Resolwe

===================

Changed
-------
- Do not require CSRF token to access uri resolver endpoint


===================

===================

Fixed
-----
- Correctly clear ``Redis`` cache on data restart

Added
-----
- Add ``clear_redis_cache`` management command
- Add modified field to the ``AnnotationValue`` model and expose it in API


===================

===================

Fixed
-----
- Add default value for ``FLOW_PROCESSES_ALLOW_LIST`` and
``FLOW_PROCESSES_IGNORE_LIST`` in case of missing settings.
- User defined slug must not be changed
- Raise ``ValidationError`` when trying to get non-existent ``AnnotationField``
- Propagate arguments to parent save method in ``BaseCollection`` model

Added
-----
- Add ``extend_lock`` method to redis cache in listener
- Add extending processing lock task to the listener
- Add ``FLOW_KUBERNETES_PULL_POLICY`` setting to set the image pull policy in
``Kubernetes`` instead of using ``Always``

Changed
-------
- Wrap listener command processing in a transaction
- Return ``False`` when ``has_perm`` is called with ``obj`` set no None, since
object level permissions are granted using Django ``ModelBackend``
authentication backend
- Use ``keys`` instead of ``scan_iter`` in ``clear`` method of ``redis`` cache,
since ``scan_iter`` is much slower
- Do not allow deleting annotation field if annotation values are present


===================

===================

Changed
-------
- Remove mapping of probe_ids to ENSEMBL ids and add extra variables in
``methylation-array-sesame`` process


===================

===================

Changed
-------
- Redis cache in listener is updated when data fields are retrieved from the
database
- Bulk annotations on entity endpoint now accept field path instead of id

Added
-----
- Add processes allow and ignore list to dispatcher, controlled by
environmental variables ``FLOW_PROCESSES_ALLOW_LIST`` and
``FLOW_PROCESSES_IGNORE_LIST
- Allow ovirriding the maximal number of commands listener can process
concurrently
- Suport bulk create/update/delete on AnnotationValues endpoint

Fixed
-----
- Store random postfix to redis for use at cleaup time
- Fix ``_save_data`` method in listener that caused changes to started
and progress fields to be ignored
- Already processed messages in listener are ignored for one day so messages
are not processed twice


===================

===================

Added
-----
- Add ``restart`` method to the ``Data`` model
- Add ``restart`` endpoint to ``DataViewSet``
- Add history tracking for ``Data`` and ``Collection`` objects
- Add ``bulk_delete`` to ``data``, ``entity`` and ``collection`` endpoints

Changed
-------
- **BACKWARD INCOMPATIBLE:** Remove arguments ``inherit_collection`` and
``inherit_entity`` from ``duplicate`` endpoints
- The method ``run_process`` in Python processes returns the created ``Data``
object
- The docker container names have random postfix to support restarting data
object
- The kubernetes resources have random postfix to support restarting data
object
- Remove ``inherit_collection`` and ``inherit_entity`` parameters from
``duplicate`` methods on ``Entity`` and ``Data`` models and corresponding
query sets
- Add filtering and searching to annotation field admin page
- Annotation values are re-validated if annotation field type is changed

Fixed
-----
- Add default ordering on ``AnnotationValue`` model
- Create history object for ``Collection`` if none exists
- Use the same postfix for all kubernetes volumes for the given data object
- Fix ``handle_progress`` handler in the listener
- Fix moving data to entity in collection also sets collection on the data
object


===================