/// <summary>
/// Reads
/// <see cref="Tesseract.Pix"/>
/// from input file or, if
/// this is not possible, reads input file as
/// <see cref="System.Drawing.Bitmap"/>
/// and then converts to
/// <see cref="Tesseract.Pix"/>.
/// </summary>
/// <param name="inputFile">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <returns>
/// Pix result
/// <see cref="Tesseract.Pix"/>
/// object from
/// input file
/// </returns>
internal static Pix ReadPix(FileInfo inputFile)
{
Pix pix = null;
try {
System.Drawing.Bitmap bufferedImage = iText.Pdfocr.Tesseract4.ImagePreprocessingUtil.ReadImageFromFile(inputFile
);
if (bufferedImage != null)
{
pix = TesseractOcrUtil.ConvertImageToPix(bufferedImage);
}
}
catch (Exception e) {
// NOSONAR
LogManager.GetLogger(typeof(iText.Pdfocr.Tesseract4.ImagePreprocessingUtil)).Info(MessageFormatUtil.Format
(Tesseract4LogMessageConstant.CANNOT_CONVERT_IMAGE_TO_PIX, inputFile.FullName, e.Message));
}
if (pix == null)
{
try {
pix = Tesseract.Pix.LoadFromFile(inputFile.FullName);
}
catch (Exception e) {
// NOSONAR
LogManager.GetLogger(typeof(iText.Pdfocr.Tesseract4.ImagePreprocessingUtil)).Info(MessageFormatUtil.Format
(Tesseract4LogMessageConstant.CANNOT_CONVERT_IMAGE_TO_PIX, inputFile.FullName, e.Message));
}
}
return(pix);
}
/// <summary>Checks whether parent directories are equal for the passed file paths.</summary>
/// <param name="firstPath">path to the first file</param>
/// <param name="secondPath">path to the second file</param>
/// <returns>true if parent directories are equal, otherwise - false</returns>
private bool AreEqualParentDirectories(String firstPath, String secondPath)
{
String firstParentDir = TesseractOcrUtil.GetParentDirectory(firstPath);
String secondParentDir = TesseractOcrUtil.GetParentDirectory(secondPath);
return(firstParentDir != null && firstParentDir.Equals(secondParentDir));
}
/// <summary>
/// Gets OCR result from provided multi-page image and returns result as
/// list of strings for each page.
/// </summary>
/// <remarks>
/// Gets OCR result from provided multi-page image and returns result as
/// list of strings for each page. This method is used for tiff images
/// when preprocessing is not needed.
/// </remarks>
/// <param name="inputImage">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="outputFormat">
/// selected
/// <see cref="OutputFormat"/>
/// for tesseract
/// </param>
/// <returns>
/// list of result string that will be written to a temporary files
/// later
/// </returns>
private IList <String> GetOcrResultForMultiPage(FileInfo inputImage, OutputFormat outputFormat)
{
IList <String> resultList = new List <String>();
try {
InitializeTesseract(outputFormat);
TesseractOcrUtil util = new TesseractOcrUtil();
util.InitializeImagesListFromTiff(inputImage);
int numOfPages = util.GetListOfPages().Count;
for (int i = 0; i < numOfPages; i++)
{
String result = util.GetOcrResultAsString(GetTesseractInstance(), util.GetListOfPages()[i], outputFormat);
resultList.Add(result);
}
}
catch (TesseractException e) {
String msg = MessageFormatUtil.Format(Tesseract4LogMessageConstant.TESSERACT_FAILED, e.Message);
LogManager.GetLogger(GetType()).Error(msg);
throw new Tesseract4OcrException(Tesseract4OcrException.TESSERACT_FAILED);
}
finally {
TesseractOcrUtil.DisposeTesseractInstance(GetTesseractInstance());
}
return(resultList);
}
/// <summary>Preprocess given image if it is needed.</summary>
/// <param name="inputImage">
/// original input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="pageNumber">number of page to be OCRed</param>
/// <returns>
/// path to output image as
/// <see cref="System.String"/>
/// </returns>
private String PreprocessImage(FileInfo inputImage, int pageNumber)
{
String tmpFileName = TesseractOcrUtil.GetTempFilePath(Guid.NewGuid().ToString(), GetExtension(inputImage));
String path = inputImage.FullName;
try {
if (GetTesseract4OcrEngineProperties().IsPreprocessingImages())
{
Pix pix = ImagePreprocessingUtil.PreprocessImage(inputImage, pageNumber);
