Feat/meth array template

[anngvu]

Close #405. Basically, the current EpigeneticsAssayTemplate assumes bisulfite sequencing, but we do need to add a template for methylation array data for attributes to make more sense, such as having channels, etc. ADKP model has such a template, but we're adapting Genomics Data Common's https://docs.gdc.cancer.gov/Data_Dictionary/viewer/#?view=table-definition-view&id=raw_methylation_array because it is a better standard.

Update: here is an example with metadata filled in (translated from GEO) for Harish Vasudevan dataset.

[github-actions]

Test Suite Report

Template Generation

template	result	link
AnimalIndividualTemplate	❌
BiospecimenTemplate	😄	template link
ClinicalAssayTemplate	😄	template link
EpigeneticsAssayTemplate	😄	template link
FlowCytometryTemplate	😄	template link
GenomicsAssayTemplate	😄	template link
GenomicsAssayTemplateExtended	😄	template link
HumanCohortTemplate	😄	template link
ImagingAssayTemplate	😄	template link
LightScatteringAssayTemplate	😄	template link
MethylationArrayTemplate	😄	template link
MRIAssayTemplate	😄	template link
PharmacokineticsAssayTemplate	😄	template link
PlateBasedReporterAssayTemplate	😄	template link
ProcessedAlignedReadsTemplate	😄	template link
ProcessedExpressionTemplate	😄	template link
ProcessedVariantCallsTemplate	😄	template link
ProteomicsAssayTemplate	😄	template link
ProtocolTemplate	😄	template link
RNASeqTemplate	😄	template link
ScRNASeqTemplate	😄	template link
UpdateMilestoneReport	😄	template link
WESTemplate	😄	template link
WGSTemplate	😄	template link

Manifest Validation

manifest	result	expectation
GenomicsAssayTemplate_0.csv	😄	Lists can be blank if attr not required using ‘list like’ rule
GenomicsAssayTemplate_1.csv	😄	Mixing blanks and regular list values works
GenomicsAssayTemplate_2.csv	😄	Conditional validation for attributes is currently not supported
ScRNASeqTemplate_0.csv	😄	Single list val works by using ‘list like’ rule
ScRNASeqTemplate_1.csv	😄	Fail because of missing data in required field libraryStrand

[allaway]

@anngvu - isn't this the same as platform? Could we consolidate with that concept (and perhaps also we could specify GEO platform ids for arrays....? :) (

[allaway]

LGTM. I made one optional suggestion regarding chipID

[anngvu]

Thanks @allaway -- GDC actually already had platform as a separate field with the expected enums, so I think chipID is an optional user-specified field intended for batch management.