoir_na_liongailise-0.03-fx+sm+tb.xpi", "dictionaries/ln-CD.dic", "dictionaries/ln-CD.aff",
  112. "macedonian_mk_mk_spellchecker-1.2-fn+tb+fx+sm+sb.xpi", "dictionaries/mk-MK-Cyrl.dic", "dictionaries/mk-MK-Cyrl.aff",
  113. "macedonian_mk_mk_spellchecker-1.2-fn+tb+fx+sm+sb.xpi", "dictionaries/mk-MK-Latn.dic", "dictionaries/mk-MK-Latn.aff",
  114. "malagasy_spell_checker-0.3-fn+tb+fx+sm+sb.xpi", "dictionaries/mg_MG.dic", "dictionaries/mg_MG.aff",
  115. "marathi_dictionary-9.3-sm+tb+sb+fx.xpi", "dictionaries/mr-IN.dic", "dictionaries/mr-IN.aff",
  116. "ndebele_south_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/nr-ZA.dic", "dictionaries/nr-ZA.aff",
  117. "nepali_dictionary-1.2-fx+tb.xpi", "dictionaries/ne_NP.dic", "dictionaries/ne_NP.aff",
  118. "norsk_bokmal_ordliste-2.0.10.2-fx+tb+sm.xpi", "dictionaries/nb.dic", "dictionaries/nb.aff",
  119. "norsk_nynorsk_ordliste-2.1.0-sm+fx+tb.xpi", "dictionaries/nn.dic", "dictionaries/nn.aff",
  120. "northern_sotho_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/nso-ZA.dic", "dictionaries/nso-ZA.aff",
  121. "oriya_spell_checker-0.3-fn+tb+fx+sm+sb.xpi", "dictionaries/or-IN.dic", "dictionaries/or-IN.aff",
  122. "polski_slownik_poprawnej_pisowni-1.0.20110621-fx+tb+sm.xpi", "dictionaries/pl.dic", "dictionaries/pl.aff",
  123. "punjabi_spell_checker-0.3-fx+tb+sm+sb+fn.xpi", "dictionaries/pa-IN.dic", "dictionaries/pa-IN.aff",
  124. "romanian_spellchecking_dictionary-1.14-sm+tb+fx.xpi", "dictionaries/ro_RO-ante1993.dic", "dictionaries/ro_RO-ante1993.aff",
  125. "russian_hunspell_dictionary-1.0.20131101-tb+sm+fn+fx.xpi", "dictionaries/ru_RU.dic", "dictionaries/ru_RU.aff",
  126. "sanskrit_spell_checker-1.1-fx+tb+sm+sb+fn.xpi", "dictionaries/sa_IN.dic", "dictionaries/sa_IN.aff",
  127. "scottish_gaelic_spell_checker-2.7-tb+fx+sm.xpi", "dictionaries/gd-GB.dic", "dictionaries/gd-GB.aff",
  128. "serbian_dictionary-0.18-fx+tb+sm.xpi", "dictionaries/sr-RS-Cyrl.dic", "dictionaries/sr-RS-Cyrl.aff",
  129. "serbian_dictionary-0.18-fx+tb+sm.xpi", "dictionaries/sr-RS-Latn.dic", "dictionaries/sr-RS-Latn.aff",
  130. "slovak_spell_checking_dictionary-2.04.0-tb+fx+sm.xpi", "dictionaries/sk-SK.dic", "dictionaries/sk-SK.aff",
  131. "slovak_spell_checking_dictionary-2.04.0-tb+fx+sm.xpi", "dictionaries/sk-SK-ascii.dic", "dictionaries/sk-SK-ascii.aff",
  132. "slovar_za_slovenski_jezik-0.1.1.1-fx+tb+sm.xpi", "dictionaries/sl.dic", "dictionaries/sl.aff",
  133. "songhay_spell_checker-0.03-fx+tb+sm.xpi", "dictionaries/Songhay - Mali.dic", "dictionaries/Songhay - Mali.aff",
  134. "southern_sotho_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/st-ZA.dic", "dictionaries/st-ZA.aff",
