2016/04/06 - 0.3.2:
* paperwork-chkdeps: Fix check for python-gi-cairo. When python-gi-cairo
  is not installed, sometimes, an exception pops up at an unexpected moment
  and the script remained stuck.


2016/02/25 - 0.3.1.1:
* Fix crappy dependency list


2016/02/25 - 0.3.1:
* Fix label learning
* Fix headerbar widget sizes


2016/02/16 - 0.3.0.1:
* Fix Paperwork packaging (.css files were not included)


2016/02/15 - 0.3.0:
* Whole GUI redesigned
* Added: dialog to make advanced searches
* New dependency: simplebayes
* Removed dependencies:
  - scikit* (replaced by simplebayes)
  - numpy* (replaced by simplesbayes)
  - gir1.2-gladeui (obsolete)


2015/11/25 - 0.2.5:
* Scanner support: Fujitsu scanners: handle options 'page-height' and
  'page-width'
* Scanner support: Brother MVC-J410: set mode correctly
  (value = '24bit Color' instead of 'Color' ...)
* Documents: add support for new label format that will be used in
  Paperwork >= 0.3.0
* paperwork-chkdeps: look for required icon themes
* Fix: work even if the spinner icon is not available
* Fix: paperwork-chkdeps: work even if Gtk is not yet installed
* Fix: PDF: reduce file descriptor leak
* Fix: With Pillow >= 3.x, calls to Image.rotate() must specify expand=True
* Fix: At startup, when updating the index, prevent infinite loop


2015/04/21 - 0.2.4:
* Fix python-whoosh 2.7 support


2015/04/03 - 0.2.3:
* Whenever possible, page orientation detection is now done using OCR tool
  feature (Tesseract >= 3.3.0). It's much faster and reliable.
* Fix doc indexation: last and first words of each lines weren't split
  correctly


2015/01/11 - 0.2.2:
* PDF + OCR: text wasn't indexed correctly
* Img doc: indexed text contains extra and useless data. As a side-effect,
  label prediction accuracy was strongly reduced.
  Rebuilding your index is strongly recommended
  ("rm -rf ~/.local/share/paperwork" + restarting Paperwork)


2014/12/19 - 0.2.1:
* Settings window : add help links
* Install process :
  - Extra dependencies are now detected by another script than setup.py
  - More missing dependencies are detected (aspell, tesseract, language
    packs, etc)
* Bug fixes :
  - Button 'open parent directory' doesn't remain stuck anymore when using
    the file manager Thunar
  - Settings window : Fix the way the file chooser is used (avoid
    selecting the wrong work directory by mistake)
  - Scanners support : Make it possible to use scanners even if some basic
    options are missing (source, resolution, etc)
  - When starting, don't remove empty directories anymore
  - Searching : Make sure diacritics characters are not a problem anymore
  - Import : accept file path containing spaces


2014/09/21 - 0.2:
* Improved search : whoosh.FuzzyTerm is now used
* Label look has been improved
* Menubar has been removed and replaced by an application menu
* Label prediction : when a new document is scanned, predicted labels are
  automatically set on it
* Pages are not displayed separately anymore
* New settings: scan source, number of orientations to try, OCR can be
  disabled
* Scans are displayed in real time


2014/07/08 - 0.1.3:
- Fix scanner support : don't try to set scanner options that are not active


2013/12/29 - 0.1.2:
- Improve scanner support: option names and values cases are not always the
  same on all the scanners
- Multiscan: fix multiscan end
- Translations: add german translations
- Settings window: display correctly Tesseract languages like 'deu-frak'


2013/10/03 - 0.1.1:
- Page list: fix display of page list longer than 100 pages
- Scanner support:
  - Fix support of scanners returning the supported resolutions as a range
    instead of an array
  - Fix: Always make sure the scan area is as big as it can be
- Fix: When OCR is disabled, fix scan and page editing
- Fix "no scanner found" popup (partial backport only, still slightly buggy)
- Scripts:
  - Add script scripts/obfuscate.py
  - Fix/Improve the output of scripts/stats.py


2013/08/08 - 0.1:
- Initial release
