Tags give the ability to mark specific points in history as being important
- Sort by
- Name
- Oldest updated
- Last updated
-
v1.6.3Release v1.6.3
- !335 Update docker tag in user doc: This patch fixes the tag used in the explanation of a C++ algorithm creation. Fixes #552
- !336 Add additional validation for environments/queues: When creating a new experiment, there's an additional check that is done for the validity of the environments/queues combination selected that is performed on save. This currently does not play nice with the API side as it triggers an error 500 which should rather be a validation error. Therefore the check is now done in the validate method so that this use case is better handled and allows for a correct error message to be returned. Also, now if we get an error when creating a new experiment, an APIError is generated with it.
- !337 Fix analyzer edition: The online editor still allows to have parameters for analyzer which is something that has been removed since they should not have any. This patch fixes this.
- !338 Add proper file api endpoint for cxx: This merge request implements a new endpoint for the upload and download of the "code" for binary algorithms. Currently C++ algorithms. The current method used by the website is not the most cleanest implementation and should be replaced by the endpoint in the long run. This will also allow to properly push and pull from beat/beat.cmdline> As a drive by fix, it also add the missing writing_serializer_class variable to the RetrieveUpdateDestroyExperimentView. Part of beat.cmdline#73
- !339 Cannot show environment with plus sign in name: This merge request extends the URL pattern for the environment detail view so that the C++ environment is properly supported. Fixes #556
- !340 Fix spur put experiments: This merge request fixes an issue appearing only with Safari that can send a PUT command while at the start doing a GET. That PUT call does not go through but still does enough to trigger an error since the writing serializer for the experiment is set to none as the PUT is currently custom handled. Adding a dummy serializer here allows for the call to succeed. Fixes #557
- !341 Update beat packages versions: This merge request updates the versions of beat packages used on the platform and for the editor.
-
v1.6.2Release v1.6.2
- !333 Update execution environment: Move all the execution environment definition to the new one provided by beat/beat.env.builder> This not only concerns the tests but also the qsetup script which is used to prepare the test environment. Follows beat.core!120, beat.backend.python!72 and beat.cmdline!95
- !334 Activity stream notification for supervision requests: This MR targets the notification handling for new supervision requests in the activity stream. The goal is to help supervisors notice in their activity stream (when connected) any supervision request he needs to address. Fixes #496
-
v1.6.0Release v1.6.0
- !308 [docs] added validation prodedure for web interface Bug 485 to the admin documentation: This merge request adds information given in beat/beat.web#485 about validation procedure for web interface for web admin documentation.
- !309 Fix all the beat dependencies used for CI testing to match the versions used by buildout: Since beat.core is fixed at 1.8.0, the others must also be pinned to their corresponding versions. See release.cfg
- !310 removing Biometrics Evaluation from BEAT definition as the platform has become more generic than that.
- !311 Don't make the test database in memory for beat.cmdline tests. This allows for re-use during the tests rather than recreate everthing from scratch every time.
- !312 Postpone date of users validation and clean up user profiles.
- !313 Patch the cleaning of invalid users: This patch avoids the crash of the cronjob during the cleanup of invalid profiles, depends on !312
- !314 Cleanup model representation: This patch updates the models to show proper string representation with Python 3. This is useful for getting information on Django's admin pages
- !315 Remove google analytics: Fixes #543
- !316 Remove the beta channel from the list in dev.yml: dev.yml is used to test against the latest stable version of the beat packages. The CI build system now uses the new strict channel policy of conda which means that it does not search further after a package is found in a channel. Thus if the specified version is not found, it ends up with an error. See !315 for details
- !317 Improve database map command output: The scripts now lists by default only the users with an active account. There's an option that will print all users with their activate state. The user email has been added to the output.
- !318 Remove .conda and .condarc from docker image when doing deployment: The current conda-concourse-ci image includes wrong configuration
- !319 Add usage time to database: This merge request implements accessibility time handling for database objects. Without dates the database behaviour is the same as now. If dates are added they become only accessible within the range of dates. Experiment configured with these databases won't be re-run after access expiration. Part of beat/beat.web#541
- !320 Add accessibility edition to admin page: Add the missing fields for editing the accessibility date and time in the admin page. Part of beat/beat.web#541
- !321 Fix new experiment save: When pushing a completely new experiment on the platform, an error occurs when calling update_blocks on the cache content check. If the object does not exists, then an exception is rightfully raised. This patch changes that by using get_or_create like of the outputs handling.
