Skip to content

Latest commit

 

History

History
566 lines (435 loc) · 21.4 KB

CHANGELOG.md

File metadata and controls

566 lines (435 loc) · 21.4 KB

Changelog

Issue numbers (e.g., #43) refer to GitHub issues: https://github.com/nilsreiter/CorefAnnotator/issues

2.1.1

  • Fixed a bug that prevented flags from visible if there are a lot of them #387
  • Fixed a bug that made flags undeletable #388

2.1.0

  • New: Multiple files can be exported at once into CSV format #370
  • Fixes an issue that made deleted annotations to re-appear if the file is re-opened #380
  • Fixes an issue that caused a freeze when deleting an annotation #381
  • Updated some dependencies

Thanks Börge for taking care of these!

2.0.1

  • Fixes an issue that prevented expanding the tree properly in some imported files #373
  • Fixes an issue in the QuaDramA/TEI importer #374

2.0.0

  • New: Discontinuous annotations. A single mention can now consist of multiple spans. Mention parts, that have been used for relative clauses and such are gone. Add a new span to a mention by dragging it onto the mention in the tree. #293, #104
  • New: Analysis functions. From the main window, files can now be opened in the analyzer view, which offers visualizations and some statistics for the annotations. #292
  • New: Annotation guidelines can now be linked via a profile. They appear directly in the Help menu. #12
  • Files are now saved with a specific file extension to avoid confusion and accidental decompression. The file extension is .ca2 for uncompressed, git-friendly files and .ca2z for gzip compressed files. Compression
    makes the files roughly ten times smaller. #335
  • Search is now case-insensitive by default #320
  • Various UI tweaks
  • Main window now shows a section if a version is opened for the first time
  • Updated Dependencies
    • log4j: 2.12.1 -> 2.13.3
    • guava: 28.1-jre -> 30.1-jre
    • reflections: 0.9.11 -> 0.9.12
    • mockito: 3.1.0 -> 3.4.4
    • dkpro: 1.9.0 -> 2.1.0
    • uima: 2.10.2 -> 3.1.1
    • eclipse collections: 9.2.0 -> 10.2.0
    • javafx: 13.0.1 -> 14.0.2.1
    • ikonli: 2.1.1 -> 11.5.0
    • poi: 4.1.1 -> 4.1.2

1.15.2

  • Command line tool ExportMultipleFiles no longer tries to descend into sub directories #367

1.15.1

  • Files without segments no longer show empty parentheses in tree view #311
  • Save and close now works properly for imported files #357

1.15.0

  • The entity tree can now also be sorted based on last change (to the entity). Thanks @andreasvc! #153 #349
  • The maximum length of displayed entity labels is increased to 100 characters. This is not a long-term solution though. #353
  • Pressing cmd-t now displays a mention in the tree, if a) the cursor is placed in a single mention or b) a single mention is selected exactly. #351
  • Fixed (some) undo behaviour when de-emphasizing singletons #347

1.14.5

  • Fixed behaviour of de-emphasising singletons. Thanks @andreasvc #347
  • Fixed upating of status bar if selection changes #350
  • Fixed the bug that prevented automatically removing empty entities #352
  • XMI files are now pretty printed, making them more git-friendly. Thanks @andreasvc #339

1.14.4

  • Fixed a bug caused by old flags no longer being generated automatically #340

1.14.3

  • Added that wide-spaced text is now displayed with a serif font #315
  • Fixed a bug with the alignment of line numbers #316
  • Fixed a bug that prevented formatting info (italic etc.) to be displayed correctly in the SFB1391 importer #314

1.14.2

  • Fixed a bug what caused the auto-save to continue saving a file even after the document had been closed #309

1.14.1

  • Fixes a bug that prevented the export of full mention tables to CSV or Excel #301
  • Export CSV/Excel files now contain segment information (e.g., strophes) if included in the file #306
  • If no predefined line numbers are present, we fall back to calculated line numbers #303

1.14.0

  • Added a new preference called sticky markers. If enabled, changing the markers of any mention also changes the markers of all mentions that cover the (exact) same span. Thus, the markers are "sticky" to their position. New mentions in other entities do not automatically get all the flags, but it's now possible to duplicate mentions and then move them to a new entity. #271
  • Added a new view to show simple summary statistics, and an exporter that exports the same info as CSV file. #286
  • Automatic saving every 5 minutes #294

