CHANGES AND BUGFIXES
------------------------------

 0.1 -> 0.2: (19.06.2014)

-> Splitting now cuts through child elements, e.g. splitting a text region also splits parent lines and words
-> foregroundAlpha and backgroundAlpha values can now be set in the View->Change viewing settings menu 
	or via the config.properties file
-> F1-F5 keys can be used to toggle the visibiliy of printspace/text regions/lines/baselines/words
-> Virtual keyboards. Custom special characters can be added to the file "specialChars.txt"
-> Page-type was not saved bug fixed
-> Original filenames are shown in the titlebar and the "Documents" tab on the left
-> Bug fixed: could not apply text styles to multiple selected elements
-> MAC Support

0.2.1:

-> Some minor bugfixes and add-ons

0.3: (11.07.2014)
-> Update system (Help -> Check for updates)
-> Duplicate ID Problem fixed (I hope so at least...)
-> Transcript source can be opened via GUI (beside pages toolbar)
-> Different versions of a transcript could not be loaded -> fixed
-> Multiselect
-> specialChars.txt replaced with virtualKeyboards.xml --> added functionality to add multiple custom virtual keyboards
-> Rendering of different styles in transcription widgets can be turned toggled
-> Global font style is stored in config file
-> BUILD.txt specifies the current version and the build date and time

0.3.2: (23.07.2014)
-> Applying the font-styles now behaves different: font-size and kerning of 0 and an empty font-family  
	are NOT applied to sub-elements to avoid erasing this information!
 	NOTE: boolean values (serif yes/no, monospace yes/no, bold yes/no etc.) are still applied
 	 to each level and thus the original information gets erased!
-> Different apply buttons for structure and text-styles

0.4:
-> Indexed tagging and general custom tags

0.4.1:
-> Bugfixes
-> Changed SWT version from 4.3 to 4.4

0.4.2:
-> Bugfixes
-> Removed menu bar and moved it to "home" toolbar icon
-> Introduced undockable left/right/bottom widgets 

0.4.3:
-> Bugfixes & update to new REST interface
-> Minor changes in custom tag widget
-> added version tab / restricted version combo to 5 items

0.4.4:
-> Introduced simple bug report / feature request system -> access via "bug" button at the very top-right
-> Tagging of multiple different words selected in canvas
-> Bugfixes, amongst others: no more handles crash

0.4.5:
-> Bugfixes: editing in transcription widgets (word and line based) was broken in some cases
-> added text undo/redo in transcription widgets
-> Integrated XML source file viewer now has search functionality

0.4.6:
-> Changes over multiple lines should now be possible in the line transcription widget and behave as such:
	- if the number of lines inserted is greater than the number of affected lines (i.e. the marked lines), 
		the excessive newline characters are replaced with whitespaces to guarantee insertion of the whole text
	- if the number of lines inserted is less than the number of affected lines, the remaining affected lines will be cleared
-> Removed bug that selected parent region after any edit operation
-> Fixed another memory leak problem
-> Added login info (username & server) in doc-overview tab

0.4.7:

-> Some urgent bugfixes: 
	* Deleting words in word transcription editor not possible
	* Tagging / text style adding was buggy too
	* concurrent modification exceptions in doc view
-> Save dialog appears on exit if unsaved changes exist
-> first version of PDF export feature (via acrobat button on top toolbar)
-> Auto update search feature

0.4.8:
	-> Sorting functionality in doc overview
	-> doc overview: do not show public documents
	-> Program updater can overwrite config files
	-> Install a specific version of the tool via Menu->Install a specific version...
	
0.4.9:
	-> Reading order editing via the "Structure" view. 
		* Note: Once a reading order is defined inside a page / region / line, the elements get sorted  by this reading order first and then the coordinates!
		* You can auto-set the reading order according to the coordinates using the respective buttons in the "Structure" tab
		* The reading order is implemented on shape level via a custom tag called "readingOrder" with an integer value that determines the order.
		* If the reading order is equal for two shapes at a certain level, shapes are sorted by their coordinates.
		* If a reading order is set for some shape and not set for the other, the shape with the reading order comes first.
		* When saving the file a reading order on region level is also saved to the corresponding PAGE XML element ('ReadingOrder').
		Reading orders on other levels (lines / words) are not saved in this way, since the file cannot be loaded with Alethia anymore. 
		* Also, a reading order on region level set by the Alethia tool is synchronized with the custom tags on a page load.
		
	-> Adding of baselines, lines and words without their parent region -> if the parent region is not found, 
	the user is asked if he wants to automatically add the corresponding parent region.
	There is also a new toggle button (square shape with a plus) that when toggled disables this question.
	The corresponding property in the config.properties file is 'autoCreateParent'.
	
	-> Drag-and-Drop support in the "Structure" tab: elements can be moved from one parent to another via drag and drop, 
		i.e: lines can be moved to another region, words can be moved to another line.
		The drop has to be performed *onto* a region / line
		Note: it is not checked, whether the new parent region geometrically overlaps its new children -> use this option with care!
		
	-> Pasting multiple lines in the line transcription editor now fills up empty lines after the selection index if possible
	
0.5.0:
	-> Bugfixes (as always...)
	-> Finish a shape on double-click
	-> Polyrectangles (instead of just rectangles) are automatically created as parent shapes for baselines
	-> Choose to remember decision of auto-creating parent shapes in dialog
	-> Can now toggle the auto-centering feature in transcription widget (property is: centerCurrentTranscriptionLine)
	
0.5.3:
	-> TEI Export, 1st version
	-> upload via FTP to fix upload of very large files
	-> baseline extraction
	-> removed a bug: when selecting multiple elements in canvas, text style was not displayed correctly
	-> font family in text style editor is now a combobox with suggestions: Antiqua, Gothic, Normal
	
0.5.4:
	-> Some minor bugfixes: font family could not be changed, page lock icon always appeared, document could not be closed
	-> Other regions types from PAGE can now be displayed and its location changed: tables, images, graphics etc.
	-> First version of wordgraph editor
	
0.5.5:
	-> fixed bug report system: can now send bug reports from everywhere...
	-> changed logging framework from log4j to logback
	-> Can now add points on shapes by "dragging them away from the boundary"
	-> Bugfix: tags were lost when saving on server
	-> Some appearance modifications; most important: type of transcription (line or word based) is now changed via a combobox in the toolbar of the transcription widget
	-> Server side: introduced collections
	
0.5.6:
	-> introduced "tip of the day" widget
	-> changed start script and package names to "Transkribus.*"
	-> Transkribus.exe program file available for windows (using launch4j)
	-> bugfix: splitting of text-regions was duplicating child line elements
	-> splitting of a region, line or word now assigns all the text to the first element and none to the second -> no text duplication!
	
0.5.7:
	-> auto set email for bug request for logged-in user
	-> removed old starting scripts (TrpGui.sh, TrpGui.command)
	-> bugfix: image files starting with a dot (e.g. .xyz.jpg) will now be ignored on local documents -> relevant for mac!
	-> bugfix: structure types were not applied to multiple selection
	-> bugfix: job overview widget lost focus when updated
	-> first version of collection manager (create new collections, delete empty collections, find / add users to collections, add / remove documents in collections)
	
0.5.8:
	-> no "my collection" is added on register automatically
	-> on upload of document you can now specify the collection to upload to; if no collection with necessary rights exists, you can create one
	-> changing name of collections possible
	(-> resolved nasty "invisible top toolbar on window resize in windows when document loaded" bug - hopefully...)
	-> The program updater now also works via https only (default) -> property 'useFtpProgramUpdater' in config.properties adjusts if program updates are done via ftp or https
	
0.5.9:
	-> introudced options item in canvas toolbar (wrench symbol) -> summarizes all options for drawing new elements
	-> new shapes are added to overlapping shapes by default -> to add a shape only to a selected shape by default turn on the corresponding option in the canvas menu or set the 'addToOverlappingParentLine', 'addToOverlappingParentLine' properties properly in config.properties
	-> remodelled top menu bar to hopefully finally remove the invisible items bug
	-> can move image with left mouse button: if no shape is selected or if mouse not over selected shape
	-> multi-select with selection rectangle now only possible when holding *ctrl* button simultaneously
	-> adjust if shape is selected after creation (property: selectNewlyCreatedShape, default=false)
	
0.6.0:
	-> Remember credentials feature
	-> Changed image loader from jface/swt to jai due to mysterious white pages when loading group4 compressed images
	-> Minor bugfixes on shape selection behaviour
	
0.6.1:
	-> Beta version of new tagging system including properties and basic search functionality
	-> new pdf export
	-> multiple bug fixes
	
0.6.2:
	-> Bugfixes in tagging widget / search form
	-> Fixed 'abbrev' tag added
	-> added encoding for special chars in tag attributes
	
0.6.3:
	-> Fixed bug that too many search results in tag search widget crashes app -> results are now loaded "lazily"
	-> Introduced document search functionality
	-> Export list of tags as xslx file
	-> Change color of tags, save tag definitions to config.properties file

0.6.4:
	-> Search text (in same widget as tag search)
	-> Search next / previous tag and text search functionality	
	-> CATTI server integration. Tick "CATTI" checkbox in transcription widget to enable it. 
		ctrl-arrowdown retrieves next best hypotheses (rejectSuffix), edits in text can lead to different endings of sentence (setPrefix)

0.6.5:
	-> Enhanced export features: 
		- text pages in pdf
		- tags highlighting in pdf
		- rtf: tag overview page
		- TEI export with tags
		- excel export for tags
	-> tag normalization (in tag search widget)
	-> comparison between two versions
	-> first version of editorial declaration
	
0.6.6:
	-> Paging widgets
	-> Server reconstruction: introduces page-ids
	-> Page image replacement
	-> Write from right-to-left in transcription widgets
	-> support for empty tags (e.g. 'gap' tag)
	-> adding comments
	-> show different versions of image: orig, view, bin
	-> ftp upload
	-> blackenings
	-> pdf export bug resolved
	
0.6.7:
	-> bugfix: could not open local document images any more in windows
	
0.6.8:
	-> bugfix: could not save after deleting a version ("no transcript loaded" error message)
	-> setting file permissions in assembly plugin to 0755 -> starting scripts should now be executable by default
	-> remodelled updater dialogs

0.6.9:
	-> bugfix: "widget is disposed" error message after closing collection manager dialog once and reloading list of collections
	-> line bullet for current line in transcription widget is now bolded
	-> fixed bug in CATTI server (overruling when deleting whitespace)
	-> state and docId filter in jobs view
	-> fixed minor security leak: only admin users can view all jobs
	-> bugfix: nullpointerexception when exporting document with line that has no baseline

0.7.0:
	-> bugfix: editorial declaration dialog did not show any documents any more
	-> fixed some CATTI issues
	-> (re)introduced http upload of a single doc due to multiple 'connection timeout' problems when using apache wagon-ftp client - can now choose between http and ftp upload of single doc!
	-> unified upload dialog
	-> reading order improved
	
0.7.1:
	-> several bug fixes
	-> multiple java versions on mac problem solved in Transkribuc.command script: should now always take the newest one!
	-> improved TEI export:
		- word based TEI export: exporting zones and/or text on word level now possible
		- can now respect blackenings
		- includes all region types (graphic, images, blackening regions etc.) 
	-> user can view how is currently browsing which page ("People" button beside collection manager button in doc overview)
	-> receive copy of bug report / feature request option
 	-> paged table in search docs
 	-> CATTI bug with puncuation marks fixed by using tokenization / detokenization of prefix / suffix
 	-> bugfix: could not save text styles over whole line
 	-> bugfix: could not delete gap tag (and any other empty tag) when selection area was starting from this tag onwards
 	
 0.7.2:
	-> bugfix: baselines were assigned to wrong parent element when split
 	-> new "text recognition" interface
	-> bugfix: change of colors from config file were not applied after restart
	-> event messages
	-> bugfixes in PDF export
	-> Loupe function
	-> parent id of saved transcription gets stored 
	
0.7.3:
	-> save transcript with commit message (save button with balloon or ctrl-c)
	-> document copy functionality via collection manager (in my-documents table on bottom right)
	-> bugfix: loupe did not work for images that were scaled internally (i.e. images with more than 15 million pixels)
	-> render control characters (whitespace, tab, end of line) into the text (1st version...)
	-> bugfix: button that determines which styles are rendered did not work properly
	-> bugfix: dropdowntoolitem with radio style items fired selection event twice
	
0.7.4:
	-> Bugfix: crash when opening and closing collection manager
	-> new admin area functionality: apply affine transformations to doc; batch replace images for remote doc
	
0.7.5:
	-> fixed severe tagging bug: characters from other alphabets not recovered on transcript reload
	-> collection manager remodelled:
		add to collection, remove from collection and duplicate document now also possible for collection owner
		removed my-docs table and merged it with the documents table
	-> rtf export replaced by docx export
	
0.7.6:
	-> fixed minor bug in remember credentials functionality: last logged in user is remembered
	-> deleting a page added
	-> proxy settings
	-> display status and user in thumbnail overview
	
0.7.7:
	-> bugfix: could not load local document any more
	
0.7.8:
	-> rotate image 90 degrees
	-> bugfix: baseline focus with angle >180 degrees were displayed upside down
	-> export bugs fixed
	-> drawing direction arcs for baselines
	
0.7.9:
	-> adding new points on boundary only possible when holding down ctrl/cmd key (prevents adding new points by accident)
	-> bugfix: adding gap tag at the end of last line of region led to crash 
	-> rendering of polyrectangle around baselines (and selection of the baselines within those)
	-> locking zoom on focus possible (via wrench tool on canvas toolbar)
	-> auto login feature
	-> fixed catti bug (whitespace from received suffix removed)
	-> catti server made reentrant (multiple users from different clients now possible)
	-> adding paragraphs in textfield possible via shift+enter or dedicated button on transcription toolbar
	
0.8.0:
	-> profiles
	-> show complete img and transcript url in DocOverviewWidget

0.8.1:
	-> multiple bugfixes
	-> bugfix: after tag was added, "Tags under cursor" and "Properties" were not updated
	-> deleteLineIfBaselineDeleted property (configure via wrench tool)
	-> virtual keyboard configurable in gui (1st version)
	-> moved tag & text search to main search dialog in documents view
	-> find text and tag it
	
0.8.2:
	-> bugfixes: "no transcript loaded" when saving twice & virtual keyboards not working
	
0.8.3:
	-> splitting regions bug fixed
	-> swapped horizontal / vertical splitting semantic
	-> split parent line too if splitting baselines
	-> bugfix: group 4 compressed local images were displayed blank white
	-> Collection manager usability: document search possibility, load collections with double click,...
	-> Comments tool rearranged
	-> Document Table: mouse over shows complete document info 
	-> edit of virtual keyboard chars in gui saves back to virtualKeyboards.xml
	-> tei tags for Person, Place and Organization changed to "persName", "placeName" and "orgName"
	-> new tei tag "add" for addition
	-> tei export option: tag lines via <l>...</l> or <lb/> tag

0.8.4:
	-> gamma correction
	-> tif bug on windows fix
	-> Transcription and Segmentation profiles added
	-> createThumbs for local docs flag (configure via Menu -> File -> Create thumbs when opening local folder)
	-> Export: Possibility to export all chosen export formats in a ZIP file
	-> Docx export - consider soft breaks
	-> PDF import
	-> HTR based on RNN by Rostock integrated (alpha stage)
	
0.8.5:
	-> fixed a bug that occurred on duplicated documents with editorial declaration (collection-specific features)
	-> fixed a bug in the date setting widget in document metadata
	-> PAGE XML viewer can now edit XML and reload page from it
	
0.8.6:
	-> extended docx export settings for abbreviations
	-> PDF: images only export
	-> PDF: unique! tag values at end of PDF plus expansion for abbrevs
	-> export: choose current page and choose versions with their status
	-> docx index creation for tags (press F9 to create list in Word), export settings are shown in title page
	-> change to next/previous region in transcription widget with alt+left/right key
	-> batch layout analysis dialog
	-> right click behavior added for versions and transcripts (texts) 
	-> Login via OAuth2: Google
	
0.8.7:
	-> bugfixes for local doc reader & thumbnail widget deadlock
	-> local RGBA imgs bugfix
	-> select and load recently used documents from doc tab
	-> load last used document after login
	
0.8.8:
	-> removed auto-load of last used doc
	-> RNN HTR activated
	-> prevent install of same version which breaks application on windows 
	
0.9.0:
	-> move doc upload and search button from left document tab to the main menu bar
	-> colored background in thumbnail overview: light green=transcribed lines are there, 
		light yellow=now transcription (but lines segmented), light red=no line segmentation
	-> thumbnail overview also shows number of transcribed lines (per page and per document)
	-> automatic right to left text export for arabic, hebrew and so on
	-> included context menu in canvas (especially for tables)
	-> collapsed buttons to add elements to one dropdown button
	-> more strict version numbering (major version is first number, minor version is second number, hotfixes are third number; snapshot versions of any of those versions are identified by an additional fourth number and a '-SNAPSHOT' suffix)
	-> change to java 8
	-> first draft of table editor included (saving only works on local documents for now - until the server accepts the new PAGE format)
	-> removed obsolete writing orientation button
	-> redesigned tabs (no more right tab view)
	
0.9.1:
	-> save btn fix
	-> no more handles (transform) fix
	
0.9.2:
	-> reload doc btn reintegrated
	-> fixed reading order bug during split
	
0.10.0:
	-> vertical canvas edit toolbar
	-> removed all dropdown tool items and partially replaced by new toolboxes
	
0.10.1:
	-> "loadThumbs" property determines whether thumbnails are loaded (to prevent 'no more handles' issue)
	-> first version of local autosave
	-> pagination pagesize index out of bounds bug fixed
	-> setting heap space in windows starters (.exe & .bat) reversed due to "no more handles" bug
	-> Compare and show text differences of two versions in a text file (Tools - Compare)
	-> split/merge baselines is done for lines too; bugs fixed for undoing those actions
	
1.0.0:
	-> Deleting versions optimized
	-> Session expired check during recent doc load
	-> Change reading order optimized
	-> Refine reading order circles appearance
	-> Export tables into Docx and Excel (in excel tables only)
	-> Export as a rest service
	-> reading order behavior during region split
	-> show download link after server export to the user
	-> supplied tag handling for docx export
	-> option to show filename in thumbs manager
	-> included tag search
	-> jump to version 1.0.0 
		-> third number is for differentiating between snapshot versions for a specific release (i.e. 1.0.1-SNAPSHOT is the first snapshot for release 1.0)
		-> second number of for urgent bugfixes
		-> first number if the main release version
		
1.1.0:
	-> bugfix: NPE when empty description in HTR models
	-> bugfix: when replacing image on server, returned TrpPage could not be unmarshalled due to java.sql.Timestamp fields
	-> bugfix: update of page when tool job finished
	-> support for 'Supplementary Characters' (5 digits) in virtual keyboard
	-> bugfix: merge regions and lines is working again
	-> bugfix: seldom PDF export bug fixed
	-> included new la methods: ncsr (new), cvl & citlab (restricted access)
	-> Upload from private FTP: validation of directories now happens separately instead of during read process on server 
	-> new 'div' tag
	
1.1.1:
	-> determine CITlabLaJob availability in UI via server call
	-> remove "new" (which was actually the old one) and buggy NCSR LA method

1.1.2:
	-> private FTP upload bugfix
	-> bugfix: preserve reading order during export
	
1.2.0:
	-> remove "analyze structure" from tools
	-> add "other tools" area and integrate polygon to baselines as first such tool
	-> redesigned HtrTrainingDialog
	-> more details on HTR models in TextRecognitionConfigDialog
	-> context menu for doclist widget (delete, duplicate)
	-> default sorting for docs in doclist is doc-id descending
	-> collection-combo-viewer enhanced
	-> filter for doclist viewers
	-> removed collection manager and moved all of its functionality into the doclist viewer
	
1.2.1:
	-> collapse open-local-doc, import, export and find btns in server widget to one drop-down btn
	-> revert to old collection combo viewer (new one did not work on mac)
	
1.2.2:
	-> encode METS URL before sending it
	-> reload doclist bug fixed
	
1.2.3:
	-> fixed bug that prevented starting the LA without any regions
	-> HTR training interface modified
	-> collection metadata editor
	
1.3.0:
	-> new collection selector widget
	-> redesigned transcription widget toolbar
	-> table-cells are focused individually (formerly: whole table was focused when double-clicking table cell)
	-> bugfix: selection of shapes in canvas did not work when moving a little bit during click
	-> iterate through table cells using arrow keys in canvas
	-> bugfix: duplicate document possible again
	-> bugfix: pages selector could not be opened in tools widget
	-> "delete tags for current selection" added to transcription widget context menu
	-> autosave: prompt user if he wants to load autosaved version if newer than loaded transcript
	
1.3.1:
	-> fixed delete document
	 
1.3.2:
	-> integrate baseline2polygon tool
	-> collection-wise / multiple documents server export
	-> include TranskribusMetadata element with transcript metadata in export
	-> fixed TEI export issues regarding XML escaping (issue TranskribusCore#15)
	-> fixed export issues regarding duplicate filenames (issue TranskribusCore#13)
	-> fixed issue regarding export of documents with backslashes in title (issue TranskribusCore#6)

1.3.3:
	-> drag & drop to move elements between different parent shapes in 'Layout' tab
	-> button in 'Layout' tab that moves subshapes to parent shapes according to geometrical overlap
	-> do not check for local autosaves by default; introduce additional save dropdown where autosaves can be loaded explicitly
	-> show history of finished server exports not older than 2 weeks in export dialog
	-> bugfix: merging baselines (issue TranskribusSwtGui#21)
	-> align shapes in canvas according to text alignment (enable via transcription widget toolbar)
	-> HTR CER chart: tick-units fixed on y-axis, annotation for stored HTR, minor changes
	-> crowdProject management
	
1.3.4:
	-> fix NullpointerException in TrpMessageDialog	
	
1.3.5:
	-> normalizing tag attributes available again
	-> TrpDocUploadHttp: no file size limits, resumable uploads
	-> import local docs with text files only (in txt subfolder, one text file for each image, dummy lines with size of image are created for each line)
	-> CommonExportDialog: allow to choose image quality (original | JPEG)
	-> HtrTrainingDialog: fix bugs in selection of base model
	-> IO stuff: broken image files can now be handled (local/remote docs)
	-> fix high resolution issue
	-> fix PDF import bug
	
1.3.6:
	-> fix Abbyy XML import bug
	-> Export: fix version chooser
	-> show correct error message when a job can't be started from tools tab
	-> Fix a bug that blocked the HTR config. dialog when a non-existent HTR is selected
	-> KWS
	-> new default parameters for HTR training
	-> CITlab Advanced LA is now the default selection

1.3.7:
	-> fix NPE in Text2Image
	-> add "value" column in tag search&replace
	
1.4.0:
	-> add BIDI support in transcription widget
	-> refined PDF export
	-> Docx export considers tables resp. text in tables
	-> adapted tag search and replace
	-> new export option: simple txt export
	-> export images without loading transcripts
	-> Transcripts: Revert to the version(s) before the last (batch) job was completed
	-> add 'What is new' in Transkribus (make CHANGES.txt visible to the user)
	-> add search for documents by collection ID
	-> remove outdated TipsOfTheDay 
	-> add direct link to Transkribus HowTo guides
	-> extended virtual keyboard: Hebrew and Arabic are new
	-> change text based tagging system
	-> take local tmp directory as default for pdf upload
	
1.4.1:
	-> CITlabAdvanced layout analysis method can now be configured in user interface
	-> KWS: experimental test feature
	-> bugfix: cannot open local document
	-> bugfix: apply properties for tags over all continuations in property editor 
	-> bugfix: cannot open page due to widget closed exception
	-> bugfix: deletion of documents was sometimes not possible
	-> All tag definitions get stored in DB on server for later use in web interface
	-> Status of pages can be set for an entire document in the 'Document Manager'
	-> fix slow image rendering on Windows systems
	-> first draft of new structure type editor
	-> DOCX export: allow to force page breaks for each single page in a document
	-> sync dialog: allow synchronizing from txt files
	-> bugfix: memory and path reading errors for pdf upload 
	-> bugfix: resetting symbolic image during deletion of document
	-> PDF export: select your preferred font or use the default one ('FreeSerif') - mainly to cover the different used character sets
	-> compare versions now also considers tables
	-> Add 'saveTranscriptDialog' before tools start their job
	-> updated config for test server with separate imagestore
	-> Save changes dialog is now triggered also for table border mark-ups on leaving the current page
	
1.4.2:
	-> bugfix for structural tagging with custom defined structure types

1.5.0:
	-> In the export dialog the server export is set as 'default' tab
	-> TEI export (1): restricted to server export
	-> TEI export (2): We use a xslt (from https://github.com/dariok/page2tei) to transform page XMLs into TEI XMLs
	-> Customer specific transformations can be integrated as well
	-> Compare versions tool for several pages resp. a complete document 
	-> CITlabAdvanced and CVL are bottom-up approaches and thus detect always lines: check boxes are therefore preselected and disabled
	-> User manager: Roles are considered in order to determine what is allowed and what is not allowed
	-> show last save action on a document in the document manager (time, page number, user)
	-> Enable possibility to choose image quality for PDF export
	-> integrated more user-friendly table border markup box (called from context menu and border button) 
	-> Adapt the GUI appearance to the role a user has in the currently loaded collection, e.g. 'Reader' can not use the text editor or use 'save button' etc.
	-> Recycle bin for documents: contains deleted documents where they can be restored or deleted irreversible
	-> Stray documents list: the documents of a deleted collection can be added to another collection or irreversible deleted
	-> extended document metadata for archives (authority, hierarchy, backlink and external ID)
	-> bug fix: Error messages in the user manager were not closable 
	-> bug fix: merging of two shapes keeps the reading order and adds a space between them
	-> bug fix: selecting a tag from the table now also displays the right position in the image
	-> bug fix: Sorting lines according to coordinates now also works for tables (R-button in Layout tab)
	-> bug fix: Server export of tags and tables (xlsx) wasn't transparent to the user if there weren't any
	-> bug fix: Table export into xlsx also works fine for highly structured tables
	-> bug fix: vanishing properties in text tag property editor
	-> new textfield to quickly open a document with a certain id
	-> re-engineer the docking state buttons to quickly show/hide the control or transcription view with one click
	-> remove flickering in transcription widget while typing and scrolling
	-> add "show all lines of a page in the transcription widget" setting (enable via transcription widget toolbar wrench icon)
	-> add possibility to move transcription widget to the right side (enable via corresponding docking state dropdown menu button, i.e. third button on top left)
	-> PDF import: show progress during image extraction
	-> PDF import: path for temporary files is now correctly built in Linux
	-> bug fix: closing export dialog immediately after opening caused an exception (widget disposed)
	-> config.properties logHttp setting: enable to get full INFO level HTTP traffic log in console and log file 
	-> implement scrollbar mode for canvas and set it to default scrolling mode 
	-> convert from tables to simple textregions for all cells without losing lines and their transcriptions (via "Metadata->Structural" tab)
	-> CITlab HTR+ basic integration
	 
1.5.1:
	-> bugfix: UI crashed when opening advanced error rate tool and switching page afterwards
	
1.5.2:
	-> bugfix: user could not edit last line if empty (index out of bounds exception)

1.5.3:
	-> bugfix: nullpointer exception when merging table cells
	-> bugfix: disable automatic bidi processing, which is broken and leads to a program crash, and let the user manually switch between left-to-right and right-to-left orientation
	-> bugfix: fix broken baseline merging when merging lines with words
	-> bugfix: fix broken client side export (update saxon version)
	-> keep zoom when switching between pages
	-> bugfix: unable to add some unicode values to virtual keyboard
	
1.5.4:
	-> bugfix: broken image display on Linux systems
	-> enhancement/bugfix: remove unused jars from libs folder on program start
	
1.6.0:
	-> Displayed images are rotated according to EXIF orientation tag value
	-> Support user defined empty tags (like the <gap/> tag)
	-> Support java versions > 8 
	-> bugfix: virtual keyboard shortcut issues
	-> Add login btn to toolbar
	-> bugfix: clear doc table filter when loading another collection
	-> bugfix: Speed up 'Add page(s) to document' in document manager
	-> Enable export of versions created by some 'tool', User get selection of all 'toolnames' which have run on that document
	-> User activity: Shows a edit history per collection - more specifically sees the latest 'saves' and 'status changes' for each page and can also load that page with double click
	-> 'Compare Text Differences Tool' does not block, and (if wished) it shows the line numbering like the text editor
	-> Automatic saving tool: Stores transcript of a page from time to time locally to the user directory. If user loads a remote page it checks if a local version exists and if it is newer then the version on the server - if so it can be stored to the server. Settings can be made via the 'Save button' drop down menu.
	-> Option to change the reading order for all subsequent shapes as well, e.g. ro '35' changes to '3', afterwards '36' to '4', '37' to '5', '38' to '6' till the end of the shape list with the same parent.
	
1.7.0:
	-> several bugfixes
	-> P2PaLA recognition functionality for pre-trained models under ToolsWidget->Other Tools
	-> fix image exif orientation bug for MacOS, see bug TranskribusSwtGui#276
	-> simplified Text2Image tool (without training step) integrated in ToolsWidget->Other Tools 
	-> add feature to move pages of document by image filename list; start either via Main Menu->Document or the Document-Manager context menu on a loaded document
	
1.7.1:
	-> fix scrollbar layout issue in tools widget
	-> fix broken export for local documents
	-> CITlab HTR training obeys Unicode 12.1.0 character specification instead of Unicode 6.2
	-> usability: show configuration before running CITlabAdvanced layout analysis
	
1.8.0:
	-> additional parameters for Text2Image
	-> new sync txt-files and PAGE-XMLs with existing doc dialog (including: apply txt-files to existing layouts)
	-> text editor: 'Backspace' and 'Ctrl+Return' behavior - moves text one line up or down; Use case: applying txt-files to existing layouts may start at the wrong line which can be corrected
	-> HTR+ training using base models
	-> enable gzip encoding for HTTP API connection
	-> new HTR options regarding the line segments produced by CITlab HTR
		-> "Do polygon simplification": reduce the complexity of line segments, saving bandwidth and storage space. 
		-> "Keep original line polygons": copy the recognized text to the line segments without altering them.
	-> fix bug that caused image rotation according to EXIF thumbnail section
	-> remove small text-regions feature
	-> async page/transcript loading (do not show progress bar anymore)
	-> quickload of collection by id
	-> CITlab HTR: creating and storing ConfMats can be disabled if keyword spotting functionality is not required
	-> HTR models detail view bugfix: final CER values shown in text fields did not match the best net stored in all cases 
	
1.9.0:
	-> tag-search: minimum number of characters in tagname reduced to 1 (to be able to search for tags like e.g. "hi" or "p")
	-> bugfix: do *not* select underlying shape if image is moved with left mouse button
	-> ground truth management features
	-> rename test set to validation set as it is the more accurate term
	-> Allow deletion of HTR model via context menu (HTR overview table -> right-click -> Delete model...)
	-> HTR model metadata can be edited by owner/curator of model
	-> first draft of P2PaLA training (for interested early adopters only)
	-> bugfix: jumping to location of search result did not work anymore if page had to be changed
	-> bugfix: column-sensitive sorting resulted in errors -> skip those cases and sort by y-x coords only
	-> fix several layout issues
	
1.9.1:
	-> bugfix: force merge of baselines in horizontal direction
	
1.10.0:
	-> HTR ground truth data sets may be copied to a new document via right-click context menu
	-> HTR sharing dialog accessible via right-click context menu
	-> CITlab HTR configuration allows to select integrated dictionary or language model built from its train data set (available for models trained after 6.11.2018 [sic])
	-> bugfix: baselines were associated to wrong parent shapes on splitting with custom line
	-> CITlab HTR+ display bugfix: the model used for recognition is the one from the last epoch => display CER values from final training iteration.
	-> added Text2Image options
	
1.10.1:
	-> P2PaLA train dialog - "Analyze structure types" fixes
	-> HTR configuration dialog bugfix: did not load when stored configuration contained a deleted model
	-> fix a memory leak in CanvasImage
	-> xlsx export: concat continued tags
	
1.11.0:
	-> integrate baseline error rate tool
	-> improve P2PaLA train dialog - check structures and merged-structures consistency
	-> status change of one page is checked for reasonableness and creates new transcript version
	-> label existing layouts using P2PaLA
	-> Export option to export a text file for each selected page
	-> severe bugfix: update JRE for MacOS to version 8u251 to support server certificate 
	-> paged htr model tables
	
1.12.0:
	-> (PyLaia) training from compressed viewing files 
	-> bugfix: reading order after merge
	-> bugfix: editing pagination in the document manager dialog could lead to errors in the following save operations 
	-> PDF export: PDFs are much smaller now! for binarized images choosing 'Original' as image type delivers the smallest size, in most other cases 'JPEG'
	-> add collection doclist export as CSV (via green arrow-down button in the doclist toolbar)
	-> main menu: user can choose to load last modified document in Transkribus at start time
	-> HTR model lists are now all represented as tree or table widget pagination
	-> Undo a job (see the 'Jobs Overview': 'Undo button' will be enabled if the docID of the job corresponds to the ID of the currently loaded document)
	-> Documents moved to the recycle bin are automatically discarded after two weeks
	-> New cool feature ;) can be found here: 'Canvas tool bar' - 'Other segmentation tools' - 'Copy shapes to other pages': after the user has drawn tables or text regions he wants to copy this table/text regions for all the other pages as well because they look pretty much the same
	-> And another one: same same but different: 'Canvas tool bar' - 'Other segmentation tools' - 'Remove small text lines': deletes lines which are very small in relation to their region - needs to be used carefully and checked afterwards
	-> Last but not least in this series: 'Canvas tool bar' - 'Other segmentation tools' - 'Merge small text lines': automatically merges small text lines in the same row (should be used carefully)
	-> Editorial Declaration: export your selected transcription options directly in the 'Editorial Declaration Dialog' into XLSX
	-> automatic split of training set and validation set (choose if you want 2%, 5% or 10% of validation pages)
	-> bugfix: handwritten text recognition restriction on selected structures had no effect
	-> first integration of Document Understanding with a fixed model 
	
1.13.0:
	-> User can set the size of the 'reading order circles' in the viewing settings
	-> User management: take selected users over to another collection, roles are kept as they are and can be adapted in this other collection
	-> Document manager: store older transcript versions as latest by choosing the status or toolname. Use-case example: store 'Ground Truth' versions as latest
	-> Credit Manager Dialog: misc. usability fixes. Now shows the data for the collection, selected in the collection table, not the actually loaded collection when opened.
    -> HTR+ training: add option to skip lines with unclear and/or gap tags
    -> Event message dialog: is resizable & allows scrolling
    -> integrate new and simplified PyLaia preprocessing including line dewarping
    -> bugfix: UI freezes when HTR model was set for a collection and then deleted in HTR config dialog
    -> moving multiple shapes via ctrl/cmd+drag
    -> Login with Google is no longer supported. You can log in to your account after setting a new password on our website, using your Google email address as username. See the Login Dialog for more information.
    
 1.13.1:
 	-> bugfix: MacOS table redraw issue
 	-> (minor) bugfixes for MacOS: invisible Doc-ID text field and size of paging text fields
 	-> fixed credit manager usability Issues with Splitting Packages 
 	-> bugfix: change of collection did not work on first try
	-> remove OCR option for unauthorized users
	
1.14.0:
	-> PDF export: option to export PDF/A
	-> DOCX export: Users can choose to add the image name and page number before the text of each page
	-> batch processing: users can start jobs for several documents at the same time
	-> duplicate document(s) extended (1): user can choose status and only pages with that status get copied, use case: copy all GT pages into a new document
	-> duplicate document(s) extended (2): user can select several documents and copy them into one new document
	-> add final CER values in models table (Tools -> Models)
	-> perform PyLaia recognition on regions tagged with structure types only (as in CITlab recognition)
	-> Layout-Analysis: choose if lines are clustered or a single bounding-box is drawn (via advanced options)
	-> bugfix: textual tag-table redraw issue on MacOS
	-> bugfix: virtual-keyboard shortcuts were not activated on restart (only when opening virtual-keyboards dialog once)
	-> sample creation: added options to disregard short lines (specify a min fraction of the page width) & keep text of lines
	-> credit consumption display for a job can be accessed from Job Overview (the Transactions tab of the Credit Manager was removed)
	-> enhance and fix comments widget
	-> add code-signing of Transkribus.exe to prevent security warnings from virus scanners 
	
1.15.0:
	-> text reversion during training (Citlab and PyLaia)
	-> PyLaia Training: omit lines with gap or unclear tag
	-> export ALTO XML: update to schema version 4.2
	-> TEI server export: two variants of TEIs can be exported (added option as used from local export)
	-> Overview tab: GT pages are shown in a dark green background color
	-> bugfix: MacOS redrawing issues (search results & finding users) 
	-> bugfix: starting jobs on multiple docs was not possible
	-> bugfix: Excel Tag export: continued tags over multiple lines created multiple entries & wrong context for continued tags
	-> integrate Transkribus-OCR job for all users (block segmentation + HTR+ print model)
	-> bugfix: tag export -> unknown tags were not exported. Now: either export all tags in document or choose tag names from list; missing tag names can be added in text field
	
1.15.1:
	-> HTR+, PyLaia and P2PaLA trainings are now free for all (added links to tutorials in UI)
	
1.16.0:
	-> bugfix: opening document with broken images lead to UI deadlock
	-> introduce collection tags: beside user defined tags there are tag definitions for each collection - stored in the DB and shared with Transkribus Lite
	-> manage collection tags: create new tags, delete tags, take over tags from user tag definitions
	-> changes to collection tags are seen by all users of this collection and Transkribus Lite
	-> newer PAGE XML formats are converted to Transkribus import format 'PAGE 2013'
	-> Java 16 compatibility
	-> support Alto 4.2 import/export
	-> handle article information during split and merge
	-> Metadata tab: new possibility to assign metadata for each individual page of a document, gets stored in the PAGE XML (Alpha)
	
1.16.1:
	-> FTP Upload: folders and PDFs not ingested will be automatically deleted from the FTP server after 14 days
	-> PyLaia training: use textFeats as default preprocessing again (may have caused poor performance of some models due to missing deslanting option)
	
1.17.0:
	-> Main menu - Document - two alternative sorting methods (reading order) implemented:
	---> (1) first method: lines at the same height (within a region) are sorted according to X coordinate, otherwise Y coordinate
	---> (2) second method: sort lines/regions column wise, can fail because of overlapping shapes and this leads to the violation of transitivity.
	---> (3) both methods work for RTL too
	---> (4) both methods should be tested on single pages to see if useful or not before starting on a complete document
	-> bug fix for IIIF import
	-> Run HTR dialog: button to get all pages without (HTR) text: pages with text in lines and empty pages are omitted
	-> text recognition can be restricted by the user to pages without text
	-> text recognition can be restricted by the user to pages with status IN_PROGRESS, NEW (GT, FINAL, DONE excluded)
	-> Yes-No Dialog, that is shown before HTR is started, includes a info message on the nr. of credits that will be consumed
	-> fixed a bug that may cause document uploads to abort
	-> fixed a bug that caused misoriented layout on images with wrong width and height in EXIF data
	-> search/replace method (main menu -> Document), replacement rules are in a txt file, comma separated
	---> first the string to be replaced and after a comma the new string
	---> example: ſ,s (all occurences of long s will be replaced by s)
	---> further replacements are separated with newline
	-> own colors for status 'DONE' (light blue) and 'FINAL' (moderate blue)
	-> List of most recently used htr models in 'Run Text Recogntion'
	-> Set PyLaia HTR training as default
	-> Login dialog: add password reset button, remove account type selection	
	-> layout editor: added baseline extension for current page (Segmentation toolbar -> Other segmentation tools -> Extend baselines)
	-> P2PaLA bugfix: trained structure types of model were not displayed correctly (also: delete model button made more visible)
	-> PyLaia bugfix: display charset of model
	
1.18.0:
	-> Export of page related metadata into Excel
	-> Document Manager: "set transcripts ... as latest" -> add "New" state to enable reset to initial upload state
	-> bugfix in Document Manager which caused inability to start any jobs
	-> Page related metadata: it's possible to copy one property to other pages (beside to copy the complete set)
	-> enhance UI for phrase search in fulltext search
	-> easy start of a PyLaia training from HTR+ model data: right click on the model data -> train and validation data and settings are taken over
	-> 'Separator Detection' as new LA method
	-> Document metadata: user can add any values (name of metadata and value)
	-> allow changing color of Separators
	-> bug fix: remove lines without baseline after split
	-> bug fix: remove baselines with single point after split (remove parent line as well)
	-> Training of custom tags: helpful for e.g. to train expansions of abbreviations; trained model extracts the 'abbrev' tag with the expansion during recognition
	--> further use case: text styles: italic, bold, underlined etc. can be learned
	-> Added methods for working with tables:
	--> prepare table GT (works best with columns only, rows are difficult to learn); 50-100 pages without text lines
	--> method 'transform table to regions' creates input for P2PaLA (table becomes text regions)
	--> P2PaLA recognition produces regions: now use 'transform regions to table' to get a table again
	--> and 'create table rows' (experimental) helps to get the rows if the lines can be used for this task
	---> for the latter e.g. each line in the first column starts a new table row until the next line occurs
	-> sample-creation: add an option to add a maximum number of lines per page to the final sample
	-> add support for new ISO based model languages
	-> collection export on server: to avoid memory problems export is limited to 15 000 pages per export job
	--> an export of more pages is split into several jobs automatically
	--> for API users: please split up the exports as well 

1.18.1:
	-> bugfix: reading order was not editable in layout overview widget
	-> page metadata: auomatically create new MD set if none was set in PAGE-XML (--> s.t. no properties are lost!)  
	
