/// <summary>
/// Performs tesseract OCR using wrapper for Tesseract OCR API for the selected page
/// of input image (by default 1st).
/// </summary>
/// <remarks>
/// Performs tesseract OCR using wrapper for Tesseract OCR API for the selected page
/// of input image (by default 1st).
/// Please note that list of output files is accepted instead of a single file because
/// page number parameter is not respected in case of TIFF images not requiring preprocessing.
/// In other words, if the passed image is the TIFF image and according to the
/// <see cref="Tesseract4OcrEngineProperties"/>
/// no preprocessing is needed, each page of the TIFF image is OCRed and the number of output files in the list
/// is expected to be same as number of pages in the image, otherwise, only one file is expected
/// </remarks>
/// <param name="inputImage">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="outputFiles">
///
/// <see cref="System.Collections.IList{E}"/>
/// of output files
/// (one per each page)
/// </param>
/// <param name="outputFormat">
/// selected
/// <see cref="OutputFormat"/>
/// for tesseract
/// </param>
/// <param name="pageNumber">number of page to be processed</param>
internal override void DoTesseractOcr(FileInfo inputImage, IList <FileInfo> outputFiles, OutputFormat outputFormat
, int pageNumber)
{
ScheduledCheck();
try {
ValidateLanguages(GetTesseract4OcrEngineProperties().GetLanguages());
InitializeTesseract(outputFormat);
OnEvent();
// if preprocessing is not needed and provided image is tiff,
// the image will be paginated and separate pages will be OCRed
IList <String> resultList = new List <String>();
if (!GetTesseract4OcrEngineProperties().IsPreprocessingImages() && ImagePreprocessingUtil.IsTiffImage(inputImage
))
{
resultList = GetOcrResultForMultiPage(inputImage, outputFormat);
}
else
{
resultList.Add(GetOcrResultForSinglePage(inputImage, outputFormat, pageNumber));
}
// list of result strings is written to separate files
// (one for each page)
for (int i = 0; i < resultList.Count; i++)
{
String result = resultList[i];
FileInfo outputFile = i >= outputFiles.Count ? null : outputFiles[i];
if (result != null && outputFile != null)
{
try {
using (TextWriter writer = new StreamWriter(new FileStream(outputFile.FullName, FileMode.Create), System.Text.Encoding
.UTF8)) {
writer.Write(result);
}
}
catch (System.IO.IOException e) {
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_WRITE_TO_FILE
, e.Message));
throw new Tesseract4OcrException(Tesseract4OcrException.TESSERACT_FAILED);
}
}
}
}
catch (Tesseract4OcrException e) {
LogManager.GetLogger(GetType()).Error(e.Message);
throw new Tesseract4OcrException(e.Message, e);
}
finally {
if (tesseractInstance != null)
{
TesseractOcrUtil.DisposeTesseractInstance(tesseractInstance);
}
if (GetTesseract4OcrEngineProperties().GetPathToUserWordsFile() != null && GetTesseract4OcrEngineProperties
().IsUserWordsFileTemporary())
{
TesseractHelper.DeleteFile(GetTesseract4OcrEngineProperties().GetPathToUserWordsFile());
}
}
}
/// <summary>
/// Check whether tesseract executable is installed on the machine and
/// provided path to tesseract executable is correct.
/// </summary>
/// <param name="execPath">path to tesseract executable</param>
private void CheckTesseractInstalled(String execPath)
{
try {
TesseractHelper.RunCommand(execPath, JavaCollectionsUtil.SingletonList <String>("--version"));
}
catch (Tesseract4OcrException e) {
throw new Tesseract4OcrException(Tesseract4OcrException.TESSERACT_NOT_FOUND, e);
}
}
public virtual void TestDetectAndFixBrokenBBoxes()
{
FileInfo hocrFile = new FileInfo(TEST_DOCUMENTS_DIRECTORY + "broken_bboxes.hocr");
IDictionary <int, IList <TextInfo> > parsedHocr = TesseractHelper.ParseHocrFile(JavaCollectionsUtil.SingletonList
(hocrFile), TextPositioning.BY_WORDS_AND_LINES);
TextInfo textInfo = parsedHocr.Get(1)[1];
NUnit.Framework.Assert.AreEqual(383.0f, (float)textInfo.GetBbox()[0], 0.1);
NUnit.Framework.Assert.AreEqual(101.0f, (float)textInfo.GetBbox()[1], 0.1);
NUnit.Framework.Assert.AreEqual(514.0f, (float)textInfo.GetBbox()[2], 0.1);
NUnit.Framework.Assert.AreEqual(136.0f, (float)textInfo.GetBbox()[3], 0.1);
}
/// <summary>
/// Performs OCR using provided
/// <see cref="iText.Pdfocr.IOcrEngine"/>
/// for the given list of
/// input images and saves output to a text file using provided path.
