Release v1.3.7

Bumped some packages to address security issues

Release v1.3.6

Fixed editing functionality for relations

Release v1.3.5

A very minor release addressing compatibility issue with django-scientific-survey.

Release v1.3.4

Minor release addressing compatibility and security issues of the open source packages that Textinator relies on.

Minor features

• Batch revisions are now exportable and searchable

Release v1.3.3

Fixed a bug so that pinned elements are now included in the saved context with a fixed divider >>pin<<.

Release v1.3.2: Added demo

Textinator now has a demo instance! 🎉 🎉 🎉

Please do not store any sensitive data on that instance though, as it will be regularly reset. Also please be kind to our servers, they are cheap and hence not so powerful 😉

Minor improvements:

• Search in the editing pane now allows looking up the exact match
• Search now also has full support for Ukrainian, you can enable it on the running instance by issuing the following command:
  bash postgres/exec_sql.sh prod postgres/docker-entrypoint-initdb.d/01-create-configurations.sql

QoL improvements:

• postgres/exec_sql.sh script now supports loading the environmental variables for both prod and dev environments

Release v1.3.1: Fixing Extras

This is a minor release, featuring fixing bugs with storing extra information optionally associated with labels and bumping django-scientific-surveys to 0.3.2.

Release v1.3.0: Maintenance Mode

Major features

• Added maintenance mode (activated via tools/maintenance.sh by passing an argument on or off)
• Added the possibility of site-wide announcements (used via admin interface)
• Data importer -- it is now possible to re-import project into Textinator using previously exported JSON file.
• Improvements to the editing pane:
  • Search allows for boolean AND clauses!
  • Additionally it is now possible to search by annotation number, only among flagged datapoints, get a random sample of already annotated cases for random checks
  • Now it is possible to enable sequential annotation mode for mass re-annotation work. Textinator then switches to the next annotation batch in line automatically without the need to click on it.
  • Search through texts is now faster

Minor features

• New project setting:
  • allow_post_editing -- indicates whether editing of own annotations AFTER finishing all of them should be allowed.
• Now it is possible to export flags and annotation numbers, if necessary.

QoL features

• A new data source type: interactive datasets! Type text as you go and get it instantly ready for annotation. Mostly good for the purposes of testing the annotation environment, while the data is not available yet. Maybe good for something else though? Remains to be seen :)
• The element you have just edited now scrolls into view instead of the page restarting from the very beginning.
• Flagged datapoints in the data explorer can now be searched by the content of the texts they flag.
• Dotted JSON paths (key1.key2.key3) can now be used in the key field for JSON data sources.
• Added an action button that clears all annotated labels instantly

Experimental features

• Experimental toolbox: now bits and pieces that could potentially be included as new annotation tasks or tools in Textinator will first appear in the experimental toolbox (can be enabled via "Profile" -> "Settings")

---

Fixed issues:

• Substantially improved stability while annotating nested (with more than 2 levels) and/or multi-paragraph labels.
• Fixed UI bugs to make editing board more responsive

Full Changelog: v1.2.0...v1.3.0

Release v1.2.1: Updated surveys

This is a minor release, featuring the reorganization of the code structure and bumping django-scientific-surveys to 0.3.0.

Release v1.2.0: Reviewer #2 suggests making charts larger

Major features

• Annotators can now revise the annotations of other annotators (if allowed in project settings)
  • Every revision is saved as a separate entry, linked to the original annotations. Only the revised fields are stored to optimize the storage space.
• Improved editing pane
  • All previously submitted annotations are now searchable in the editing pane.
  • The annotations that are believed to have been saved incorrectly can now be flagged.
• Charts in data explorer are now being handled asynchronously on back-end as well using Redis - Celery configuration. No more hanging pages while someone else has opened data explorer!
• Updated the survey functionality:
  • sliders can now be used as question types
  • each category in the survey can used as an independent sub-survey, allowing, for instance, to create a number of batches of questions, assign each batch to its own category, and then assign each category randomly to a participant

Minor features

• New project settings:
  • allow_reviewing -- this is the setting that turns on reviewing possibility
  • editing_title_regex supplies the regular expression to be used for searching the annotated texts and using the first found result as a title of the batches to be edited
  • editing_as_revision basically disallows modifying the data directly. By default editing happens directly in the annotated objects. If this setting is turned on, the original objects will remain intact and separate reivison objects will be created.

QoL features

• The chart that provides statistics on the lengths of inputs and labels will now only count the meaningful ones, excluding the inputs with fixed pre-defined length, e.g., radio buttons, checkboxes

---

Fixed issues: #8 , #9 , #10 , #11 , #12 , #13 , #14 , #15 , #16 , #18 , #19
Full Changelog: v1.1.0...v1.2.0