public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/json/test.json");
        final ExtractData extractData = jsonExtractor.getText(in, null);
        CloseableUtil.closeQuietly(in);

        final String content = extractData.getContent();
        logger.info(content);

        // Verify content contains expected fields
        assertTrue(content.contains("title"));

            File temporaryFile = ((TemporaryFileInputStream) is).getTemporaryFile();
            assertTrue(temporaryFile.exists());
            CloseableUtil.closeQuietly(is);
            assertFalse(temporaryFile.exists());
        } else {
            CloseableUtil.closeQuietly(is);
        }
    }

    private ResponseData getTestData1_OK() {
        final ResponseData responseData = new ResponseData();

        final ExtractData extractData = pdfExtractor.getText(in, null);
        final String content = extractData.getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
        assertEquals("Writer", extractData.getValues("Creator")[0]);

    public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/test.txt");
        final String content = textExtractor.getText(in, null).getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
    }

    public void test_getText_null() {
        try {
            textExtractor.getText(null, null);
            fail();

                        }
                        parser.parse(in, new BodyContentHandler(writer), metadata, parseContext);
                    } finally {
                        CloseableUtil.closeQuietly(in);
                    }
                }, contentEncoding, normalizeText);
                if (StringUtil.isBlank(content)) {
                    if (resourceName != null) {

    public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/lha/test.lzh");
        final String content = lhaExtractor.getText(in, null).getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
        assertTrue(content.contains("テキスト"));
    }

    public void test_getText_maxSize() throws IOException {

        }
    }

    /**
     * Closes a {@link Closeable} quietly, suppressing any {@link IOException}.
     *
     * @param closeable the closeable object
     */
    public static void closeQuietly(final Closeable closeable) {
        if (closeable == null) {
            return;
        }
        try {
            closeable.close();
        } catch (final IOException e) {
            // ignore
        }

            final CountDownLatch latch = new CountDownLatch(3);
            final Process process = jobProcess.getProcess();
            new Thread(() -> {
                try {
                    CloseableUtil.closeQuietly(process.getInputStream());
                } catch (final Exception e) {
                    logger.warn("Could not close a process input stream.", e);
                } finally {
                    latch.countDown();

    public void test_getText() {
        final InputStream in = ResourceUtil.getResourceAsStream("extractor/msoffice/test.doc");
        final String content = msWordExtractor.getText(in, null).getContent();
        CloseableUtil.closeQuietly(in);
        logger.info(content);
        assertTrue(content.contains("テスト"));
    }

    public void test_getText_null() {
        try {
            msWordExtractor.getText(null, null);
            fail();

                responseData.setContentLength(0);
            }
        } catch (final CrawlerSystemException e) {
            CloseableUtil.closeQuietly(responseData);
            throw e;
        } catch (final Exception e) {
            CloseableUtil.closeQuietly(responseData);
            throw new CrawlingAccessException("Could not access " + uri, e);
        }

        return responseData;
    }