void doOCROnCurrentPage() throws IOException, TikaException, SAXException {
if (config.getOCRStrategy().equals(NO_OCR)) {
return;
}
TesseractOCRConfig tesseractConfig =
context.get(TesseractOCRConfig.class, DEFAULT_TESSERACT_CONFIG);
TesseractOCRParser tesseractOCRParser = new TesseractOCRParser();
if (!tesseractOCRParser.hasTesseract(tesseractConfig)) {
throw new TikaException(
"Tesseract is not available. "
+ "Please set the OCR_STRATEGY to NO_OCR or configure Tesseract correctly");
}
PDFRenderer renderer = new PDFRenderer(pdDocument);
TemporaryResources tmp = new TemporaryResources();
try {
BufferedImage image = renderer.renderImage(pageIndex, 2.0f, config.getOCRImageType());
Path tmpFile = tmp.createTempFile();
try (OutputStream os = Files.newOutputStream(tmpFile)) {
// TODO: get output format from TesseractConfig
ImageIOUtil.writeImage(image, config.getOCRImageFormatName(), os, config.getOCRDPI());
}
try (InputStream is = TikaInputStream.get(tmpFile)) {
tesseractOCRParser.parseInline(is, xhtml, tesseractConfig);
}
} catch (IOException e) {
handleCatchableIOE(e);
} catch (SAXException e) {
throw new IOExceptionWithCause("error writing OCR content from PDF", e);
} finally {
tmp.dispose();
}
}
@Test
public void testImageMagick() throws Exception {
InputStream stream =
TesseractOCRConfig.class.getResourceAsStream("/test-properties/TesseractOCR.properties");
TesseractOCRConfig config = new TesseractOCRConfig(stream);
String[] CheckCmd = {config.getImageMagickPath() + TesseractOCRParser.getImageMagickProg()};
assumeTrue(ExternalParser.check(CheckCmd));
}
/*
If Tesseract is found, test we retrieve the proper number of supporting Parsers.
*/
@Test
public void offersTypesIfFound() throws Exception {
TesseractOCRParser parser = new TesseractOCRParser();
DefaultParser defaultParser = new DefaultParser();
ParseContext parseContext = new ParseContext();
MediaType png = MediaType.image("png");
// Assuming that Tesseract is on the path, we should find 5 Parsers that support PNG.
assumeTrue(canRun());
assertEquals(5, parser.getSupportedTypes(parseContext).size());
assertTrue(parser.getSupportedTypes(parseContext).contains(png));
// DefaultParser will now select the TesseractOCRParser.
assertEquals(
TesseractOCRParser.class, defaultParser.getParsers(parseContext).get(png).getClass());
}
/*
Check that if Tesseract is not found, the TesseractOCRParser claims to not support
any file types. So, the standard image parser is called instead.
*/
@Test
public void offersNoTypesIfNotFound() throws Exception {
TesseractOCRParser parser = new TesseractOCRParser();
DefaultParser defaultParser = new DefaultParser();
MediaType png = MediaType.image("png");
// With an invalid path, will offer no types
TesseractOCRConfig invalidConfig = new TesseractOCRConfig();
invalidConfig.setTesseractPath("/made/up/path");
ParseContext parseContext = new ParseContext();
parseContext.set(TesseractOCRConfig.class, invalidConfig);
// No types offered
assertEquals(0, parser.getSupportedTypes(parseContext).size());
// And DefaultParser won't use us
assertEquals(ImageParser.class, defaultParser.getParsers(parseContext).get(png).getClass());
}