@Test public void testDefault() throws IOException { Dictionary dic = loadAbbDictionary(); char[] eos = {'.', '?'}; SentenceModel sdModel = train(new SentenceDetectorFactory("eng", true, dic, eos)); SentenceDetectorFactory factory = sdModel.getFactory(); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); ByteArrayOutputStream out = new ByteArrayOutputStream(); sdModel.serialize(out); ByteArrayInputStream in = new ByteArrayInputStream(out.toByteArray()); SentenceModel fromSerialized = new SentenceModel(in); factory = fromSerialized.getFactory(); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); }
@Test public void testNullDict() throws IOException { Dictionary dic = null; char[] eos = {'.', '?'}; SentenceModel sdModel = train(new SentenceDetectorFactory("eng", true, dic, eos)); SentenceDetectorFactory factory = sdModel.getFactory(); Assert.assertNull(factory.getAbbreviationDictionary()); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); ByteArrayOutputStream out = new ByteArrayOutputStream(); sdModel.serialize(out); ByteArrayInputStream in = new ByteArrayInputStream(out.toByteArray()); SentenceModel fromSerialized = new SentenceModel(in); factory = fromSerialized.getFactory(); Assert.assertNull(factory.getAbbreviationDictionary()); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); }
@Test public void testDefaultEOS() throws IOException { Dictionary dic = null; char[] eos = null; SentenceModel sdModel = train(new SentenceDetectorFactory("eng", true, dic, eos)); SentenceDetectorFactory factory = sdModel.getFactory(); Assert.assertNull(factory.getAbbreviationDictionary()); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(Factory.defaultEosCharacters, factory.getEOSCharacters())); ByteArrayOutputStream out = new ByteArrayOutputStream(); sdModel.serialize(out); ByteArrayInputStream in = new ByteArrayInputStream(out.toByteArray()); SentenceModel fromSerialized = new SentenceModel(in); factory = fromSerialized.getFactory(); Assert.assertNull(factory.getAbbreviationDictionary()); Assert.assertTrue(factory.getSDContextGenerator() instanceof DefaultSDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DefaultEndOfSentenceScanner); Assert.assertTrue(Arrays.equals(Factory.defaultEosCharacters, factory.getEOSCharacters())); }
@Test public void testDummyFactory() throws IOException { Dictionary dic = loadAbbDictionary(); char[] eos = {'.', '?'}; SentenceModel sdModel = train(new DummySentenceDetectorFactory("eng", true, dic, eos)); SentenceDetectorFactory factory = sdModel.getFactory(); Assert.assertTrue(factory.getAbbreviationDictionary() instanceof DummyDictionary); Assert.assertTrue(factory.getSDContextGenerator() instanceof DummySDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DummyEOSScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); ByteArrayOutputStream out = new ByteArrayOutputStream(); sdModel.serialize(out); ByteArrayInputStream in = new ByteArrayInputStream(out.toByteArray()); SentenceModel fromSerialized = new SentenceModel(in); factory = fromSerialized.getFactory(); Assert.assertTrue(factory.getAbbreviationDictionary() instanceof DummyDictionary); Assert.assertTrue(factory.getSDContextGenerator() instanceof DummySDContextGenerator); Assert.assertTrue(factory.getEndOfSentenceScanner() instanceof DummyEOSScanner); Assert.assertTrue(Arrays.equals(eos, factory.getEOSCharacters())); Assert.assertEquals(factory.getAbbreviationDictionary(), sdModel.getAbbreviations()); Assert.assertTrue(Arrays.equals(factory.getEOSCharacters(), sdModel.getEosCharacters())); }
@Override public void done() { samplesQueue.add(POISON); try { modelTrained.await(); if (ioException != null) { throw new RuntimeException(ioException); } if (sentenceModel == null) { throw new RuntimeException("Error training sentence model."); } OutputStream outputStream = Files.newOutputStream(outputPath.resolve("sentence.bin"), CREATE, TRUNCATE_EXISTING); sentenceModel.serialize(outputStream); } catch (InterruptedException e) { throw new RuntimeException("Interrupted before model could be saved."); } catch (IOException e) { throw new RuntimeException("Failed to write out model."); } }
@Override public void collectionProcessComplete() throws AnalysisEngineProcessException { try { stream.close(); } catch (IOException e) { throw new AnalysisEngineProcessException(e); } SentenceModel model; try { model = future.get(); } catch (InterruptedException | ExecutionException e) { throw new AnalysisEngineProcessException(e); } try (OutputStream out = new FileOutputStream(targetLocation)) { model.serialize(out); } catch (IOException e) { throw new AnalysisEngineProcessException(e); } } }
try { modelOut = new BufferedOutputStream(new FileOutputStream(modelFile)); model.serialize(modelOut); } finally { if (modelOut != null)
model.serialize(new FileOutputStream(outFile));
model.serialize(new FileOutputStream(outFile));
mod.serialize(outStream);
mod.serialize(outStream);