Automatic generator of dataflow descriptor languages.

The purpose of this interface is to enable the use of QIIME 2 plugins in environments which support Dockstore tools/workflows.

To see rendered tools, visit:

qiime2/dockstore-tools (in progress)

• Workflow Description Language
• Common Workflow Lanauage

• Automatic .dockstore.yml registration
• NextFlow (TBD)

A similar interface exists for Galaxy Tool definitions called q2galaxy, which may be integrated into q2dataflow as well (although publication of these tools to Dockstore is not currently possible as only .ga files (not tools) are supported).

q2dataflow {cwl | wdl} template {all | builtins} {output directory}

or

q2dataflow {cwl | wdl} template plugin {plugin_id} {output directory}

q2dataflow requires installation of the following packages:

• qiime2, most conveniently installed via a distro
• click
• pytest

Then to install q2wdl, activate an environment with QIIME 2 and run:

pip install 'q2dataflow @ git+https://github.com/qiime2/q2dataflow.git'

Additionally, the WDL-related tests require:

1. an installation of the WDL emulator miniwdl
2. a docker environment containing an installation ofq2dataflow and qiime2

If running on a Mac, it is very important for #1 to follow the Mac-specific install instructions described in chanzuckerberg/miniwdl#145; miniwdl will fail in somewhat inscrutable ways if the export TMPDIR=/tmp step near the end is not performed.

For #2, a dockerfile is provided that will install q2dataflow and also the latest qiime2 distribution available on quay.io/repository/qiime2/amplicon.

Using this dockerfile, a docker image can be built from the command line in the top q2dataflow directory using the command

docker build docker/ --build-arg CACHEBUST=$(date +%s) REF=main --tag testq2dataflow

The name of the image, testq2dataflow, is important: it is used in generating the test commands in q2dataflow\languages\wdl\usage.py. Currently, it is set in a variable in this module and should presumably be refactored to elsewhere.

The tests generate wdl files and parameter files and then run each wdl file (using the parameters in the associated parameters file) through miniwdl on the testq2dataflow docker image. If the tests hang, go to a terminal, enter the conda environment containing miniwdl, and run miniwdl run_self_test. If this hangs due to errors stating "error while loading TLS certificate in /var/lib/docker/swarm/certificates/swarm-node.crt", then run docker swarm leave --force and try again.