/// </summary>
/// <param name="inputImages">
///
/// <see cref="System.Collections.IList{E}"/>
/// of images to be OCRed
/// </param>
/// <param name="txtFile">file to be created</param>
public virtual void CreateTxtFile(IList <FileInfo> inputImages, FileInfo txtFile)
{
LogManager.GetLogger(GetType()).Info(MessageFormatUtil.Format(Tesseract4LogMessageConstant.START_OCR_FOR_IMAGES
, inputImages.Count));
StringBuilder content = new StringBuilder();
foreach (FileInfo inputImage in inputImages)
{
content.Append(DoImageOcr(inputImage, OutputFormat.TXT));
}
// write to file
TesseractHelper.WriteToTextFile(txtFile.FullName, content.ToString());
}
public virtual void TestPixSavingAsPng()
{
String path = TEST_IMAGES_DIRECTORY + "numbers_01.jpg";
String tmpFileName = GetTargetDirectory() + "testPixSavingAsPng.png";
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(tmpFileName)));
Pix pix = ImagePreprocessingUtil.ReadPix(new FileInfo(path));
TesseractOcrUtil.SavePixToTempPngFile(tmpFileName, pix);
NUnit.Framework.Assert.IsTrue(File.Exists(System.IO.Path.Combine(tmpFileName)));
TesseractHelper.DeleteFile(tmpFileName);
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(tmpFileName)));
}
public virtual void TestImageSavingAsPng()
{
String path = TEST_IMAGES_DIRECTORY + "numbers_01.jpg";
String tmpFileName = GetTargetDirectory() + "testImageSavingAsPng.png";
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(tmpFileName)));
System.Drawing.Bitmap bi = (System.Drawing.Bitmap)System.Drawing.Image.FromStream(new FileStream(path, FileMode.Open
, FileAccess.Read));
TesseractOcrUtil.SaveImageToTempPngFile(tmpFileName, bi);
NUnit.Framework.Assert.IsTrue(File.Exists(System.IO.Path.Combine(tmpFileName)));
TesseractHelper.DeleteFile(tmpFileName);
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(tmpFileName)));
}
public virtual void TestDoTesseractOcrForNonAsciiPathForExecutable()
{
String path = TEST_IMAGES_DIRECTORY + "tèst/noisy_01.png";
FileInfo imgFile = new FileInfo(path);
FileInfo outputFile = new FileInfo(TesseractOcrUtil.GetTempFilePath("test", ".hocr"));
Tesseract4OcrEngineProperties properties = new Tesseract4OcrEngineProperties();
properties.SetPathToTessData(GetTessDataDirectory());
properties.SetPreprocessingImages(false);
Tesseract4ExecutableOcrEngine engine = new Tesseract4ExecutableOcrEngine(GetTesseractDirectory(), properties
);
engine.DoTesseractOcr(imgFile, outputFile, OutputFormat.HOCR);
NUnit.Framework.Assert.IsTrue(File.Exists(System.IO.Path.Combine(outputFile.FullName)));
TesseractHelper.DeleteFile(outputFile.FullName);
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(outputFile.FullName)));
}
/// <summary>Reads data from the provided input image file.</summary>
/// <param name="input">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="outputFormat">
///
/// <see cref="OutputFormat"/>
/// for the result returned
/// by
/// <see cref="iText.Pdfocr.IOcrEngine"/>
/// </param>
/// <returns>
///
/// <see cref="ITesseractOcrResult"/>
/// instance, either
/// <see cref="StringTesseractOcrResult"/>
/// if output format is TXT, or
/// <see cref="TextInfoTesseractOcrResult"/>
/// if the output format is HOCR
/// </returns>
private AbstractTesseract4OcrEngine.ITesseractOcrResult ProcessInputFiles(FileInfo input, OutputFormat outputFormat
)
{
IDictionary <int, IList <TextInfo> > imageData = new LinkedDictionary <int, IList <TextInfo> >();
StringBuilder data = new StringBuilder();
IList <FileInfo> tempFiles = new List <FileInfo>();
AbstractTesseract4OcrEngine.ITesseractOcrResult result = null;
try {
// image needs to be paginated only if it's tiff
// or preprocessing isn't required
int realNumOfPages = !ImagePreprocessingUtil.IsTiffImage(input) ? 1 : ImagePreprocessingUtil.GetNumberOfPageTiff
(input);
int numOfPages = GetTesseract4OcrEngineProperties().IsPreprocessingImages() ? realNumOfPages : 1;
int numOfFiles = GetTesseract4OcrEngineProperties().IsPreprocessingImages() ? 1 : realNumOfPages;
for (int page = 1; page <= numOfPages; page++)
{
String extension = outputFormat.Equals(OutputFormat.HOCR) ? ".hocr" : ".txt";
for (int i = 0; i < numOfFiles; i++)
{
tempFiles.Add(CreateTempFile(extension));
}
DoTesseractOcr(input, tempFiles, outputFormat, page);
if (outputFormat.Equals(OutputFormat.HOCR))
{
IDictionary <int, IList <TextInfo> > pageData = TesseractHelper.ParseHocrFile(tempFiles, GetTesseract4OcrEngineProperties
().GetTextPositioning());
if (GetTesseract4OcrEngineProperties().IsPreprocessingImages())
{
imageData.Put(page, pageData.Get(1));
}
else
{
imageData = pageData;
}
result = new AbstractTesseract4OcrEngine.TextInfoTesseractOcrResult(imageData);
}
else
{
foreach (FileInfo tmpFile in tempFiles)
{
if (File.Exists(System.IO.Path.Combine(tmpFile.FullName)))
{
data.Append(TesseractHelper.ReadTxtFile(tmpFile));
}
}
result = new AbstractTesseract4OcrEngine.StringTesseractOcrResult(data.ToString());
}
}
}
catch (System.IO.IOException e) {
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_OCR_INPUT_FILE
, e.Message));
}
finally {
foreach (FileInfo file in tempFiles)
{
TesseractHelper.DeleteFile(file.FullName);
}
}
return(result);
}
/// <summary>
/// Performs tesseract OCR using command line tool for the selected page
/// of input image (by default 1st).
