guess instrument model in demux; pass to output #260
Conversation
| @@ -164,6 +164,7 @@ workflow demux_deplete { | |||
| File spikein_counts = spike_summary.count_summary | |||
|
|
|||
| String run_date = illumina_demux.run_info[0]['run_start_date'] | |||
| String instrument_model = select_first(flatten([illumina_demux.instrument_model,[""]])) | |||
There was a problem hiding this comment.
miniwdl check is erroring in travis because it doesn't like workflow output variables having the same name as workflow input variables -- perhaps change the name slightly?
There was a problem hiding this comment.
This was residual, but I'll keep it as an override—much better. Thanks for taking a look at this.
There was a problem hiding this comment.
I think the continuing CI issue though is that the output variable name is the same as the input variable name. Maybe change the outputs to instrument_model_inferred or something?
| @@ -116,7 +116,7 @@ workflow demux_deplete { | |||
| platform = "ILLUMINA", | |||
| paired = (illumina_demux.run_info[0]['indexes'] == '2'), | |||
| out_name = "sra_metadata-~{flowcell_id}.tsv", | |||
| instrument_model = select_first([instrument_model]), | |||
| instrument_model = select_first(flatten([illumina_demux.instrument_model,[""]])), | |||
There was a problem hiding this comment.
I'm a little confused what the flatten is for and why this can't be constructed a little simpler… (like select_first([illumina_demux.instrument_model,""])) and, actually, if we're going to keep the instrument_model) workflow level input, maybe the workflow level input should be optional (String?) and this line could become:
instrument_model = select_first([instrument_model, illumina_demux.instrument_model])
Meaning that a user-specified value would override, but absent that, it defaults to auto-inferring it.
There was a problem hiding this comment.
illumina_demux.instrument_model is an array as returned by the scatter-gather of demux jobs, no? I was under the impression we'd need to flatten it since select_first() would still return an Array[String] in the proposed change. Or maybe I'm misunderstanding the WDL docs?
allow for override of instrument_model with instrument_model_user_specified
…assembly comment output instrument_model from demux_deplete in sarscov2_illumina_full assembly comment; pass in overrride to to demux_deplete via instrument_model_user_specified
…titute/viral-pipelines into ct-guess-instrument-model
rename instrument_model to instrument_model_inferred in outputs to avoid collision with task-level output naming
|
This is currently pending including of viral-core as a required module in this repo 2.1.24 (#274), and following that manual check of the resulting CI demux job on DNAnexus. |
|
Thanks, I switched to use |
pipes/WDL/tasks/tasks_demux.wdl
Outdated
| String instrument_model = read_string("MACHINE_MODEL") | ||
| String flowcell_lane_count = read_string("LANE_COUNT") |
There was a problem hiding this comment.
| String instrument_model = read_string("MACHINE_MODEL") | |
| String flowcell_lane_count = read_string("LANE_COUNT") | |
| String instrument_model = read_json("~{out_base}-runinfo.json")["sequencer_model"] | |
| String flowcell_lane_count = read_json("~{out_base}-runinfo.json")["lane_count"] |
pipes/WDL/tasks/tasks_demux.wdl
Outdated
| illumina.py flowcell_metadata --inDir $FLOWCELL_DIR flowcellMetadataFile.tsv | ||
|
|
||
| # output machine model and lane count | ||
| grep "machine" flowcellMetadataFile.tsv | cut -f2 | tee MACHINE_MODEL | ||
| grep "lane_count" flowcellMetadataFile.tsv | cut -f2 | tee LANE_COUNT | ||
|
|
There was a problem hiding this comment.
I think you can drop these lines if you make the changes below in the output section
WIP: this depends on the unreleased viral-core:2.1.24 (or later) with the following merged in: