WDL workflows, input files and configuration files for routine tasks.
The following configurations should be supported for each of the workflows:
- local
- SLURM on Rivanna
- PBS-TORQUE
Each workflow should have a WDL file, and an input json skeleton. Also, every workflow that is to be run independently should have a markdown file describing the workflow.
An example:
java -Xmx4g -Dconfig.file=local.conf -jar cromwell-36.jar run hg_alns_to_gvcfs.wdl --inputs hg_alns_to_gvcfs.json
In the absence of a style guide, the following should be followed:
- Tab is equal to 2 spaces
- All the analyses is to be carried out in the "cromwell-executions" folder, since the duplication strategy is easier to implement this way. At the end of the workflow, all output files should be copied to a location which is determined using the variable "final_out_dir".
- All filenames in the input json file are to be absolute paths.