add workflows for isnvs_merge_to_vcf #864
Conversation
|
Hm.. just realized looking at this: what’s the difference between the two tasks besides the boolean input? If there’s no difference perhaps they should be consolidated to one task? The Boolean input could even be optional if we want it to be. Separately, should we rename all our tasks/xxx.wdl import files to tasks_xxx.wdl ? |
|
Nice catch; the two had carried over from the Snakemake rules. I'll consolidate them. We could rename the task files to avoid collisions. I can submit that as a separate PR. |
…/viral-ngs into ct-add-merge-to-vcf
tomkinsc
changed the title
| --parse_accession | ||
|
|
||
| $naive_filter \ | ||
| --parse_accession && \ |
There was a problem hiding this comment.
Actually, instead of stringing everything in the WDL command section together with &&, I've found it easier just to set -ex -o pipefail in any WDL task that's more than a one-liner.
|
|
…/viral-ngs into ct-add-merge-to-vcf
errors occur on Travis that so far have not been reproducible locally
| @@ -3,7 +3,7 @@ | |||
| # A wrapper script to run viral-ngs docker images | |||
| # The following paths have to be modified according to end-user environment | |||
| NOVOALIGN_PATH="/opt/novocraft" # Directory where novoalign.lic license file can befound | |||
| GATK_PATH="/opt/GenomeAnalysisTK-3.6" # Directory where the correct GATK jar file can be found | |||
| GATK_PATH="/opt/GenomeAnalysisTK-3.8" # Directory where the correct GATK jar file can be found | |||
There was a problem hiding this comment.
There was a problem hiding this comment.
Has anyone used rundocker.sh recently? It doesn’t get used for anything more pipelined, but for ad hoc command line work, is it useful beyond what you get from the usual “docker run -it —rm”?
There was a problem hiding this comment.
I haven't used it recently.
allow passage of stderr handle in run_and_print, defaulting to stdout redirection as before if not specified
tomkinsc
dismissed
dpark01’s stale review
this also needs a fix related to sample name inference
| @@ -59,7 +59,7 @@ def execute(self, command, args, JVMmemory=None, stdin=None, stdout=None): # | |||
| ] + args | |||
|
|
|||
| _log.debug(' '.join(tool_cmd)) | |||
| return util.misc.run_and_print(tool_cmd, stdin=stdin, buffered=True, silent=("databases" in command), check=True) | |||
| return util.misc.run_and_print(tool_cmd, stdin=stdin, stderr=stderr, buffered=True, silent=("databases" in command), check=True) | |||
There was a problem hiding this comment.
Any reason we can just revert this to subprocess.check_call?
There was a problem hiding this comment.
We probably should revert. Let's save it for a separate PR? This one is already larger than it should be.
No description provided.