ocr-layer-curation - [no description]

Age	Commit message (Collapse)	Author
2014-08-03	fix invert_align	Guillaume Horel

2014-08-01	add function to inverse alignment	Guillaume Horel

2014-07-29	Webapp now show three columns	Guillaume Horel
	image, original text and corrected text. The highlighting is functional as well.
2014-03-01	preliminary alignment_to_sexp	Guillaume Horel
	* right now just output a list of pairs (corrected_word, coords) * need to generate a sexp file if we want to reinsert into the djvu * bounding boxes are not smart at all right now (no merge or splits)
2013-08-21	small simplifaction	Guillaume Horel

2013-08-21	Simplify datastructure	Guillaume Horel
	An alignment is now a list of list. Empty list means word maps to nothing, and len(list) greater than one means a word maps to multiple words. This removes the artificial distinction between index and tuple.
2013-08-18	try to fix the alignment_to_sexp function	Guillaume Horel

2013-08-17	simplify Thibaut's code	Guillaume Horel

2013-08-17	add function for converting alignment to sexp	Guillaume Horel

2013-08-17	Take line jumps into accounts when grouping words	Thibaut Horel

2013-08-17	Some tweaks	Thibaut Horel

2013-08-06	Split words which map to two words	Thibaut Horel

2013-08-06	Adding some comments	Thibaut Horel

2013-08-05	Use C implementation of the Levenshtein distance	Thibaut Horel
	Requires the python-Levenshtein package on PyPI
2013-08-05	Use a Needleman-Wunsch type algorithm for text alignment	Thibaut Horel

2013-08-05	use new functions in compare.py	Guillaume Horel

2013-08-04	Add some string utils functions	Thibaut Horel
	Levenshtein distance and word hyphenation