Hotfix: Version numbers

Fix:

• Version numbers weren't updated in hocr output and setup.py in v1.3.2 #270

v1.3.2

Added:

• Add -f/--file option to ocropus-rtrain to read input filenames from a file #275

Fixed:

• Do not add $datarootdir/ocropus to model search path in windows #268

Changed:

• Use numpy functions instead of C implementations of sumprod/sumouter #265 #276
• Code and docs for on-the-fly compilation of C code removed #274
• Remove unused lru function annotaiton #273

v1.3.1

Python coding:

• Standardize Imports - Part I, #176 #206
• split functions from nlbin #244

Other features:

• Clip exponential in ffunc to avoid overflow #201
• Ignored empty lines in fonts list file #233
• Change checks for write_page_segmentation #220 (allow also small images for segmentation)
• Expand tests for coverage and CI

Bugfixing: #251, #252

v1.3.0 - Testing, CI, Python coding

Testing, continuous integration:

• Test page workflow and confidence measure, update coverage test #145
• circle.yml file for Circle CI testing, run-test-ci #149
• Travis CI builds #37
• Unit tests #209

Python coding:

• Cleanup imports in common.py, lstm.py, extract exceptions #154
• py3k: Use print function instead of statement #155
• py3k: Use new-style exceptions. #175

Other features:

• Fix behaviour of maxcolseps parameter in ocropus-gpageseg # 172
• Update characters for training #188
• Print summaries to stdout instead of stderr #170

Bugfixing: #133, #131, #179, #218
Cleanup: #180, #207, #181, #216, #219
Documentation: #185, #193, #194, #196, #205, #217

v1.2.0

• Improve installation process: #108, #110, #111, #117, #146, #148, #152
• Add new option --probabilities in ocropus-rpred #135
• Fix and improve hocr metadata: #105, #160

Bugfixing: #103, #123, #140
Cleanup: #84, #124, #143, #150, #156
Documentation: #120

v1.1.1

New feature:

• Added better print methods for distinguishing between info and error messages #53

Bugfixing and cleanup: #40, #43, #44, #75, #76

Expand documentations: #34, #39

v1.1.0 - Connect to CLSTM

New features:

• Connect to CLSTM training and recognition
• Added ocropus-lpred
• Added ocropus-ltrain
• Added ocropus-dewarp
• Added ocropus-linegen
• Added Apache license
• Allow "extract" outside original image bounds #19

Documentation:

• Add solution for OS X (clang) #28
• Add instructions for installing ocropy into a virtualenv #29

Reorganization:

• Updated download path for models
• moved ocropus-gtedit back to main directory

Bugfixing

v1.0

A cleaned up version of OCRopy with everything but the new RNN recognizer retired.

Library files that aren't needed anymore have been removed (actually, moved into OLD for the time being).

The installation process has been simplified.

A simple example of training has been added.

classic-ocropy-0.1.1

Classic ocropy with a few smallish fixes to make the test cases work after removing files.

classic-ocropy-0.1

This release contains the segmenting recognizer, language modeling, beam search, tree-VQ recognizer, and LSTM recognizer, plus character database editing tools. This was part of the OCRopus release as of late 2013.