/// </summary>
/// <remarks>
/// Performs tesseract OCR using command line tool for the selected page
/// of input image (by default 1st).
/// Please note that list of output files is accepted instead of a single file because
/// page number parameter is not respected in case of TIFF images not requiring preprocessing.
/// In other words, if the passed image is the TIFF image and according to the
/// <see cref="Tesseract4OcrEngineProperties"/>
/// no preprocessing is needed, each page of the TIFF image is OCRed and the number of output files in the list
/// is expected to be same as number of pages in the image, otherwise, only one file is expected
/// </remarks>
/// <param name="inputImage">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="outputFiles">
///
/// <see cref="System.Collections.IList{E}"/>
/// of output files
/// (one per each page)
/// </param>
/// <param name="outputFormat">
/// selected
/// <see cref="OutputFormat"/>
/// for tesseract
/// </param>
/// <param name="pageNumber">number of page to be processed</param>
internal override void DoTesseractOcr(FileInfo inputImage, IList <FileInfo> outputFiles, OutputFormat outputFormat
, int pageNumber)
{
ScheduledCheck();
IList <String> @params = new List <String>();
String execPath = null;
String imagePath = null;
try {
imagePath = inputImage.FullName;
// path to tesseract executable
if (GetPathToExecutable() == null || String.IsNullOrEmpty(GetPathToExecutable()))
{
throw new Tesseract4OcrException(Tesseract4OcrException.CANNOT_FIND_PATH_TO_TESSERACT_EXECUTABLE);
}
else
{
if (IsWindows())
{
execPath = AddQuotes(GetPathToExecutable());
}
else
{
execPath = GetPathToExecutable();
}
@params.Add(execPath);
}
CheckTesseractInstalled(execPath);
// path to tess data
AddTessData(@params);
// validate languages before preprocessing started
ValidateLanguages(GetTesseract4OcrEngineProperties().GetLanguages());
// preprocess input file if needed and add it
imagePath = PreprocessImage(inputImage, pageNumber);
AddInputFile(@params, imagePath);
// move to image directory as tesseract cannot parse non ascii
// characters in input path
IList <String> moveToDirectoryParams = MoveToImageDirectory(imagePath);
// output file
AddOutputFile(@params, outputFiles[0], outputFormat, imagePath);
// page segmentation mode
AddPageSegMode(@params);
// add user words if needed
AddUserWords(@params, imagePath);
// required languages
AddLanguages(@params);
if (outputFormat.Equals(OutputFormat.HOCR))
{
// path to hocr script
SetHocrOutput(@params);
}
// set default user defined dpi
AddDefaultDpi(@params);
OnEvent();
TesseractHelper.RunCommand(IsWindows() ? "cmd" : "bash", CreateCommandList(moveToDirectoryParams, @params)
);
}
catch (Tesseract4OcrException e) {
LogManager.GetLogger(GetType()).Error(e.Message);
throw new Tesseract4OcrException(e.Message, e);
}
finally {
try {
if (imagePath != null && !inputImage.FullName.Equals(imagePath))
{
TesseractHelper.DeleteFile(imagePath);
}
}
catch (SecurityException e) {
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_DELETE_FILE
, imagePath, e.Message));
}
try {
if (GetTesseract4OcrEngineProperties().GetPathToUserWordsFile() != null && GetTesseract4OcrEngineProperties
().IsUserWordsFileTemporary())
{
TesseractHelper.DeleteFile(GetTesseract4OcrEngineProperties().GetPathToUserWordsFile());
}
}
catch (SecurityException e) {
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_DELETE_FILE
, GetTesseract4OcrEngineProperties().GetPathToUserWordsFile(), e.Message));
}
}
}