Resolwe

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

Changed
-------
- **BACKWARD INCOMPATIBLE:** Use Elasticsearch version 6.x
- **BACKWARD INCOMPATIBLE:** Bump ``Django`` requirement to version ``2.2``
- **BACKWARD INCOMPATIBLE:** Remove not used ``django-mathfilters``
requirement

Added
-----
- Support Python 3.7
- Support forward and reverse many-to-one relations in Elasticsearch
- Add ``collection_names`` field to ``DataSerializer``
- Add test methods to ``ProcessTestCase`` that assert directory structure and
content: ``assertDirExists``, ``assertDir``, and ``assertDirStructure``
- Add ``upload-dir`` process


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

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

Changed
-------
- Bump STAR aligner version to 2.7.0c in ``resolwebio/rnaseq:4.2.2``
- Processes ``alignment-star`` and ``alignment-star-index`` now use Docker
image ``resolwebio/rnaseq:4.2.2`` which contains STAR version ``2.7.0c``
- Persistence of ``basespace-file-import`` process changed from ``RAW`` to
``TEMP``

Added
-----
- Make ``prepare-geo-chipseq`` work with both
``data:chipseq:callpeak:macs2`` and
``data:chipseq:callpeak:macs14`` as inputs

Fixed
-----
- Report correct total mapped reads and mapped reads percentage in
prepeak QC report for ``data:alignment:bam:bowtie2`` inputs in
``macs2-callpeak`` process


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

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

Changed
-------
- Enable multithreading mode in ``alignment-bwa-aln`` and
``alignment-bwa-sw``
- Lineary lower the timeout for BigWig calculation when running on
multiple cores

Fixed
-----
- Remove ``pip`` ``--process-dependency-links`` argument in testenv
settings
- Fix walt getting killed when ``sort`` runs out of memory. The ``sort``
command buffer size was limited to the process memory limit.


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

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

Changed
-------

Added
-----
- Add the ``FASTQ`` file validator script to the ``upload-fastq-single``,
``upload-fastq-paired``, ``files-to-fastq-single`` and
``files-to-fastq-paired`` processes
- Add ``spikein-qc`` process
- Add to ``resolwebio/rnaseq:4.1.0`` Docker image:

- ``dnaio`` Python library
- Add to ``resolwebio/rnaseq:4.2.0`` Docker image:

- ERCC table
- common Genialis fonts and css file
- spike-in QC report template
- Set ``MPLBACKEND`` environment variable to ``Agg`` in
``resolwebio/common:1.0.1`` Docker image

Fixed
-----
- Fix the format of the output ``FASTQ`` file in the ``demultiplex.py``
script
- Fix NSC and RSC QC metric calculation for ATAC-seq and paired-end
ChIP-seq samples in ``macs2-callpeak`` and ``qc-prepeak`` processes


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

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

Fixed
-----
- Pin ``django-priority-batch`` to version ``1.1`` to fix compatibility issues


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

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

Changed
-------
- **BACKWARD INCOMPATIBLE:** Access to DataField members (in Python process
input) changed from dict to Python objects. For example,
``input_field.file_field['name']`` changed to
``input_field.file_field.path``.
- **BACKWARD INCOMPATIBLE:** Filters that are based on ``django-filter``
``FilterSet`` now use dict-declaring-syntax. This requires that subclasses
of respective filters modify their syntax too.
- Interactively save results in Python processes

Added
-----
- Add get_data_id_by_slug method to Python processes' Process class
- Python process syntax enhancements:

- Support ``.entity_name`` in data inputs
- Easy access to process resources through ``self.resources``
- Raise error if ViewSet receives invalid filter parameter(s)
- Report process error for exceptions in Python processes
- Report process error if spawning fails
- Automatically export files for spawned processes (in Python process syntax)
- Import files of Python process FileField inputs (usage:
`inputs.src.import_file()`)

Fixed
-----
- Interactively write to standard output within Python processes
- Fix writing to integer and float output fields
- Allow non-required ``DataField`` as Python process input


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