0.6.0

Features

Major

• Adds Diff View to ease filtering and comparing differences between recognition and ground truth text
• Completely refactors PAGE XML reading and writing so that the input annotation isn't discarded on saving but instead merged with the annotation in LAREX (when possible)
• Allows directly opening OCR-D workspaces by parsing the associated METS file
• Library UI overhauled to leverage the above mentioned implementation (Library View now also uses the direct request interface which was completely overhauled as well)
• Allows displaying / editing images and segmentation according to the orientation set in the PAGE XML files
• Adds displaying baselines
• Adds basic metadata editing (Creator, Comments, ...)
• Adds basic development enviroment
• Adds visibility toggle for TextLine polygons and baselines

Minor

• Adds fullscreen mode
• Updates OpenCV to v4.3.0-2
• Updates PaperJS to 0.12.15
• Updates Prima-Core-Libs to latest Git build
• Replaces custom config file with .properties file
• Refactors backend models (Polygons, Elements, ...)
• Extracts widely used code to functions and methods
• Outputs pretty printed XML files
• Adds more concrete messages to some error messages

Bugfixes

• Fixes bug which caused an error on saving when a TextLine is missing from the Region Order
• TextLines are now correctly removed in the frontend when TextRegion (or subtype) is changed to a region type which doesn't allow TextLines
• Now correctly uses XML files in savedir when loading annotations (thanks to @Witiko)
• Increased session timeouts
• Minor JS bugfixes

0.6-RC1

Features

Major

• Adds Diff View to ease filtering and comparing differences between recognition and ground truth text
• Completely refactors PAGE XML reading and writing so that the input annotation isn't discarded on saving but instead merged with the annotation in LAREX (when possible)
• Allows directly opening OCR-D workspaces by parsing the associated METS file
• Library UI overhauled to leverage the above mentioned implementation (Library View now also uses the direct request interface which was completely overhauled as well)
• Allows displaying / editing images and segmentation according to the orientation set in the PAGE XML files
• Adds displaying baselines
• Adds basic metadata editing (Creator, Comments, ...)
• Adds basic development enviroment
• Adds visibility toggle for TextLine polygons and baselines

Minor

• Adds fullscreen mode
• Updates OpenCV to v4.3.0-2
• Updates PaperJS to 0.12.15
• Updates Prima-Core-Libs to latest Git build
• Replaces custom config file with .properties file
• Refactors backend models (Polygons, Elements, ...)
• Extracts widely used code to functions and methods
• Outputs pretty printed XML files
• Adds more concrete messages to some error messages

Bugfixes

• Fixes bug which caused an error on saving when a TextLine is missing from the Region Order
• TextLines are now correctly removed in the frontend when TextRegion (or subtype) is changed to a region type which doesn't allow TextLines
• Now correctly uses XML files in savedir when loading annotations (thanks to @Witiko)
• Increased session timeouts
• Minor JS bugfixes

0.5.0

Features

• Added Batch Processing capabilities (Segmentation, Reading Order, Export)
• Added possibility to visually hide existing segmentation
• Added PAGE XML 2019 as default schema version

Bugfixes

• Fixed bug which made it impossible to load certain TIFF images
• LAREX won't be stuck when a local XML file is invalid anymore and lets the user see/edit the page instead
• Fixed a bug where users couldn't manually add lines to TextRegion elements without a specific subtype
• Fixed a bug which caused a crash during exporting annotations to PAGE XML files when no reading order for text lines existed
• Updated prima-core-libs version number
• Fixed minor GUI issues and inconsistencies
• Pages with empty annotations are now loaded correctly

0.5-RC4

Bugfixes

• Pages with empty annotations are now loaded correctly
• LAREX now collects local page annotations prior to running batch processing
• Various other fixes related to batch processing

0.5-RC3

Bugfixes

• Fixes a bug where users couldn't manually add lines to TextRegion elements without a specific subtype
• Fixes a bug which caused a crash during exporting annotations to PAGE XML files when no reading order for text lines existed
• Updated prima-core-libs version number

0.5-RC2

Fixes minor GUI issues and inconsistencies

0.5-RC1

Features

• Added Batch Processing (Segmentation, Reading Order, Export)
• Added possibility to visually hide existing segmentation
• Added PAGE XML 2019 as default schema version

Bugfixes

• TIFF images are now correctly loaded
• LAREX won't be stuck when a local XML file is invalid anymore and lets the user see/edit the page

0.4.0

Features

• Added subtraction rectangle/polygon for segments and textlines
• Added move mode for segments and textlines
• Overhauled page dropdown and added e.g. search functionality
• Added keyboard shortcut info modal
• Added possibility to discard ground truth with keyboard shortcut (Alt + D)
• The reading order info section can now be completely disabled from automatically opening
• Page Up and Page Down shortcuts added for navigating to the previous/next page
• Added option to disable reading order info panel
• Segmentation is now undoable
• marginalia and page_number were removed from the default regions

Bugfixes

• Changed unintuitive "textline fade" shortcut from Shift to Alt
• The reading order info section won't be triggered on changes which don't affect the reading order anymore
• Actually load existing annotations when the according button is clicked
• TextRegions without a subtype can now be imported, labeled and exported
• marginalia and page_number no longer reappear in the sidebar on pange change after deleting them

0.3.1

Hotfix

• Removes unused ConfigController which contained a RequestMapping that caused a mapping error

0.3.0

Features

• Different image types (binary, greyscale, despeckled) can now be directly selected within LAREX
• LAREX now provides a small selection of preconfigured virtual keyboards
• Adds page selection dropdown to allow easier navigation in large projects
• Adds OCR4all UI mode which hides UI elements which aren't needed for usage in conjunction with OCR4all
• Adds textline selection and deselection via the mouse cursor during Ground Truth Production

Bugfixes

• Last selected image type (bin, grey, …) will now be remembered on page change
• Fixes UnknownRegion related bug described in #175
• Fixes various typos and faulty tooltips
• Includes updated prima-core-libs to fix various bugs during PAGE XML reading and writing
• The last selected mode of the Text Viewer will now be correctly remembered on page change
• Textlines now have a minimum width to selection problems when the recognized text is empty or only consists out of very few characters