- !322 Fix Sign-up/Sign-in form and documentation pointers: This MR fixes sign-up and sign-in forms and improves documentation pointers Fixes #544, #545
- !323 Improve experiment failure/success notification email: This MR targets the notification of experiments success/failure by having a login direction followed by a re-direction to the experiments page. Fixes #546
- !324 Return compatibility information for algorithms and libraries: This merge request implements the retrieval for execution environment used for an algorithm. It cleans the current implementation providing that information to the views and makes it also available on demand through the API. Fixes #548
- !325 Improve new version handling: This merge request refactors the new version handling API wise. It now does not guess what the new version should be. The version value shall be send along the rest of the data so the content of the platform can properly be checked. After this update people shall update to beat/beat.cmdline> version 1.7.1 Related to beat.cmdline#67 Fixes #547
- !326 Fix object existence check in common serialiser: The check worked with SQLite but not with PostgreSQL. In any case, since we have a user object here, there's no reason to not use it directly. Part of #547
- !327 Refactor update creation api: This merge request does the following: - Raise DRF errors rather than returning response to take advantage of the DRF exception handling mechanism - Create serialisers for PUT request to simplify the code - Improve validation checks in several places - Fixes several issues in DRF infrastructure uses - Fixes JSON fields handling - Remove routes that do no fully qualify objects - Implement a custom DRF exception handler that will allow to log errors - Improve testing - Improve permission management
- !328 Improve automatic emails with temporary urls: This MR targets the improvement for automatic emails concerning validation of new users by supervisors/re-validation of users. Fixes #542 Fixes #550
- !329 Fix supervision lists expand: This MR targets expansion of supervision lists and tables Fixes #549
-
!330 Fix parallel testing on ci: This merge request fixes parallel CI builds in the most simple way. For Linux host,
/dev/shmis used to store the test database and prefix to speedup tests execution. However, on the Gitlab CI, the same machine may run several builds in parallel which means that one run might overwrite data from another one. In order to avoid that two solutions are possible: - Use the database and prefix "in place" (i.e. in the same folder as the code) - Use the various Gitlab provided environment variables to build a path in/dev/shmto ensure each build has its data in its own directory The first solution was chosen has it is the easiest to put in place. - !332 Update beat.backend.python version
-
v1.5.3Release v1.5.3
- !288 Smarter access logic for reports' experiments
- !289 Don't use CDN: dependencies are already available through static files.
- !290 Don't let reports have unused experiments
- !291 [ui] Show warning about saving docs changes in reports
- !292 [reports][js] Fix empty values in tables for experiments with changed aliases
- !293 [doc] Fixed external and internal links
- !294 [reports][js] Show the proper experiments table when locked/published with only 1 group
- !295 [doc] Fixes some warning issues with the documentation
- !297 Fix report's experiment table name/URL construction & restrictions
- !298 Implement database access information print command
- !299 Fix analyzer method update in editor
- !300 Add root folder information to output of database information command
- !301 Replacing django-rest-swagger package that is deprecated with new drf-yasg package
- !302 Update beat/bob-devel to 2019.08.26
- !303 Update beat/bob-devel to 2019.08.28
- !304 Fix documentation link on icon
- !305 [sphinx-openapi] increased the version of sphinxcontrib-openapi to include new feature response object rendering
- !306 Implement experiment reset api: APIs are already available for start and cancel, this merge request adds reset
- !307 [swagger] changed the permission to access the swagger page only to registered users
-
v1.5.2Release v1.5.2
- !230 fix url prefix handling to properly construct URLs
- !283 Remove redundant name segment in dowloaded plot filenames
-
!284 Fix malformed URLs when
URL_PREFIXis set in report - !285 Ensure no URL_PREFIX is used for test
- !286 Add GPU environment to qsetup to allow proper installation and run of the GPU example
-
v1.5.0Release v1.5.0
- !263 New CI more in line with the other packages. Will use conda to build test environment.
- !267 Update bob/beat-devel dependencies to enforce the latest beat/bob-devel packages to be used for building this package.
- !268 Strip database_name from split chars to avoid empty dataset name generation and thus an invalid index error.
- !269 Fix install command documentation
- !270 Fix the ui/views.py file as it is missing the Http404 import
- !271 Make the top bar documentation link configurable
- !272 Update CI and documentation linking
- !273 Add graphviz to dev.yml for beat.cmdeline tests
- !274 Fix queryset related issue as latest versions of django rest framework makes it mandatory to either provide a queryset property or get_queryset method.
- !275 ZMQ refactoring that implements a new version of the scheduling in beat.web taking advantage of the new implementation in beat.core. There are now additional django commands that allows to manage the scheduler, broker and worker directly from there
- !276 Update bob/beat-devel to fix numpy-base issue
- !277 Fix backup/restore by implementing natural key and foreign key handling for Profile and AccountSettings
- !279 Restore command: validate sqlite version': SQLite version 3.26 introduced a change which is breaking the migration mechanism in Django thus the command will bailout since it's known to break. As an alternative to !278
- !281 Implement password reset
-
v1.4.1Release v1.4.1
- !257 Fix typos and text in ui registration templates
- !258 Set algorithm schema_version to 2 if not provided: This can happen when converting a v1 analyser.
- !259 Fix database retrieval
- !260 Fix mail handling: remove generic try except and update the template generation code to use dict in place of Context
- !262 Update docker image name
- !261 Fix email sending: This patch goes through all of beat.web mail sending and cleans them
- !264 Improve README and update dev environment: Latest update and sync with other READMEs available in the project.
- !265 Merge new documentation to master
- !266 Fix links in the documentation
- !263 Change the CI infrastructure to run the tests and build the documentation
-
v1.4.0Release v1.4.0
- !236 Fix backup and restore commands
- !238 Upgrade to django 1.11
- !239 Add cache deletion support: WARNING, use with caution, it may break experiments
- !240 Improve database handling when loading scheduler
- !241 Add exclude field in serializer for django rest framework compatibility
- !242 Refactor code to make it work with both Python 2 and Python 3
- !243 Update lambda use and key search in dictionaries
- !245 Fix anchors in documentation
- !246 Fix Python 2/3 use of zip
- !247 Add command to get current active user on the site
- !248 Fix handling of missing field in experiment panel
- !249 Fix various issues that have been found while doing tests using the web site and the online platform.
- !250 Improve the buildout handling
- !251 Fix bugs in the database API and improve the tools to set them up.
- !252 Update documentation for account reactivation for blocked users
- !253 Improve block state change
- !254 Refactor version fingerprinting for code mirror widgets
- !255 1.4.x: Merge the 1.4.x branch to make the 1.4 version the new stable release.
- !256 Fix context handling for template rendering in navigation: dict must now be used