5.14.0

What's Changed

Added

• Added a link to the I-analyzer user survey (#1682)

Changed

• Updated repository URL (after migrating to the CentreForDigitalHumanties organisation) (#1685)
• Overhaul GitHub actions once more (#1678)

Fixed

• Fixed sorting by date in the Times corpus (#1686)
• Fixed visualisations by date in DBNL corpus (#1684)
• Fixed field definitions in some parliamentary corpora (#1677)

Server configuration

• For text fields, the term_vector mapping parameter in Elasticsearch must be set to to with_positions_offsets or with_positions_offsets_payloads. This mapping was already the default, but this version drops support for other configurations (#1677)

---

Contributors: @BeritJanssen, @lukavdplas

Full Changelog: 5.13.0...5.14.0

5.13.0

Added

• Jewish Migration histogram visualizations (#1669)
• Jewish Migration: update data from remote url (#1654)
• Named Entity documentation (#1637)
• European Parliament corpus (#1631)
• Word model paths for Times and Guardian-Observer corpora (#1664)
• Neighbouring word visualization in various corpora (#1674)

Changed

• Improve fetching of map data (#1657)

Bugfixes

• Fix crashes of indexing operation (#1658)
• Fix versioning of Elasticsearch indices (#1660)
• Fix check for availability of Named Entities (#1667)

Security updates

• Improved registration security and validation (#1666)
• Add registration and password reset rate limit (#1673)
• Library updates (#1661, #1662)

Contributors: @BeritJanssen, @ar-jan
Full Changelog: 5.12.0...5.13.0

5.12.0

Added

• Map visualization using [VEGA{(https://vega.github.io/vega/) https://github.com//pull/1629
• Versioning commands #1636
• Corpus documentation editing through API #1640
• Frontend import shortcuts #1643

Changed

• Move download functionality to separate tab #1635
• Cache GitHub actions #1619
• Updated XML corpus definitions #1556
• Various accesibility improvements #1644 #1645 #1646

Fixed

• Resolved a bug with query comparison #1602
• Resolved a bug with NER fields #1642

Security updates

#1653, #1647, #1652, #1641, #1628, #1627, #1617, #1608

Contributors: @lukavdplas, @BeritJanssen, @JeltevanBoheemen, @ar-jan

Full Changelog: 5.9.0...5.12.0

5.10.0

What's Changed

Interface changes

• Downloading search results is now handled in a separate tab (#1635)
• Resized the image for the Times corpus (#1632)

Code quality

• Cache Docker images in github actions (#1619)
• Updated definitions for XML corpora (#1556)
• Improved commands for updating the version in the source code (#1636)

Contributors: @BeritJanssen, @JeltevanBoheemen, @lukavdplas

Full Changelog: 5.9.0...5.10.0

5.9.0

Added

• Named entity highlighting in search interface
• First version of JSON corpus interface

Fixed

• Fixed a problem where unauthenticated users could not access functionality they should be able to.
• Fixed various minor bugs

Full Changelog: 5.8.0...5.9.0

5.8.0

Added

• Added U-Blad corpus to the application
• Show navigation links for users that are not logged in

Fixed

• Resolved a bug in the ngram component that disallowed compensation selection

Changed

• Migrated to Angular 17
• Changed manual layout
• Documentation on first time setup
• Bump braces from 3.0.2 to 3.0.3 in /frontend by @dependabot in #1597
• Bump tornado from 6.3.3 to 6.4.1 in /backend by @dependabot in #1595

Full Changelog: 5.7.0...5.8.0

5.7.0

What's Changed

Updates to corpora

• Enabled sorting results by date in the Guardian/Observer corpus (#1577)
• Added citation guidelines for Times, Dutch Newspapers (public), and People & Parliament (Netherlands) corpora (#1572, #1548)

Accessibility

• Page titles are set appropriately (#1591)

Bug fixes

• Fixed the wordcloud not loading for fields with dynamic language value; fixed the wordcloud table not loading (#1568)

Preparation for planned features

• Database-only corpora can now be imported and exported as JSON definitions through the backend API (#1547, #1569, #1571)
• Enabled indexing data for database-only corpora (#1567)

• In the specification for a date or range filter, the upper/lower bounds are now optional (#1546).
• Add Elasticsearch annotated-text-plugin to Docker configuration (#1586)

Code quality and CI

• Removed unused properties in frontend corpus class (#1587)
• Reorganised and corrected syntax in Github actions (#1585)
• Centralised corpus definitions for unit tests (#1544)

Contributors: @BeritJanssen, @JeltevanBoheemen, @lukavdplas

Full Changelog: 5.6.2...5.7.0

5.6.2

Security fixes.

Full Changelog: 5.6.1...5.6.2

5.6.1

What's Changed

Documentation

• Added documentation about loading corpora before indexing (#1555)
• Added documentation on Django project settings (#1530)

Bugfixes

• Fixed an issue where field coverage did not load for anonymous users (#1552)

Preparation for planned features

• Added module to parse a JSON corpus definition (#1519)
• Added module to get the source data reader corpora without a Python definition (#1541)

Changes in server configuration

The data_directory for a corpus is now validated when the corpus is loaded into the database. It must either be empty, or an existing directory. If the server configuration specifies nonexistent directories, they should be removed.

---

Contributors: @JeltevanBoheemen, @tijmenbaarda, @lukavdplas

Full Changelog: 5.6.0...5.6.1

5.6.0

What's Changed

New features

• Users can now access I-analyzer without logging in (#1484, #1498, #1539)
• The site footer now displays the version of I-analyzer (#1479, #1492)
• Added new options for corpus documentation: citation, licence, and terms of service (#1488)
• Added a tag overview menu (#1463)
• Updated the layout of document previews in search results (#1476)

Accessibility

• Standardised keyboard controls for dropdown menus (#1465)
• Set the lang property of a field's content based on the corpus definition (#1335)

Updates to corpora

• Various corpora: renamed some fields named "URL" for clarity (#1491)
• People & Parliament Ireland: enable adding word2vec models (#1490)
• Jewish Migration: allow loading local source data (#1510)
• DBNL: fixed the chapter order when opening "view book" (#1503)

• Updated images for several corpora (#1485)

Bug fixes

• Fixed paragraph breaks not showing up in the document view (#1477, #1511)
• Fixes to controls for viewing scanned documents (#1493)
• Fixed rendering of geopoint array (#1444)
• Fixed the layout of barchart options menu on mobile devices (#1487)

Code quality

• Avoid repeated downloading of NLTK stopwords (#1481)
• Update elasticsearch in docker configuration (#1482)
• Refactored visualisation selection menu (#1467)
• Added backend dependence on elasticsearch service in docker configuration (#1471)
• Refactored wordcloud component (#1475)
• Forbid 'scan' and 'tab-scan' field names in corpus definitions (#1496)
• Updated documentation about Python versions (#1499)
• Default sorting is now a property of a Corpus object, rather than its fields (#1501)
• Updated the file structure in the addcorpus module (#1502)
• Improved corpus validation: distinguish validation at indexing and publishing stage (#1504)
• Retrieve wordcloud stopwords from the field language, rather than corpus settings (#1514)
• Allow checking whether a corpus has a Python definition (#1515)
• Corpus images and documentation are now saved in the database (#1526, #1531)
• Added workflow for updating CITATION.cff (#1534)

Preparations for planned features

These changes add functionality for planned features.

• Added component for map visualisation (#1507)
• Added schema for JSON definitions of corpora (#1497)

Changes in server configuration

These changes are required when updating a server to this version:

In the backend settings.py:

• configure the MEDIA_ROOT setting. Corpus images will be saved here.

In the frontend environment.ts:

• add directDownloadLimit property (set to 1000 to keep the same behaviour)
• add version and sourceUrl properties. See the development environment for an example.

Contributors: @BeritJanssen, @JeltevanBoheemen, @ar-jan, @lukavdplas

Full Changelog: 5.5.1...5.6.0