TesseractOcrUtil.SavePixToTempPngFile(tmpFileName, pix);
if (!File.Exists(System.IO.Path.Combine(tmpFileName)))
{
System.Drawing.Bitmap img = TesseractOcrUtil.ConvertPixToImage(pix);
if (img != null)
{
TesseractOcrUtil.SaveImageToTempPngFile(tmpFileName, img);
}
}
}
if (!GetTesseract4OcrEngineProperties().IsPreprocessingImages() || !File.Exists(System.IO.Path.Combine(tmpFileName
)))
{
TesseractOcrUtil.CreateTempFileCopy(path, tmpFileName);
}
if (File.Exists(System.IO.Path.Combine(tmpFileName)))
{
path = tmpFileName;
}
}
catch (System.IO.IOException e) {
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_READ_INPUT_IMAGE
, e.Message));
}
return(path);
}
public virtual void TestReadingPageFromInvalidTiff()
{
String path = TEST_IMAGES_DIRECTORY + "example_03.tiff";
FileInfo imgFile = new FileInfo(path);
Pix page = TesseractOcrUtil.ReadPixPageFromTiff(imgFile, 0);
NUnit.Framework.Assert.IsNull(page);
}
public virtual void TestImageSavingAsPngWithError()
{
String path = TEST_IMAGES_DIRECTORY + "numbers_01.jpg";
System.Drawing.Bitmap bi = (System.Drawing.Bitmap)System.Drawing.Image.FromStream(new FileStream(path, FileMode.Open
, FileAccess.Read));
TesseractOcrUtil.SaveImageToTempPngFile(null, bi);
}
/// <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());
}
}
}
public virtual void TestInitializeImagesListFromInvalidTiff()
{
String path = TEST_IMAGES_DIRECTORY + "example_03.tiff";
FileInfo imgFile = new FileInfo(path);
TesseractOcrUtil tesseractOcrUtil = new TesseractOcrUtil();
tesseractOcrUtil.InitializeImagesListFromTiff(imgFile);
NUnit.Framework.Assert.AreEqual(0, tesseractOcrUtil.GetListOfPages().Count);
}
public virtual void TestPreprocessingConditions()
{
Pix pix = null;
NUnit.Framework.Assert.IsNull(TesseractOcrUtil.ConvertToGrayscale(pix));
NUnit.Framework.Assert.IsNull(TesseractOcrUtil.OtsuImageThresholding(pix));
NUnit.Framework.Assert.IsNull(TesseractOcrUtil.ConvertPixToImage(pix));
TesseractOcrUtil.DestroyPix(pix);
}
public virtual void TestNullSavingAsPng()
{
String tmpFileName = TesseractOcrUtil.GetTempFilePath(GetTargetDirectory() + "/testNullSavingAsPng", ".png"
);
TesseractOcrUtil.SaveImageToTempPngFile(tmpFileName, null);
NUnit.Framework.Assert.IsFalse(File.Exists(System.IO.Path.Combine(tmpFileName)));
TesseractOcrUtil.SavePixToTempPngFile(tmpFileName, null);
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 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 TestGetOcrResultAsStringForFile()
{
String path = TEST_IMAGES_DIRECTORY + "numbers_01.jpg";
String expected = "619121";
FileInfo imgFile = new FileInfo(path);
Tesseract4LibOcrEngine tesseract4LibOcrEngine = GetTesseract4LibOcrEngine();
tesseract4LibOcrEngine.SetTesseract4OcrEngineProperties(new Tesseract4OcrEngineProperties().SetPathToTessData
(GetTessDataDirectory()));
tesseract4LibOcrEngine.InitializeTesseract(OutputFormat.TXT);
String result = new TesseractOcrUtil().GetOcrResultAsString(tesseract4LibOcrEngine.GetTesseractInstance(),
imgFile, OutputFormat.TXT);
NUnit.Framework.Assert.IsTrue(result.Contains(expected));
}
public virtual void TestTesseract4OcrForPix()
{
String path = TEST_IMAGES_DIRECTORY + "numbers_02.jpg";
String expected = "0123456789";
FileInfo imgFile = new FileInfo(path);
Pix pix = ImagePreprocessingUtil.ReadPix(imgFile);
Tesseract4LibOcrEngine tesseract4LibOcrEngine = GetTesseract4LibOcrEngine();
tesseract4LibOcrEngine.SetTesseract4OcrEngineProperties(new Tesseract4OcrEngineProperties().SetPathToTessData
(GetTessDataDirectory()));
tesseract4LibOcrEngine.InitializeTesseract(OutputFormat.TXT);
String result = new TesseractOcrUtil().GetOcrResultAsString(tesseract4LibOcrEngine.GetTesseractInstance(),
pix, OutputFormat.TXT);
NUnit.Framework.Assert.IsTrue(result.Contains(expected));
}
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>Performs basic image preprocessing using buffered image (if provided).</summary>
/// <remarks>
/// Performs basic image preprocessing using buffered image (if provided).
