/**
* Test of doOCR method, of class Tesseract.
*
* @throws Exception while processing image.
*/
@Test
public void testDoOCR_List_Rectangle() throws Exception {
File imageFile = null;
String expResult = "The (quick) [brown] {fox} jumps!\nOver the $43,456.78 <lazy> #90 dog";
String result = "<empty>";
try {
logger.info("doOCR on a PDF document");
imageFile = new File(this.testResourcesDataPath, "eurotext.pdf");
List<IIOImage> imageList = ImageIOHelper.getIIOImageList(imageFile);
result = instance.doOCR(imageList, null);
logger.info(result);
assertEquals(expResult, result.substring(0, expResult.length()));
} catch (IOException e) {
logger.error(
"Exception-Message: '{}'. Imagefile: '{}'",
e.getMessage(),
imageFile.getAbsoluteFile(),
e);
fail();
} catch (TesseractException e) {
logger.error(
"Exception-Message: '{}'. Imagefile: '{}'",
e.getMessage(),
imageFile.getAbsoluteFile(),
e);
fail();
} catch (StringIndexOutOfBoundsException e) {
logger.error(
"Exception-Message: '{}'. Imagefile: '{}'",
e.getMessage(),
imageFile.getAbsoluteFile(),
e);
fail();
}
}
@Test
public void testFolderExtraction() {
File tessDataFolder = null;
try {
/** Loads the image from resources. */
String filename = String.format("%s/%s", "/test-data", "eurotext.pdf");
URL defaultImage = getClass().getResource(filename);
File imageFile = new File(defaultImage.toURI());
/** Extracts <code>tessdata</code> folder into a temp folder. */
logger.info("Loading the tessdata folder into a temporary folder.");
tessDataFolder = LoadLibs.extractTessResources("tessdata");
/** Gets tesseract instance and sets data path. */
ITesseract instance = new Tesseract();
if (tessDataFolder != null) {
logger.info(tessDataFolder.getAbsolutePath());
instance.setDatapath(tessDataFolder.getAbsolutePath());
instance.setLanguage("por");
}
/** Performs OCR on the image. */
String result = instance.doOCR(imageFile);
logger.info(result);
} catch (TesseractException e) {
logger.error(e.getMessage());
logger.error(e.getMessage(), e);
} catch (URISyntaxException e) {
logger.error(e.getMessage(), e);
}
// checks if tessdata folder exists
assertTrue(tessDataFolder != null && tessDataFolder.exists());
}
