- accept sentences without
# text = ...(but issue warning)
- table view: Ctrl-click on Word ID keeps Word marked until sentence is changed or ID clicked again (to accelerate edtiing using shortcuts)
- corrections:
- add non-standard columns when adding words and MWT
- unmark word when changing current sentence (n in table view)
- use
*instead of_for empty non-standard columns (CoNLL-U plus) - new tests
- clear button also empties comment search
- table view:
- keep modified column width
- FORM, LEMMA, UPOS, XPOS are also modifiable
- table view:
- buttons to change display width of FEATS, DEPS and MISC column
- short cuts work here as well now
- compiling with openjdk 17
- unitary tests upgraded to jupiter.junit 5
- dependencies upgraded to latest version
- new tests for server API
- if
show basic in enhancedis activated, this also applies for LaTeX output - new test
Feat:VerbForm=orMisc:SpaceAfter=(no value given) searches for words which have the feature/misc name with any value (usenot Feat:VerbForm:to look for words which do not have the given feature)- new tests
- bug concerning --rootdir corrected
- MISC shortcut correction,
mod addmiscadded - new test
- undo/redo bug corrected
- check json encoded error messages in tests
- shortcuts for MISC column added
- update
# text = ...aftersentsplit,sentjoin,compose,splitandjoin - tests updated
- corrected
bin/replace.shfor MacOS users
- warning if the
# text = ...does not correspond to the concatenated forms (respecting Space(s)After) - inputfield to edit value for
# textor set from concatenated forms - tests updated
- new tests
- added a search fonction which implements (partially) grew match search patterns
- search menu acces changed to combo box
- new tests
- corrected Locale issue in tests
- bug corrected in joining two tokens
- option
-rno longer needed (using another way to calculate the position ofconllueditor/guifrom the compiled.jar-file
- correction in legacy function
- correction in metadata editing test
- allow compilation without
.git
- validation rules (if a list of conditions applies to word than another list of conditions must apply too)
- new tests
- more tests for value access
- added compatibility test
~in addition to strict comparison with= - validator shortcut changed from
=to! - check for absent features/misc-keys
- extension to mass-edit/complex search&replace: possibility to search heads/childs etc with same Feature value or same UPOS etc
- new tests
- new option --shortcutTimeout to set the maximal time (milliseconds) between to keys of a shortcut sequence
- shortcuts sequence must be taped with max 300ms between keys. This permits having short and long shortcuts (e.g. A, and AV)
- bug correction in complex search: child(child(...))
- more file statistics, doc updated
- search&replace save problem corrected, doc updated
- add feature/values in statistics, update numbers if file was edited
- show CoNLL-U file statistics (sentences, tokens, words, MWTs, UPOS, deprels)
- in totally empty features/misc field in editor is replaced with
_ - test extended and new test
IsMWTandIsEmptyin complex search fixed (issue #16)- new tests
- make
transliterate.pymore robust against invalid CoNLL-U files - shortcuts to scroll the tree that a given token is visible in the current viewport
- optional json output for ConlluComporator
- test corrected
- use commons-cli to parse command-line options for ConlluEditor and ConlluComparator
- new tests
- corrected confusion: ConlluEditor deals with MWT (multiword tokens) and not with _MTW_s...
- added a script to find similar or identical sentences in a single or multiple CoNLL-U files
- updated versions of dependencies in
pom.xmlto latest version available
- possibility to modify
deprelalso in the word edit menu
- mass editing
- error corrected
- condition
Emptyrenamed inIsEmpty, new conditionsIsMWT,HeadID,EUD - new changeable colums:
Head,EUD - new columns to get data from:
HeadId - new tests
- add/edit sentence metadata (
sent_id,newdoc,newpar, transliteration, translations) - button to initialise sentence metadata
translitfromMISC:Translitfield if present - script to add an initial version of Form transliteration (
MISC:Translit)
- adding grammar for replacements (and changed format for search and replace)
- correction with subclass issue
- adding a complex search mode (syntax as in mass editing) and search-and-replace
- mass-editing:
- allow brackets in feature names like
Number[psor]=... - delete features if value is empty
- allow brackets in feature names like
- parser client configuration: accept definition of headers
- Lemma: and Form: can have filenames as argument
- added
Misc:...to search and replace tool
- Lexer/Grammar based search and replace
- check for missung root nodes, cycles and invalid head ID's and list errors when starting server
- new tests
- help page updated,
_key deletes features of active word
- shortcut display (key
?) corrected
- added subtree search (in ConLL-U or sd-parse format)
- importing subtrees of a given word
- tests
- sort features case insensitive in all cases (even when duplicating a token)
- bug in maven tests corrected (using
--saveAfter 1in tests to have output file written immediately)
- shortcuts can have any length, shortcuts to set Feature=Value pairs
- new tests
- new default for saving: without the option
--saveAfterthe conllu file is saved whenever a new sentence is chosen (next, prec, ...) the old behaviour (save every modification immediately) can be achieved using the option--saveAfter 1 - shortcuts can be defined for two keys
mod compose startid length [form]optional form of the new MWT
- correction: replace
\ninMISC:SpacesAfter=by space in# text =metadata line
- options
--featuresand--deprelscan read official specifications for valid features and deprels in https://github.com/UniversalDependencies/tools/tree/master/data (features.jsonanddeprels.json) - new option
--language(to read the feature/deprel definitions for a given language (if not given, use only universal features/deprels) - new option
--include_unused(use features declared as unusued for given language)
- help text slightly corrected
- process
SpacesBeforecorrectly - junit 4.12 -> 4.13.1
- add
--compareoption to Docker image - make
MISCautocompletion work when editing MWT - new command to transform existing word in MWT
- new tests
- Warning improved
- bug corrected in MWT creation (delete Space(s)After from MWT member tokens and add value from last memeber token to the MWT token
- put concatenated forms of created MWT to the FORM column of the MWT
- Error if edited file is not controlled under git AND there is already a backup-file (.2) in order to avoid overwriting the output of a preceding editing
- bug corrected when inserting new MWT and serializing MWTs with Space(s)After to CoNLL-U
- search: find word on multiple criteria
- error corrected for sentsplit of sentences with enhanced dependencies and/or empty words
- display roots of enhanced deps (in flat view)
- new test
- add and delete empty words (enhanced dependencies)
- delete normal words
- deleted causes for warnings
- table edit form
- case insensitive ordering of features (names) in CoNLL-U
- allow comma seperated feature value list
- new option
--featuresto specify a list of files which define valid feature=value pairs and lists of valid features for a given upos - autocomplete for feature=value pairs and misc entries
- multi word token larger, bug for variable size corrected
- fixing issue #7 (if two words overlapping with a MWT are joined, the MWT will be deleted)
- fixing issue #6 (joining/splitting sentences sets correctly
# text =metadata) - new tests, tests updates
- option "auto-adapt" which sets the horizontal width of words to the size needed
- added test whether we are on MacOS in start script
- test error corrected
- first (partial) implementation of CoNLL-U plus format. Only conllup files with 10 standard columns and additional columns can be edited
- correction for displaying data in columns > 10
- adding tool to convert any CoNLL-U/CoNLL-U Plus in any other
- deal correctly with very big number values in
MISCcolumn (likeLNumValue=4000000000)
- new field to edit MISC column for MultiToken words
- new option
--compareto compare a CoNLL-U file with the currently edited file (shown in tree and hedge mode)
- search field for morpho syntactic features
- highlight found words according to column prefix used in "find any" search (f,l,u,x,d)
- clear button for search fields
- bug correction in "find any" search
- search for sentence id
- getLaTex optimized, tests updated
- read/write correctly sentence initial empty nodes
- legacy code deleted
- new functionality: Parser ConLL-U front end in order to see the CoNLL-U result as dependency tree (or hedge)
- accept d: in 'any' search field
ConllSentence.toString(): add final empty line
- bug corrected: missing write
- bug corrected: NullPointerException at multitokens created with "compose "
- factorisation in ConllWord class
- added shortcuts: to go to next (
+) or preceding (-) sentence, run validation (=)
- layout, save button only visible when server is started with
--saveAfteroption
- add JSON export (following Spacy)
- modification to make ConlluEditor run with Edge
- layout regression in 1.14.0 corrected
- added shortcuts to change UPOS or deprel with click + key (hardwired defaults, gui/shortcuts.json file to personalise)
- add xpos and extra columns to LaTeX output
- can display columns > 10, if present (new button "extra cols")
- verbosity in test
- tests updated for the extra columns
- corrected bug for right2left display of sentences with empty words
- bug in tests (without git) corrected
- conlluedit.sh accepts filenames with spaces
- relaxed mode (corrects some formal errors in conllu-file)
- show linenumbers in error messages
- layout problem corrected
- searching chaines of deprels (up and down) corrected
- new bug from version 1.12.2 fxied
- bug corrected in undo/redo
- bug corrected for split/join function (if a MWT or enhanced dependency followed)
- minor layout modification
- support for enhanced dependencies (graphical editing)
- form search corrected
- doc updated
- Bug corrections:
- in flat mode: shift enhanced down dependencies if features are shown
- split misc field only at pipe and LF, not at comma
- larger modal for conllu/latex/validation output
- tests updated: added features to sentence with enhanced dependencies
- added support for an external validation script
- LaTeX output also for deptree.sty (see doc/)
- added split position parameter to word split function
- add support for creating, editing and deleting multiword tokens
- double click to edit a word (form, lemma, UPOS, ...), useful for tablets which usually do not have a control key
- button to edit comments (to edit empty comments)
- return version with /edit/validlists
- display version