edrn / edrn.rdf Goto Github PK

LabCAS 2.0 has different semantics for data collections and it's about time we address what @chrismattmann started a dozen years ago.

First off, all collection are the same, so we should get rid of this semantic:

<edrn:DNA_METHYLATION_LEV1 rdf:about="…">
  <edrn:DataSetName>DNA_METHYLATION_LEV1</edrn:DataSetName>
</edrn:DNA_METHYLATION_LEV1>
<edrn:FHCRC_MALDI_Dilution_Processed_Data rdf:about="…">
  <edrn:DataSetName>FHCRC MALDI Dilution Processed Data</edrn:DataSetName>
</edrn:FHCRC_MALDI_Dilution_Processed_Data>
  …

because this says we have two different objects of two different types when in reality they're just two collections with different names. This should be:

<edrn:Collection rdf:about="…">
  <edrn:DataSetName>DNA_METHYLATION_LEV1</edrn:DataSetName>
</edrn:Collection>
<edrn:Collection rdf:about="…">
  <edrn:DataSetName>FHCRC MALDI Dilution Processed Data</edrn:DataSetName>
</edrn:Collection>
  …

Then we need to address what Focus BMDB does and fix trying to interpret an XML tag name 😬 and adjust the portal too

While we're at it: use a Dublin Core title instead of our app-specific DataSetName!

The RDF being generated in https://github.com/EDRN/edrn.rdf/blob/master/edrn/rdf/edrnlabcasrdfgenerator.py creates thousands and thousands of duplicate entires (for "Automated Quantitative Measures of Breast Density Data"). It should make about 80 entires for EDRN LabCAS but is making over 4000.