1.13.0

  • CSV export has been improved in multiple ways
    • Can now be done for the entire file
    • It's possible so specify lines as context #261
    • Line numbers can be included in the export #268
  • Line numbers in which mentions appear can now be shown in the tree #266
  • Segment annotations (e.g., stanzas or chapters) can now be displayed as a table of contents on the left of the text. The TOC is fully interactive and can be used to navigate in long texts i#263
  • Compare view now supports comparing the actual entity assignment. This behavior can be enabled via the settings menu is only available if both files are loaded with the same profile, which also means they have to be in the same directory. #272 #267
  • Opening multiple files at once is now much faster by using the default dialog for opening files #275
  • Opening or importing a new document should no longer auto- create the coreference-related flags ambiguous, difficult and Nonnominal. #277
  • If line numbers are displayed in the tree, and a segmentation exist, the segments are also displayed in the tree. #279
  • Added a new exporter that directly exports into MS Excel and is more configurable. #285

1.12.2

  • Fixed the bug that the compare view hid some annotations if more than two annotators were compared, and more than two, but not all of them had annotated something. #279

1.12.1

  • Line numbers are now in sync with the lines they number #273

1.12.0

  • The app now includes a command line tool to convert multiple files at once. #242
  • Added a SFB1391 specific importer than interprets <lb> tags properly #255
  • Export mention statistics now has some configuration options. Import and export plugins can have that as well now. #76
  • Sorting of entities in the tree is now case insensitive #257
  • Shortcut keys are now displayed in brackets #256
  • We can now specify a profile when loading the application #249. The profile is stored in an XML file called profile.xml in the same directory as the XMI file to be loaded. The profile can
    • pre-define flags and entities
    • disallow operations
    • set configuration options
  • Singletons can now be displayed in light gray #177

1.11.2

  • Using the QuaDramA/TEI importer no longer generates files in the old format #246
  • Flags can now be deleted properly #247

1.11.1

  • Entities in tree are now sortable again #241

1.11.0

  • JavaFX is now optional. If JavaFX classes can't be loaded, the application uses file open/save windows from classic swing. #220
  • The app now compiles with OpenJDK properly. Tested using OpenJDK 12 #237
  • Updated guava from 28.0-jre to 28.1-jre #236
  • Updated log4j-core from 2.12.0 to 2.12.1 #234
  • Updated log4j-api from 2.12.0 to 2.12.1 #233
  • Updated mockito-core from 2.27.0 to 3.0.0 #229

1.10.0

  • With a new toggle setting, assigning a mention over an already existing mention replaces the annotation #149
  • The status bar at the bottom now shows the entity name if the cursor is within a mention of an entity. Thanks @andreasvc for the initial code contribution! #189
  • Line spacing can now be set explicitly in the View menu #226
  • Adjacent mentions that belong to the same entity can now be merged with a single action. Two mentions are considered adjacent, if there is only whitespace between them. #218
  • Entity label in context menu now has limited width #232

1.9.6

  • Improved painting of underlinings for long annotations and a lot of overlapping annotations #225

1.9.5

  • Fixed the loading of GerDraCor files #219
  • Fixed that the entity tree is not updated when mentions are moved #217
  • Entities can be hidden again properly #216

1.9.4

  • Improved stability of the grouping mechanism. #210

1.9.3

  • Fixed a bug that prevented the display of the file open dialog #207
  • Deleting many empty entities no longer creates an infinite loop on OpenJDK #192
  • Less console output and file logging (but can be activated if needed)

1.9.2

  • CoNLL export now also works if no entities have been annotated. #191
  • Fixed a bug that prevented mentions with the same begin position to be underlined. #186
  • Exporting to CoNLL format now shows the proper file extensions and allows overwriting. #187

1.9.1

  • The shortcut key to jump to mentions has been changed to alt-arrow, because shift arrow collided with extending the selection by keyboard. #185

1.9.0

  • The CoNLL 2012 importer now always reads coreference information from the last column. Thanks @andreasvc. #127
  • The README now contains a section on requirements and installation, specifying the JavaFX dependency explicitly. Thanks @andreasvc. #174
  • When searching an entity in the tree view, the tree now scrolls to top automatically. #178
  • Using shift-arrow, the user can now jump from one annotated piece to the next. This function does not entirely iterate over mentions. If two mentions have the exact same boundaries, pressing the combo again jumps to the next mention with different boundaries. #182
  • Entities can now be automatically renamed. The new name is derived from the first, last, or longest mention. This is decided ad hoc in a dialog. The renaiming is fully undoable, so users can try with different options. #173
  • Pressing backspace while a mention is selected removes it. Multiple mentions with the same span are deleted as well. ctrl-backspace deletes all mentions in the selected area. #175
  • The label of the entity of a selected mention is now displayed in the status bar #176
  • Some formatting information can now be read from TEI/XML files and will be displayed accordingly. All <head> elements are shown large, everything with a rend=italic or rend=bold will be display as such. In addition <emph> is rendered as italic. <lg> and <div> elements will be marked as segments (and if they have <head>, they're displayed in the scrollbar). #181

1.8.0

  • It's now possible to create new flags that can be assigned to entities, mentions, and detached mention parts. The flags can be created with the flag editor (in the tools menu). #126
  • The text view can now display line numbers. See the wiki for details. #169
  • The text field in the search dialog is now automatically in focus when the window is opened #159
  • The text field now maintains previous searches #161
  • The search can now be restricted to mentions only #162
  • We now use a JavaFX-based file dialog, which should look more native to the OS (no issue)

1.7.4

  • If dragging, the entity mention tree no longer reacts on moving over mentions, i.e., the text no longer jumps to the mention if something is dragged #170
  • Moving a mention into an entity group was leading to file corruption. Files could no longer be opened. This is fixed, and corrupted files can be opened again. #171

1.7.3

  • Renamed an entity is now properly undoable #167
  • Setting the language now actually works #168

1.7.2

  • Fixed a bug that prevented loading files when run on OpenJDK #166

1.7.1

  • Fixed a bug that changed the entity name to null if its' properties are changed #163
  • Fixed a bug that showed a wrong icon when editing the entity name #164
  • It's no longer possible to rename mentions

1.7.0

  • If a group contains more than two entities, the number of additional entities is now displayed in the label #156
  • Long entity names are now abbreviated, but tooltips show the full name
  • Entity names can be edited by double clicking on it in the tree view #152
  • Deleting a mention no longer collapses the entity tree, even if the tree is automatically sorted by mention count (in this case, the expanded sub tree may be moved). #155

1.6.1

  • Fixed a bug with the export to TEI. XML ids may now include special characters #157

1.6.0

  • Removed commenting system #130
  • Context menu in the text view is now flatter #146
  • Clicking on an entity now highlights all mentions of it #142
  • If segments are annotated in the XMI file, they can now be displayed in the scroll bar. This is a preparation for segment support, and not yet directly usable. #84
  • The search panel now has a second tab that allows searching for mention flags #42

1.5.6

  • Fixed an issue that prevented merging compressed files #150

1.5.5

  • Fixed an issue that resulted in copied mentions #141

1.5.4

  • Fixed that creating new documents from plain text (and others) did not set the correct file version number #137
  • Fixed that the style is not correctly set when importing QuaDramA files #139
  • Fixed that hidden entities also have their properties in gray #133

1.5.3

  • Fixed a bug that caused detached mention parts to disappear after conversion #131
  • Fixed an issue that caused entities to be disrupted after conversion #129

1.5.2

  • Fixed an issue that prevented loading of the correct style

1.5.1

  • Fixes that hiding an entity didn't de-underline its mentions #127

1.5.0

  • Searching in the tree view now behaves slightly different: If a new mention is dragged in (or added a key) while the tree view is ranked according to the search, the ranking remains after having added the mention. If the user presses enter, the ranking (and search) is removed. #100
  • Added a new command to automatically remove all singletons. Empty entities are also removed at the same time. #103
  • If a text segment has been selected, clicking on the context menu shows the option "copy formatted example". This copies the selection with their markdown or plain text formatted annotations onto the clipboard. Currently, the markdown format uses to subscript, but other formats can be included. #110
  • New type system: This version introduces a modified type system, which entails that the file format has to change. At this point, the changes are not significant, but provide the stage for later changes. When an old file is opened for the first time, it gets converted automatically. #101
  • The user can now control whether singletons are ignored for comparing two annotation files. #106
  • Compressed files: If the file name entered in the Save As dialog ends on ".xmi.gz", the file will be gzipped and therefore take much less space on disk (and is more email-friendly). Files ending on ".xmi.gz" can be loaded directly as usual. #125
  • QuaDramA/CRETA-specific importers that we used to import files from WebAnno have been removed. Import (of coreference annotations) from WebAnno can still be done using the dkpro importer. #124
  • JSON export: Coreference annotations can now be exported as JSON files. The file contains a list of tokens, and each token can be associated with one or more entities. This JSON format will be importable by rCat in the future. #120

1.4.4

  • Fixed an issue in the compare view that led to displaced underlines. Root cause was the handling of windows-style newlines \r\n #114
  • Fixed an issue in the TEI importer that caused the document title to be wrong #119

1.4.3

  • Fixed an issue with the TEI/XML handling. The base TEI IO plugin now only handles elements with @ref attribute, and does not read or write @xml:id attributes. This is in line with the documentation of the @ref attribute, which is not restricted to XML IDs. #118

1.4.2

  • Fixed that detached mention parts were not included in the tree after loading a file. #111
  • Fixed that warning dialogs were displayed even if all changes were saved. The downside is that saving now clears the history. #109
  • Delete everything now deletes everything. #97
  • Fixed that the settings menu would disappear on Windows if multiple windows are opened. #113