  135. "sownik_acinski-0.41.20110603-tb+fx+sm.xpi", "dictionaries/la.dic", "dictionaries/la.aff",
  136. "sownik_jezyka_dolnouzyckiego-1.4.8-an+fx+tb+fn+sm.xpi", "dictionaries/dsb.dic", "dictionaries/dsb.aff",
  137. "srpska_latinica-0.1-fx+tb+sm.xpi", "dictionaries/Srpski_latinica.dic", "dictionaries/Srpski_latinica.aff",
  138. "svenska_fria_ordlistan-1.1-tb+sm+fx.xpi", "dictionaries/sv.dic", "dictionaries/sv.aff",
  139. "svenska_fria_ordlistan-1.1-tb+sm+fx.xpi", "dictionaries/sv_FI.dic", "dictionaries/sv_FI.aff",
  140. "swati_spell_checker-20110323-tb+sm+fx+fn.xpi", "dictionaries/ss-ZA.dic", "dictionaries/ss-ZA.aff",
  141. "tamil_spell_checker_for_firefox-0.4-tb+fx.xpi", "dictionaries/ta-TA.dic", "dictionaries/ta-TA.aff",
  142. "telugu_spell_checker-0.3-tb+fx+sm.xpi", "dictionaries/te_IN.dic", "dictionaries/te_IN.aff",
  143. "te_papakupu_m__ori-0.9.9.20080630-fx+tb.xpi", "dictionaries/mi-x-Tai Tokerau.dic", "dictionaries/mi-x-Tai Tokerau.aff",
  144. "te_papakupu_m__ori-0.9.9.20080630-fx+tb.xpi", "dictionaries/mi.dic", "dictionaries/mi.aff",
  145. //BUG: broken file (hunspell refuses to load, too) "thamizha_solthiruthitamil_spellchecker-0.8-fx+tb.xpi", "dictionaries/ta_IN.dic", "dictionaries/ta_IN.aff",
  146. "tsonga_spell_checker-20110323-tb+sm+fx+fn.xpi", "dictionaries/ts-ZA.dic", "dictionaries/ts-ZA.aff",
  147. "tswana_spell_checker-20110323-tb+sm+fx+fn.xpi", "dictionaries/tn-ZA.dic", "dictionaries/tn-ZA.aff",
  148. //BUG: missing FLAG declaration "turkce_yazm_denetimi-3.5-sm+tb+fx.xpi", "dictionaries/tr.dic", "dictionaries/tr.aff",
  149. "turkmen_spell_checker_dictionary-0.1.6-tb+fx+sm.xpi", "dictionaries/tk_TM.dic", "dictionaries/tk_TM.aff",
  150. "ukrainian_dictionary-1.7.0-sm+an+fx+fn+tb.xpi", "dictionaries/uk-UA.dic", "dictionaries/uk-UA.aff",
  151. "united_states_english_spellchecker-7.0.1-sm+tb+fx+an.xpi", "dictionaries/en-US.dic", "dictionaries/en-US.aff",
  152. "upper_sorbian_spelling_dictionary-0.0.20060327.3-tb+fx+sm.xpi", "dictionaries/hsb.dic", "dictionaries/hsb.aff",
  153. "urdu_dictionary-0.64-fx+tb+sm+sb.xpi", "dictionaries/ur.dic", "dictionaries/ur.aff",
  154. "uzbek_spell_checker-0.3-fn+tb+fx+sm+sb.xpi", "dictionaries/uz.dic", "dictionaries/uz.aff",
  155. "valencian_catalan_dictionary-2.5.0-tb+fn+sm+fx.xpi", "dictionaries/ca-ES-valencia.dic", "dictionaries/ca-ES-valencia.aff",
  156. "venda_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/ve-ZA.dic", "dictionaries/ve-ZA.aff",
  157. "verificador_ortografico_para_portugues_do_brasil-2.3-3.2b1-tb+sm+fn+fx.xpi", "dictionaries/pt_BR.dic", "dictionaries/pt_BR.aff",
