Resolwe

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

Changed
-------
- Make sorting by contributor case insensitive in Elasticsearch endpoints
- Delete ES documents in post delete signal instead of pre delete one

Added
-----
- **BACKWARD INCOMPATIBLE:** Add on-register validation of default values in
process and schemas
- **BACKWARD INCOMPATIBLE:** Validate that field names in processes and
schemas start with a letter and only contain alpha-numeric characters
- Support Python 3.6
- Add ``range`` parameter and related validation to fields of type
``basic:integer:``, ``basic:decimal``, ``list:basic:integer:`` and
``list:basic:decimal``
- Support filtering and sorting by ``process_type`` parameter on Data API
endpoint
- Add ``dirname`` Jinja filter
- Add ``relative_path`` Jinja filter

Fixed
-----
- Add missing ``list:basic:decimal`` type to JSON schema
- Don't crash on empty ``in`` lookup
- Fix {{ requirements.resources.* }} variables in processes to take in to
the account overrides specified in Django settings
- Create Elasticsearch mapping even if there is no document to push


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

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

Changed
-------
- Use the latest versions of the following Python packages in
``resolwebio/rnaseq`` docker image: Cutadapt 1.16, Apache Arrow 0.9.0, pysam
0.14.1, requests 2.18.4, appdirs 1.4.3, wrapt 1.10.11, PyYAML 3.12
- Bump tools version in ``resolwebio/rnaseq`` docker image:

- Salmon to 0.9.1
- FastQC to 0.11.7
- Generalize the no-extraction-needed use-case in ``resolwebio/base`` Docker
image ``download_and_verify`` script

Added
-----
- Add the following Python packages to ``resolwebio/rnaseq`` docker image: six
1.11.0, chardet 3.0.4, urllib3 1.22, idna 2.6, and certifi 2018.1.18
- Add ``edgeR`` R library to ``resolwebio/rnaseq`` docker image
- Add Bedtools to ``resolwebio/rnaseq`` docker image

Fixed
-----
- Handle filenames with spaces in the following processes:

- ``alignment-star-index``
- ``alignment-tophat2``
- ``cuffmerge``
- ``index-fasta-nucl``
- ``upload-fasta-nucl``
- Fix COSMIC url template in (multisample) amplicon reports


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

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

Changed
-------
- **BACKWARD INCOMPATIBLE:** Use Elasticsearch version 5.x
- **BACKWARD INCOMPATIBLE:** Raise an error if an invalid query argument is
used in Elasticsearch viewsets
- **BACKWARD INCOMPATIBLE:** Switch ``Data`` API viewset to use Elasticsearch
- Terminate the executor if listener response with error message
- ``verbosity`` setting is no longer propagated to the executor
- Only create Elasticsearch mappings on first push

Added
-----
- Add ``sort`` argument to ``assertFile`` and ``assertFiles`` methods in
``ProcessTestCase`` to sort file lines before asserting the content
- Add ``process_slug`` field to ``DataSerializer``
- Improve log messages in executor and workload connectors
- Add ``process_memory`` and ``process_cores`` fields to ``Data`` model and
``DataSerializer``
- Support lookup expressions (``lt``, ``lte``, ``gt``, ``gte``, ``in``,
``exact``) in ES viewsets
- Support for easier dynamic composition of type extensions
- Add ``elastic_mapping`` management command

Fixed
-----
- Fix Elasticsearch index rebuilding after a dependant object is deleted
- Send response to executor even if data object was already deleted
- Correctly handle reverse m2m relations when processing ES index dependencies


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

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

Fixed
-----
* Use name-ordered BAM file for counting reads in ``HTSeq-count`` process by
default to avoid buffer overflow with large BAM files


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

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

Changed
-------
- **BACKWARD INCOMPATIBLE:** Remove Ubuntu 17.04 base Docker image due to end
of lifetime
- **BACKWARD INCOMPATIBLE:** Remove support for Jinja in ``DescriptorSchema``'s
default values
- **BACKWARD INCOMPATIBLE:** Remove ``CONTAINER_IMAGE`` configuration option
from the Docker executor; if no container image is specified for a process
using the Docker executor, the same pre-defined default image is used
(currently this is ``resolwe/base:ubuntu-16.04``)
- Add mechanism to change test database name from the environment, appending a
``_test`` suffix to it; this replaces the static name used before

Added
-----
- Add Ubuntu 17.10 and Ubuntu 18.04 base Docker images
- Add database migration operations for process schema migrations
- Add ``delete_chunked`` method to ``Data`` objects queryset which is needed
due to Django's extreme memory usage when deleting a large count of ``Data``
objects
- Add ``validate_process_types`` utility function, which checks that all
registered processes conform to their supertypes
- Add ``FLOW_CONTAINER_VALIDATE_IMAGE`` setting which can be used to validate
container image names
- Only pull Docker images at most once per process in ``list_docker_images``
- Add ``FLOW_PROCESS_MAX_CORES`` Django setting to limit the number of CPU
cores used by a process

Fixed
-----
- Make parallel test suite worker threads clean up after initialization
failures
- Add mechanism to override the manager's control channel prefix from the
environment
- Fix deletion of a ``Data`` objects which belongs to more than one ``Entity``
- Hydrate paths in ``refs`` of ``basic:file:``, ``list:basic:file:``,
``basic:dir:`` and ``list:basic:dir:`` fields before processing ``Data``
object


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

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

Fixed
-----
- Fix ``cuffnorm`` process to correctly use sample names as labels in output
files and expand ``cuffnorm`` tests


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