1.4.1

  • App now launches on a Mac with Java 9 and higher. #71
  • Mentions can now be moved into an entity group properly #108

1.4

  • Selecting text in the compare view now displays agreement calculated in this section on the right #79
  • Annotated files can now be merged. This can be used if different annotators have worked on different parts of the text. The merge is done as is -- entities need to be merged by hand after the documents have been merged. #73
  • Duplicate mentions (covering the same span) can now be deleted automatically. This is done via the context menu for entities and should be useful after files have been merged. #81
  • Mention export: Mentions of selected entities can now be exported as a CSV file. The CSV file contains the surface, the span (character positions), the mention label and number, all mention and entity properties (ambiguous, generic etc.). This can be used for quick and dirty statistical analysis of an entity (R-script available). #94
  • Entity groups can now be formed from more than two selected entities. Only the first two will be used to create the label. The 'and' in the label is now translatable. #96
  • Automatic support: The tools menu now contains a sub menu with automatic helpers. At the moment, StanfordNER can be run on a number of languages. See help window for details. #89
  • Fixes that the first mention of a chain was missing when importing from TEI #92

1.3.3

  • Fixed a bug that caused loading to crash if in previous runs, entities had been removed under certain conditions. Files can now be opened, but group members may be missing. #93
  • Fixed a bug that could cause subsequent loading issues. If an entity was part of a group, and then removed entirely, a null-reference would appear in the group and causing loading to crash. #95

1.3.2

  • Fixed a bug that caused saved files to be written in another file name after export #90

1.3.1

  • Fixed a bug that caused multiple removal actions to be individual undos #78
  • Fixed that the TEI import did not include the header #80
  • Fixed a bug that prevented the removal of an entity from an entity group to be undoable #82
  • Fixed: Add buttons in search panels are now inactive when now findings are selected #83
  • Attempt at fixing a crash bug in the search panel #70

1.3

  • Add TEI import plugin. In TEI files, everything that as a ref attribute is considered a mention. Elements with who attribute are treated specially, because dramas (i.e., the sub-ordinate speaker- element is considered the mention). #72
  • Undo support. Pressing ctrl/cmd-z undos the last action. The history is saved as long as the document is left open. #65
  • Generation of candidates. Clicking on a selected text shows the context menu, which now contains a list of candidate entities (as well as the option to create a new entity) #58
  • Compare view. A new view allows comparing the works of two annotators. This view only compares the spans of the mentions -- there is currently no way of comparing the actual entity assignments. #69.

1.2.1

  • Application now also starts on Windows and Linux #68

1.2

  • Findings can be dragged from search window and dropped onto the entity tree #49
  • Main window informs if new version is available #60
  • Annotation also works by dropping on any already annotated mention within the text view #1
  • Document language can now be set before or after importing
  • Post-correction of automatic CR: #43
  • We can now import CoNLL 2012, which is generated by IMS HotCoref DE
  • English texts can be run through Stanford Core NLP CR via an import plugin
  • Selected text can now be copied to system clipboard #61
  • Groups can be created via context menu
  • Groups have a better icon
  • Arbitrary segments of the document can be selected and commented (via the menu bar). Existing comments are visible in the text and in a comment viewer, where they can be edited and deleted. #25
  • bug fixes
    • First/last token can now be properly annotated #67

1.1.2

  • Fixed issue that files were not saved properly because overwriting didn't work #63
  • Fixed issue that button in search panel didn't became active after selecting an entity in tree #59

1.1.1

  • Fixed shortcut window #56
  • Add possibility to mark mentions as non-nominal #55
  • Fixed merging of multiple entities #54

1.1

  • If multiple things are now selected in the tree, they can be deleted at once
  • Entities can now be merged. All mentions of the smaller entity are moved to the larger one, and the smaller one gets deleted #36
  • Increased line spacing to improve readability #35
  • (re-)Sorting the tree while searching is now much faster #40
  • Export to CoNLL 2012 #37
  • It is now possible to mark findings in the search panel and annotate them at once, either to an existing or a new entity #41
  • Multiple things can be moved at once #33
  • Entities can be hidden, such that their mentions are no longer underlined #46

1.0.4

  • Fixed that multiple entities could be unintentionally merged when exporting to CRETA/Adorno #48
  • Fixed that original annotations (that have been imported) are still present in the export when using CRETA/Adorno #47

1.0.3

  • Fixed that the export dialog was not shown #44
  • Fixed a bug preventing application launch #45

1.0.2

  • Fixed a bug in which some underlines were printed on top of each other #38
  • Fixed a bug preventing loading of texts if a title was not set for the document #39

1.0.1

  • Fixed a bug when loading DKpro formatted files (including WebAnno) #32

1.0

  • First release