  158. "vietnamese_dictionary-2.1.0.159-an+sm+tb+fx+fn.xpi", "dictionaries/vi-DauCu.dic", "dictionaries/vi-DauCu.aff",
  159. "vietnamese_dictionary-2.1.0.159-an+sm+tb+fx+fn.xpi", "dictionaries/vi-DauMoi.dic", "dictionaries/vi-DauMoi.aff",
  160. "woordenboek_nederlands-3.1.1-sm+tb+fx+fn.xpi", "dictionaries/nl.dic", "dictionaries/nl.aff",
  161. "xhosa_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/xh-ZA.dic", "dictionaries/xh-ZA.aff",
  162. "xuxen-4.0.1-fx+tb+sm.xpi", "dictionaries/eu.dic", "dictionaries/eu.aff",
  163. "yiddish_spell_checker_yivo-0.0.3-sm+fn+fx+tb.xpi", "dictionaries/yi.dic", "dictionaries/yi.aff",
  164. "zulu_spell_checker-20110323-tb+fn+fx+sm.xpi", "dictionaries/zu-ZA.dic", "dictionaries/zu-ZA.aff"
  165. };
  166. public void test() throws Exception {
  167. for (int i = 0; i < tests.length; i += 3) {
  168. File f = new File(DICTIONARY_HOME, tests[i]);
  169. assert f.exists();
  170. try (ZipFile zip = new ZipFile(f, StandardCharsets.UTF_8)) {
  171. ZipEntry dicEntry = zip.getEntry(tests[i+1]);
  172. assert dicEntry != null;
  173. ZipEntry affEntry = zip.getEntry(tests[i+2]);
  174. assert affEntry != null;
  175. try (InputStream dictionary = zip.getInputStream(dicEntry);
  176. InputStream affix = zip.getInputStream(affEntry)) {
  177. Dictionary dic = new Dictionary(affix, dictionary);
  178. System.out.println(tests[i] + "\t" + RamUsageTester.humanSizeOf(dic) + "\t(" +
  179. "words=" + RamUsageTester.humanSizeOf(dic.words) + ", " +
  180. "flags=" + RamUsageTester.humanSizeOf(dic.flagLookup) + ", " +
  181. "strips=" + RamUsageTester.humanSizeOf(dic.stripData) + ", " +
  182. "conditions=" + RamUsageTester.humanSizeOf(dic.patterns) + ", " +
  183. "affixData=" + RamUsageTester.humanSizeOf(dic.affixData) + ", " +
  184. "prefixes=" + RamUsageTester.humanSizeOf(dic.prefixes) + ", " +
  185. "suffixes=" + RamUsageTester.humanSizeOf(dic.suffixes) + ")");
  186. }
  187. }
  188. }
  189. }
  190. public void testOneDictionary() throws Exception {
  191. String toTest = "hungarian_dictionary-1.6.1.1-fx+tb+sm+fn.xpi";
  192. for (int i = 0; i < tests.length; i++) {
  193. if (tests[i].equals(toTest)) {
  194. File f = new File(DICTIONARY_HOME, tests[i]);
  195. assert f.exists();
  196. try (ZipFile zip = new ZipFile(f, StandardCharsets.UTF_8)) {
  197. ZipEntry dicEntry = zip.getEntry(tests[i+1]);
  198. assert dicEntry != null;
  199. ZipEntry affEntry = zip.getEntry(tests[i+2]);
  200. assert affEntry != null;
  201. try (InputStream dictionary = zip.getInputStream(dicEntry);
  202. InputStream affix = zip.getInputStream(affEntry)) {
  203. new Dictionary(affix, dictionary);
  204. }
  205. }
  206. }
  207. }
  208. }
  209. }