/** * If no words are provided, then a set of english default stopwords is used. */ public void testDefaults() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsFilterFactory factory = new CommonGramsFilterFactory(); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); Map<String, String> args = Collections.emptyMap(); factory.init(args); factory.inform(loader); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue(words.contains("the")); Tokenizer tokenizer = new MockTokenizer(new StringReader("testing the factory"), MockTokenizer.WHITESPACE, false); TokenStream stream = factory.create(tokenizer); assertTokenStreamContents(stream, new String[] { "testing", "testing_the", "the", "the_factory", "factory" }); }
/** * If no words are provided, then a set of english default stopwords is used. */ public void testDefaults() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsQueryFilterFactory factory = new CommonGramsQueryFilterFactory(); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); Map<String, String> args = Collections.emptyMap(); factory.init(args); factory.inform(loader); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue(words.contains("the")); Tokenizer tokenizer = new MockTokenizer(new StringReader("testing the factory"), MockTokenizer.WHITESPACE, false); TokenStream stream = factory.create(tokenizer); assertTokenStreamContents(stream, new String[] { "testing_the", "the_factory" }); }
public void testInform() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsFilterFactory factory = (CommonGramsFilterFactory) tokenFilterFactory("CommonGrams", TEST_VERSION_CURRENT, loader, "words", "stop-1.txt", "ignoreCase", "true"); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 2, words.size() == 2); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = (CommonGramsFilterFactory) tokenFilterFactory("CommonGrams", TEST_VERSION_CURRENT, loader, "words", "stop-1.txt, stop-2.txt", "ignoreCase", "true"); words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 4, words.size() == 4); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = (CommonGramsFilterFactory) tokenFilterFactory("CommonGrams", TEST_VERSION_CURRENT, loader, "words", "stop-snowball.txt", "format", "snowball", "ignoreCase", "true"); words = factory.getCommonWords(); assertEquals(8, words.size()); assertTrue(words.contains("he")); assertTrue(words.contains("him")); assertTrue(words.contains("his")); assertTrue(words.contains("himself")); assertTrue(words.contains("she")); assertTrue(words.contains("her")); assertTrue(words.contains("hers")); assertTrue(words.contains("herself")); }
public void testInform() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsQueryFilterFactory factory = (CommonGramsQueryFilterFactory) tokenFilterFactory("CommonGramsQuery", TEST_VERSION_CURRENT, loader, "words", "stop-1.txt", "ignoreCase", "true"); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 2, words.size() == 2); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = (CommonGramsQueryFilterFactory) tokenFilterFactory("CommonGramsQuery", TEST_VERSION_CURRENT, loader, "words", "stop-1.txt, stop-2.txt", "ignoreCase", "true"); words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 4, words.size() == 4); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = (CommonGramsQueryFilterFactory) tokenFilterFactory("CommonGramsQuery", TEST_VERSION_CURRENT, loader, "words", "stop-snowball.txt", "format", "snowball", "ignoreCase", "true"); words = factory.getCommonWords(); assertEquals(8, words.size()); assertTrue(words.contains("he")); assertTrue(words.contains("him")); assertTrue(words.contains("his")); assertTrue(words.contains("himself")); assertTrue(words.contains("she")); assertTrue(words.contains("her")); assertTrue(words.contains("hers")); assertTrue(words.contains("herself")); }
public void testInform() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsFilterFactory factory = new CommonGramsFilterFactory(); Map<String, String> args = new HashMap<String, String>(); args.put("words", "stop-1.txt"); args.put("ignoreCase", "true"); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); factory.init(args); factory.inform(loader); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 2, words.size() == 2); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = new CommonGramsFilterFactory(); args.put("words", "stop-1.txt, stop-2.txt"); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); factory.init(args); factory.inform(loader); words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 4, words.size() == 4); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = new CommonGramsFilterFactory(); args.put("words", "stop-snowball.txt"); args.put("format", "snowball"); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); factory.init(args); factory.inform(loader); words = factory.getCommonWords(); assertEquals(8, words.size()); assertTrue(words.contains("he")); assertTrue(words.contains("him")); assertTrue(words.contains("his")); assertTrue(words.contains("himself")); assertTrue(words.contains("she")); assertTrue(words.contains("her")); assertTrue(words.contains("hers")); assertTrue(words.contains("herself")); }
public void testInform() throws Exception { ResourceLoader loader = new ClasspathResourceLoader(TestStopFilter.class); assertTrue("loader is null and it shouldn't be", loader != null); CommonGramsQueryFilterFactory factory = new CommonGramsQueryFilterFactory(); Map<String, String> args = new HashMap<String, String>(); args.put("words", "stop-1.txt"); args.put("ignoreCase", "true"); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); factory.init(args); factory.inform(loader); CharArraySet words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 2, words.size() == 2); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = new CommonGramsQueryFilterFactory(); args.put("words", "stop-1.txt, stop-2.txt"); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); factory.init(args); factory.inform(loader); words = factory.getCommonWords(); assertTrue("words is null and it shouldn't be", words != null); assertTrue("words Size: " + words.size() + " is not: " + 4, words.size() == 4); assertTrue(factory.isIgnoreCase() + " does not equal: " + true, factory .isIgnoreCase() == true); factory = new CommonGramsQueryFilterFactory(); factory.setLuceneMatchVersion(TEST_VERSION_CURRENT); args.put("words", "stop-snowball.txt"); args.put("format", "snowball"); factory.init(args); factory.inform(loader); words = factory.getCommonWords(); assertEquals(8, words.size()); assertTrue(words.contains("he")); assertTrue(words.contains("him")); assertTrue(words.contains("his")); assertTrue(words.contains("himself")); assertTrue(words.contains("she")); assertTrue(words.contains("her")); assertTrue(words.contains("hers")); assertTrue(words.contains("herself")); }