Removing optional inputs

[priyambial123]

Hello,

I use the miniwdl workflow and I get an error in running the deepVariant analysis as:

check JSON input; couldn't construct DeepVariantModel? from "WomCompositeType {\n model ->
WomCompositeType {\n data -> File\ndata_index -> File \n}\nmetadata -> File \n}? (optional)" (in
humanwgs.deepvariant_model) 

In the input.hpc.json, I gave the deepvariant version as 1.5.0. in line 65 Is there anything else to change or give the path:

  "humanwgs.deepvariant_version": "1.5.0",
  "humanwgs.deepvariant_model": "WomCompositeType {\n model -> WomCompositeType {\n data -> File\ndata_index -> File \n}\nmetadata -> File \n}? (optional)",
  "humanwgs.pbsv_call_mem_gb": "Int (optional, default = if N<=3: 64 else 96)",
  "humanwgs.glnexus_mem_gb": "Int (optional, default = 30)",
  "humanwgs.run_tertiary_analysis": "Boolean (optional, default = false)",
  "humanwgs.backend": "HPC",
  "humanwgs.preemptible": true
}

Thank you

Priya

[pb-dstucki]

Hi Priya, please remove the lines marked as (optional) from the json file and run it again.

[williamrowell]

@pb-dstucki is right. The problems here are the optional inputs. Some of the template values are the type of value, rather than the actual value (e.g. "Int", "String", "Boolean", "WomCompositeType"). These values need to be replaced or deleted, depending on whether they are optional or required. Optional inputs can be deleted, like "humanwgs.deepvariant_model", if there's no need to change them. You can remove:

• humanwgs.deepvariant_version
• humanwgs.deepvariant_model
• humanwgs.pbsv_call_mem_gb
• humanwgs.glnexus_mem_gb

If you don't need the tertiary analysis, you can remove humanwgs.run_tertiary_analysis and it default to false. If you do need tertiary analysis, change this line to:

  "humanwgs.run_tertiary_analysis": true,

If the template says Int or Boolean, then the value needs to be bare (without quotes). This is json syntax.

[gevro]

Hi all, The supplied GCP input json example is still giving errors:

Workflow input processing failed

(Caused by [reason 1 of 5]: Failed to evaluate input 'pbsv_call_mem_gb' (reason 1 of 1):
 For input string: "Int (optional, default = if N<=3: 64 else 96)")

(Caused by [reason 2 of 5]: Failed to evaluate input 'glnexus_mem_gb'
    (reason 1 of 1): For input string: "Int (optional, default = 30)")

(Caused by [reason 3 of 5]: Failed to evaluate input 'run_tertiary_analysis'
    (reason 1 of 1): No coercion defined from '"Boolean (optional, default = false)"' of type 'spray.json.JsString' to 'Boolean'.)

(Caused by [reason 4 of 5]: Failed to evaluate input 'preemptible'
    (reason 1 of 1): No coercion defined from '"Boolean"' of type 'spray.json.JsString' to 'Boolean'.) 

Caused by [reason 5 of 5]: Failed to evaluate input 'deepvariant_model'
    (reason 1 of 1): No coercion defined from '"WomCompositeType {\n model -> WomCompositeType {\n data -> File\ndata_index -> File \n}\nmetadata -> File \n}? (optional)"' of type 'spray.json.JsString' to 'WomCompositeType { model -> WomCompositeType { data -> File data_index -> File } metadata -> File }?'.)

Can you all provide an input json that works right out of the box? Or at least some documentation on how to get it working?

[williamrowell]

As in the comment above, delete the optional inputs and set preemptible to truef;

this:

{
  ...
  "humanwgs.deepvariant_version": "1.5.0",
  "humanwgs.deepvariant_model": "WomCompositeType {\n model -> WomCompositeType {\n data -> File\ndata_index -> File \n}\nmetadata -> File \n}? (optional)",
  "humanwgs.pbsv_call_mem_gb": "Int (optional, default = if N<=3: 64 else 96)",
  "humanwgs.glnexus_mem_gb": "Int (optional, default = 30)",
  "humanwgs.run_tertiary_analysis": "Boolean (optional, default = false)",
  "humanwgs.backend": "HPC",
  "humanwgs.preemptible": "Boolean"
}

becomes this:

{
  ...
  "humanwgs.backend": "HPC",
  "humanwgs.preemptible": true
}

[gevro]

Are you sure backend should be HPC and not GCP if we are on Terra?

[williamrowell]

Sorry, yes, should be GCP for Terra. I was copy/pasting from the comment above.