Changelog

0.6-0
	* Added getLSTMSymbolChoices() to access the alternative characters within each word and
	  their confidences from an LSTM recognition of the page/image.

0.5-0
	* Compatible with tesseract 5.0 thanks to work and a pull request from Wesley Brooks

	* Added level name to class of OCRResults, e.g. WordOCRResults.

	* Added OCRResultsConfidences class to OCRResults when Confidence column is included.

0.4-0
	* readMultipageTiff() function to read all pages of a TIFF file. Can pass each page to
	  tesseract() or use as a Pix object directly.

	* readImageInfo() function to read metadata from an image file, i.e., format, dimensions,
	  pixels per sampler, has a color map.

        * getAvailableLanguages() function.

	* Enhancements related to garbage collection and copying of Pix objects.

	* Use UTF-8 encoding on text obtained from OCR

	* Examples of reading Russian and Sanscrit documents.

	* Avoid garbage collection related segmentation fault related to GetImage().

	* dim() method for TesseractBaseAPI to get dimensions of current image.

	* Added oem<-() function for TesseractBaseAPI objects.

	* getLines()/findLines() now coerce the pix parameter to a Pix so can accept a
	  TesseractBaseAPI object or a path to an image file.

	* getLines()/findLines() convert the Pix to 8-bit

	* GetImageDims() works on OCRResults objects.

	* R global option "OCRConfidenceColors" controls the range of colors used in plot.OCR for displaying
	  the confidences.

	* lines() method for drawing object returned by getLines().

	* GetFontInfo() function.

0.3-1
	* Set the locale for each .Call() invocation to handle tesseract 4.0.

	* pixRead() has an addFinalizer logical argument to control whether this is added or not; related
	  to SetImage() and tesseract free'ing its image when tesseract itself is GC'ed.

	* Can treat the TesseractBaseAPI object returned by tesseract() as a named list to get/set the
	  tesseract variables. names(), $, [[, $<-, [[<-, [, [<-, methods.

0.3-0
	* getText(), getConfidences(), getBoxes(), getAlternatives() with methods for image file
	  name or TessBaseAPI objects (and others internally).

	* removed ocr() function

	* color the plot of results by confidence.

	* removed exposure of ResultIterator and Pix to the end user to avoid potential memory
	  management issues.

	* ReadConfigFile() has a debug = FALSE parameter.

	* several functions to convert the image to searchable/selectable PDF, HTML, TSV, OSD.

	* code more robust.

0.2-0
	* coerce tesseract API object to iterator as necessary

	* works with tesseract 3.05dev

	* improved configuration