1.19.0:
	-> important NOTE: expert client now needs Java >= 11 - install from here: https://www.oracle.com/java/technologies/downloads/
	-> support for Apple M1 chips (aarch64)
	-> new export feature for text export: one document is divided into several text files
	--> start tags and end tags must be annotated, each text file contains text between 'start_tag' and 'end_tag'
	--> export works also from 'start_tag' to the next 'start_tag'
	--> text file names can be extracted from the set properties, e.g. 'id' property in 'start_tag', if not, the filename is document name + unique number
	--> additionally tags can be selected which should be ignored in the text output
	-> NPE while starting collection export
	-> TEI export: tables are included in 'Variant 2' of server export and 'Client export'
	-> Document manager: bug fix - reset of transcripts with toolname 'Synced from local plaintext file...' is possible now for complete document
	-> support for newer Mac's with M1 chip (upgraded to SWT 4.23)
	-> Credit Manager: add account history tab
	-> intergrate new Models API (using SOLR)
	-> CITlab-LA training integration 
	-> Document manager: possibility to change the latest version with the previous one for the selected pages or the complete document
	--> normally this can be achieved by undo a job but not in all cases
	
1.19.1:
	-> Bugfix for "MessageBodyReader not found for media type=application/json;charset=utf-8, type=class eu.transkribus.core.model.beans.CostEstimationRepresentation" when starting new jobs
	 
1.19.2:
	-> bugfix for LA-Training (could not start training from baselines without text)
	-> fix included JREs for MacOS and added a MACOS_NOTES.txt into the package for startup instructions
	-> bugfix for export of documents without images
	-> Metadata - Structural: advanced options for structure types
	--> 1) Assign structure type to all regions without 'structure' in all selected pages
	--> 2) Delete structure type from all selected pages
	--> 3) Rename structure type in all selected pages
	-> Canvas - Other segmentation tools
	--> Beside removing small text regions it is now possible to remove 'empty' text regions as well. (Empty means without text lines)
	
1.19.3:
	-> Bugfix: could not select language model and dictionaries for HTR+ models ("eu.transkribus.core.model.beans.TrpModelMetadata.isLanguageModelExists()" is null error)
	-> added shortcuts for the segmentation view: ctrl+v/h - vertical and horizontal split, ctrl+m - merge selected shape, ctrl+b - add a baseline
	
1.20.0:
	-> Baseline (aka CITlab-LA) training integration revisted (integration for all users)
	-> fix language filter for models
	-> restructure tools widget - separate area for model training
	-> enable/disable showing structure type colors, and structure type text for selected shapes, is controlled via 'Eye' button above canvas widget
	-> Models Dialog & Model Data Viewer: add "My Models" filter mode
	-> Show credit consumption preview for Transkribus OCR and batch jobs
	-> Job type filter allows to select a value from predefined, common job names
	
1.20.1:
	-> bugfix: error during collection export (click on 'Choose documents..' was necessary)
	-> bugfix: fulltext- and tag-search were broken for regular users (NullPointerException)
	
1.21.0:
	-> enable PyLaia training from base-models with differing character set
	-> integrate TrHtr recognition for admins
	-> bugfix: ask for unsaved changes when opening new search result on different page
	-> Search/Replace job: (alpha release: testing phase for admins only)
	--> applicable via Search - Fulltext Search; the search hits can be selected and replaced;
	--> works for single word search and not for several search terms
	--> the replaced words may appear in a new search because the indexing of the new versions will take some time, please have that in mind!!
	-> new 'Undo' job; Search/Replace jobs can be undone
	-> refined tag training
	-> dark mode fixes (dark background color for text editor & cyan foreground for selected line when in dark mode)
	-> IIIF import adapted: if the manifest cannot be validated against IIIF Image API 2.1 the import tries to get the image URLs with an alternative approach
	-> assign unique IDs to multiple pages of document (Menu -> Document -> Assign unique IDs to pages...)
	
1.22.0:
	-> add option to train PyLaia and HTR+ models using existing line polygons
	-> add export option to create Alto version 2 instead of Alto version 4.2 (as default)
	-> make TrHtr recognition available for selected users
	-> new TextRegion's have structure type "paragraph" by default
	-> replace structure types for multiple pages of document or whole collection (Menu -> Document -> Replace structure types...)
	-> revision of recent docs widget
	-> selected line color of transcription widget now configurable (Menu -> View -> Change viewing settings)
	-> Advanced options for tagging: renaming/removing tags in selected pages of document (Metadata - Textual - Tag list - Advanced options button)
	-> Model tables and details include an editable "Creator" field
	-> re-activate Transkribus OCR Job using a PyLaia model
	
1.22.1:
	-> bugfix for MacOS: local documents could not be opened anymore ("directory is not writable" error)
	
1.23.0:
	-> Search and Annotate: for a search in the current document users can annotate the search hits with a specified tag
	--> e.g. search for 'Paris' and annotate with custom tag 'place'. Properties can be included and must be specified in the tag specifications
	--> specify the tags via 'Metadata - Textual - Customize - Tag specifications'
	--> works only for single words
	-> Search and Replace: for a search in the current document users can replace the search hits with another word
	--> the search hits can be selected page by page or can be done for a complete document (not for the complete collection)
	--> works only for single words
	-> Annotate and Replace options are only shown if the search was done for the current document with a single search term
	-> disable CITlab Advanced LA, Text2Image, HTR+ Training and KWS due to license restrictions
	
1.24.0:
	-> Canvas option: split baselines on region borders: now available for several pages/documents which can be selected via a dialog
	-> option for table export into Excel: a single column can be selected and it gets exported with the image crops of the corresponding cells.
	-> option for table export into Excel: all tables of the selected pages of the document are exported into one large table.
	-> baseline2Polygon: add option to simplify polygons
	
1.24.1:
	-> bugfix for 'unable to set input for text tag property editor - null argument'
	
1.24.2:
	-> more options for baseline detection w/o regions (restrict detection on regions with structure tags, split baselines on regions)
	-> PyLaia training: turn on binarization by default again & add parameter to main settings view (previously: only inside advanced parameters)
	--> reason: training data with heterogeneous background (e.g. some binarized pages, some not) led to terrible results
		
1.25.0:
	-> P2PaLA & LayoutAnalysis2: add option to keep existing regions (append new regions to current PAGE-XML)
	-> simplified parameter selection for baseline detection
	-> bugfix: baseline-detection model mixup when switching between test and prod server
	-> add image scaling parameter to baseline-detection configuration (for low/high-res images)