private static IOcrEngine CreateEngine(OcrEngineType engineType, byte[] documentWriterOptions, bool useThunk)
{
IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(engineType, useThunk);
ocrEngine.Startup(null, null, null, null);
using (MemoryStream ms = new MemoryStream(documentWriterOptions))
{
ocrEngine.DocumentWriterInstance.LoadOptions(ms);
}
RasterCodecs codecs = ocrEngine.RasterCodecsInstance;
// Use the new RasterizeDocumentOptions to default loading document files at 300 DPI
codecs.Options.RasterizeDocument.Load.XResolution = 300;
codecs.Options.RasterizeDocument.Load.YResolution = 300;
codecs.Options.Pdf.Load.EnableInterpolate = true;
codecs.Options.Load.AutoFixImageResolution = true;
// We fine-tuned our app to only run a certain number of threads. The OCR engine can spawn it is own
// threads to increase the performance of a single recognition process, however, we do not want this
// in this scenario, these extra threads will decrease the overall performance of our app and we will
// not be able to keep track of how many threads are actually running
IOcrSettingManager settingManager = ocrEngine.SettingManager;
// Disable multi-threaded recognition
if (settingManager.IsSettingNameSupported("Recognition.Threading.MaximumThreads"))
{
settingManager.SetIntegerValue("Recognition.Threading.MaximumThreads", 1);
}
// Disable multi-threaded auto-zoning
if (settingManager.IsSettingNameSupported("Recognition.Zoning.DisableMultiThreading"))
{
settingManager.SetBooleanValue("Recognition.Zoning.DisableMultiThreading", true);
}
return(ocrEngine);
}
private void ThreadProc(object stateInfo)
{
WorkItemData data = (WorkItemData)stateInfo;
IOcrEngine ocrEngine = null;
bool passedCriticalStage = false;
try
{
// See if we have canceled
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
string destinationFile = Path.Combine(data.DestinationDirectory, Path.GetFileName(data.SourceFile));
ocrEngine = data.OcrEngine;
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
// Convert this image file to a document
string extension = DocumentWriter.GetFormatFileExtension(data.Format);
destinationFile = string.Concat(destinationFile, ".", extension);
if (data.Format == DocumentFormat.Ltd && File.Exists(destinationFile))
{
File.Delete(destinationFile);
}
string sourceFile = Path.GetFileName(data.SourceFile);
try
{
// Create a document and add the pages
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(null, OcrCreateDocumentOptions.AutoDeleteFile))
{
// Get the image number of pages
int imagePageCount;
RasterCodecs codecs = ocrDocument.RasterCodecsInstance;
using (CodecsImageInfo imageInfo = codecs.GetInformation(data.SourceFile, true))
{
long maximumMemorySize = 42187;
IOcrSettingManager settingManager = ocrEngine.SettingManager;
// Get the maximum size of the bitmap from the setting
if (settingManager.IsSettingNameSupported("Recognition.MaximumPageConventionalMemorySize"))
{
int maximumConventionalMemorySize = settingManager.GetIntegerValue("Recognition.MaximumPageConventionalMemorySize");
maximumMemorySize = (long)maximumConventionalMemorySize * 1024;
}
SetRecommendedLoadingOptions(codecs, imageInfo, maximumMemorySize);
imagePageCount = imageInfo.TotalPages;
}
// Set the DocumentWriter options
using (MemoryStream ms = new MemoryStream(data.DocumentWriterOptions))
{
ocrDocument.DocumentWriterInstance.LoadOptions(ms);
}
passedCriticalStage = true;
//recognize and add pages
for (int pageNumber = 1; pageNumber <= imagePageCount; pageNumber++)
{
lock (_abortedLockObject)
{
if (_aborted)
{
return;
}
}
var image = codecs.Load(data.SourceFile, pageNumber);
using (var ocrPage = ocrEngine.CreatePage(image, OcrImageSharingMode.AutoDispose))
{
ocrPage.Recognize(null);
ocrDocument.Pages.Add(ocrPage);
}
}
// Save
ocrDocument.Save(destinationFile, data.Format, null);
}
}
finally
{
}
OnSuccess(destinationFile);
}
catch (Exception ex)
{
string message;
if (passedCriticalStage && data.FirstTry)
{
message = string.Format("Error '{0}' while converting file '{1}' (first time, quarantined)", ex.Message, data.SourceFile);
AddToQuarantine(data.SourceFile);
}
else if (passedCriticalStage && !data.FirstTry)
{
message = string.Format("Error '{0}' while converting file '{1}' (quarantined error)", ex.Message, data.SourceFile);
}
else
{
message = string.Format("Error '{0}' while converting file '{1}'", ex.Message, data.SourceFile);
}
OnError(message);
}
finally
{
if (ocrEngine != null && ocrEngine != data.OcrEngine)
{
ocrEngine.Dispose();
}
if (Interlocked.Decrement(ref _workItemCount) == 0)
{
_batchFinishedEvent.Set();
}
}
}