/// Preprocessed image will be saved in temporary directory.
/// </remarks>
/// <param name="inputFile">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="pageNumber">number of page to be preprocessed</param>
/// <returns>
/// created preprocessed image as
/// <see cref="Tesseract.Pix"/>
/// </returns>
internal static Pix PreprocessImage(FileInfo inputFile, int pageNumber)
{
Pix pix = null;
// read image
if (IsTiffImage(inputFile))
{
pix = TesseractOcrUtil.ReadPixPageFromTiff(inputFile, pageNumber - 1);
}
else
{
pix = ReadPix(inputFile);
}
if (pix == null)
{
throw new Tesseract4OcrException(Tesseract4OcrException.CANNOT_READ_PROVIDED_IMAGE).SetMessageParams(inputFile
.FullName);
}
return(TesseractOcrUtil.PreprocessPix(pix));
}
/// <summary>
/// Gets OCR result from provided single page image and preprocesses it if
/// it is needed.
/// </summary>
/// <param name="inputImage">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <param name="outputFormat">
/// selected
/// <see cref="OutputFormat"/>
/// for tesseract
/// </param>
/// <param name="pageNumber">number of page to be OCRed</param>
/// <returns>result as string that will be written to a temporary file later</returns>
private String GetOcrResultForSinglePage(FileInfo inputImage, OutputFormat outputFormat, int pageNumber)
{
String result = null;
try {
// preprocess if required
if (GetTesseract4OcrEngineProperties().IsPreprocessingImages())
{
// preprocess and try to ocr
result = new TesseractOcrUtil().GetOcrResultAsString(GetTesseractInstance(), ImagePreprocessingUtil.PreprocessImage
(inputImage, pageNumber), outputFormat);
}
if (result == null)
{
System.Drawing.Bitmap bufferedImage = ImagePreprocessingUtil.ReadImage(inputImage);
if (bufferedImage != null)
{
try {
result = new TesseractOcrUtil().GetOcrResultAsString(GetTesseractInstance(), bufferedImage, outputFormat);
}
catch (Exception e) {
// NOSONAR
LogManager.GetLogger(GetType()).Info(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_PROCESS_IMAGE
, e.Message));
}
}
if (result == null)
{
// perform ocr using original input image
result = new TesseractOcrUtil().GetOcrResultAsString(GetTesseractInstance(), inputImage, outputFormat);
}
}
}
catch (Exception e) {
// NOSONAR
LogManager.GetLogger(GetType()).Error(MessageFormatUtil.Format(Tesseract4LogMessageConstant.TESSERACT_FAILED
, e.Message));
throw new Tesseract4OcrException(Tesseract4OcrException.TESSERACT_FAILED);
}
return(result);
}
/// <summary>
/// Create list of parameters for command moving to the image parent
/// directory.
/// </summary>
/// <param name="imagePath">path to input image</param>
/// <returns>command list</returns>
private IList <String> MoveToImageDirectory(String imagePath)
{
// go the image parent directory
IList <String> @params = new List <String>();
String parent = TesseractOcrUtil.GetParentDirectory(imagePath);
String replacement = IsWindows() ? "" : "/";
parent = parent.Replace("file:///", replacement).Replace("file:/", replacement);
// Use "/d" parameter to handle cases when the current directory on Windows
// is located on a different drive compared to the directory we move to
if (IsWindows())
{
@params.Add("cd /d");
}
else
{
@params.Add("cd");
}
@params.Add(AddQuotes(parent));
return(@params);
}
/// <summary>
/// Using provided input stream there will be created
/// temporary file (with name 'language.user-words')
/// containing words (one per line) which ends with
/// a new line character.
/// </summary>
/// <remarks>
/// Using provided input stream there will be created
/// temporary file (with name 'language.user-words')
/// containing words (one per line) which ends with
/// a new line character. Train data for provided language
/// should exist in specified tess data directory.
/// NOTE:
/// User words dictionary doesn't work properly in tesseract4
/// and hidden for public usage until fix is available
/// </remarks>
/// <param name="language">
/// language as
/// <see cref="System.String"/>
/// , tessdata for
/// this languages has to exist in tess data directory
/// </param>
/// <param name="inputStream">
/// custom user words as
/// <see cref="System.IO.Stream"/>
/// </param>
/// <returns>
/// the
/// <see cref="Tesseract4OcrEngineProperties"/>
/// instance
/// </returns>
internal virtual iText.Pdfocr.Tesseract4.Tesseract4OcrEngineProperties SetUserWords(String language, Stream
inputStream)
{
SetPathToUserWordsFile(null);
if (!GetLanguages().Contains(language))
{
if (DEFAULT_LANGUAGE.Equals(language.ToLowerInvariant()))
{
IList <String> languagesList = GetLanguages();
languagesList.Add(language);
SetLanguages(languagesList);
}
else
{
throw new Tesseract4OcrException(Tesseract4OcrException.LANGUAGE_IS_NOT_IN_THE_LIST).SetMessageParams(language
);
}
}
String userWordsFileName = TesseractOcrUtil.GetTempFilePath(language, "." + DEFAULT_USER_WORDS_SUFFIX);
try {
using (StreamWriter writer = new StreamWriter(userWordsFileName)) {
TextReader reader = new StreamReader(inputStream, System.Text.Encoding.UTF8);
int data;
while ((data = reader.Read()) != -1)
{
writer.Write(data);
}
writer.Write(Environment.NewLine);
SetPathToUserWordsFile(userWordsFileName, true);
}
}
catch (System.IO.IOException e) {
SetPathToUserWordsFile(null);
LogManager.GetLogger(GetType()).Warn(MessageFormatUtil.Format(Tesseract4LogMessageConstant.CANNOT_USE_USER_WORDS
, e.Message));
}
return(this);
}
/// <summary>
/// Initializes instance of tesseract if it haven't been already
/// initialized or it have been disposed and sets all the required
/// properties.
/// </summary>
/// <param name="outputFormat">
/// selected
/// <see cref="OutputFormat"/>
/// for tesseract
/// </param>
public virtual void InitializeTesseract(OutputFormat outputFormat)
{
if (GetTesseractInstance() == null || TesseractOcrUtil.IsTesseractInstanceDisposed(GetTesseractInstance())
)
{
tesseractInstance = TesseractOcrUtil.InitializeTesseractInstance(IsWindows(), GetTessData(), GetLanguagesAsString
(), GetTesseract4OcrEngineProperties().GetPathToUserWordsFile());
}
GetTesseractInstance().SetVariable("tessedit_create_hocr", outputFormat.Equals(OutputFormat.HOCR) ? "1" :
"0");
GetTesseractInstance().SetVariable("user_defined_dpi", "300");
if (GetTesseract4OcrEngineProperties().GetPathToUserWordsFile() != null)
{
GetTesseractInstance().SetVariable("load_system_dawg", "0");
GetTesseractInstance().SetVariable("load_freq_dawg", "0");
GetTesseractInstance().SetVariable("user_words_suffix", GetTesseract4OcrEngineProperties().GetDefaultUserWordsSuffix
());
GetTesseractInstance().SetVariable("user_words_file", GetTesseract4OcrEngineProperties().GetPathToUserWordsFile
());
}
TesseractOcrUtil.SetTesseractProperties(GetTesseractInstance(), GetTessData(), GetLanguagesAsString(), GetTesseract4OcrEngineProperties
().GetPageSegMode(), GetTesseract4OcrEngineProperties().GetPathToUserWordsFile());
}
/// <summary>
/// Creates a new
/// <see cref="Tesseract4LibOcrEngine"/>
/// instance.
/// </summary>
/// <param name="tesseract4OcrEngineProperties">set of properteis</param>
public Tesseract4LibOcrEngine(Tesseract4OcrEngineProperties tesseract4OcrEngineProperties)
: base(tesseract4OcrEngineProperties)
{
tesseractInstance = TesseractOcrUtil.InitializeTesseractInstance(IsWindows(), null, null, null);
}
/// <summary>
/// Reads input file as Leptonica
/// <see cref="Tesseract.Pix"/>
/// and
/// converts it to
/// <see cref="System.Drawing.Bitmap"/>.
/// </summary>
/// <param name="inputImage">
/// input image
/// <see cref="System.IO.FileInfo"/>
/// </param>
/// <returns>
/// returns a
/// <see cref="System.Drawing.Bitmap"/>
/// as the result
/// </returns>
internal static System.Drawing.Bitmap ReadAsPixAndConvertToBufferedImage(FileInfo inputImage)
{
Pix pix = Tesseract.Pix.LoadFromFile(inputImage.FullName);
return(TesseractOcrUtil.ConvertPixToImage(pix));
}
/// <summary>Creates a temporary file with given extension.</summary>
/// <param name="extension">
/// file extension for a new file
/// <see cref="System.String"/>
/// </param>
/// <returns>
/// a new created
/// <see cref="System.IO.FileInfo"/>
/// instance
/// </returns>
private FileInfo CreateTempFile(String extension)
{
String tmpFileName = TesseractOcrUtil.GetTempFilePath(Guid.NewGuid().ToString(), extension);
return(new FileInfo(tmpFileName));
}