WRROC has invalid hasParts for some nf-core pipelines

[bentsherman]

Four other pipelines have missing payloads. These files are in hasPart, but they are missing in the Crate:

denovotranscript-1.0.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/ce/1f37ce4888f0e3305b4db32f2fcc76/okayset/all_assembled.okay.mrna' as part of its payload",
denovotranscript-1.0.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/ce/1f37ce4888f0e3305b4db32f2fcc76/okayset/all_assembled.pubids' as part of its payload",
proteinfold-1.1.1-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/8d/5f814f4ec329c4c24f51eb7e4c8460/T1024.1.fasta' as part of its payload",
proteinfold-1.1.1-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/c1/c4175ab996c015eef3f88848408521/T1026.1.fasta' as part of its payload",
viralrecon-2.6.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/13/c9227d087fffb9e03051bc06587ffc/quast/report.tsv' as part of its payload",
viralrecon-2.6.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/58/89a60f6a6959b2ad48da2dbd9cb699/quast/report.tsv' as part of its payload",
viralrecon-2.6.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/85/e6deaba6c3df726378006b31d72512/quast/report.tsv' as part of its payload",
viralrecon-2.6.0-ro-crate-1.1-required.json:            "message": "The RO-Crate does not include the Data Entity 'work/89/60cec3fa49ded79c35a944283142c8/quast/report.tsv' as part of its payload",

Originally posted by @fbartusch in #39 (comment)

[bentsherman]

It looks like some intermediate files are being included under hasPart instead of mentions.

The hasPart is defined as:

nf-prov/plugins/nf-prov/src/main/nextflow/prov/WrrocRenderer.groovy

Lines 498 to 503 in 0a4f495

	"hasPart" : withoutNulls([
	["@id": mainScriptId],
	*asReferences(datasetParts),
	*asReferences(inputFiles),
	*asReferences(outputFiles)
	]),

At a first glance, I'm not sure how a work path would get into any of these values. The inputFiles is derived from this function:

nf-prov/plugins/nf-prov/src/main/nextflow/prov/ProvHelper.groovy

Lines 96 to 109 in 0a4f495

	static Set<Path> getWorkflowInputs(Set<TaskRun> tasks, Map<Path,TaskRun> taskLookup) {
	final result = [] as Set<Path>
	tasks.each { task ->
	task.getInputFilesMap().each { name, path ->
	if( taskLookup[path] )
	return
	result << path
	}
	}
	return result
	}

Which essentially says "collect all task input files that are not an output of a previous task (i.e. not in the task lookup)". So maybe the task lookup is missing a few tasks?

I would be curious to see if these errors go away on a resumed run, similar to #40 .