MIAPA-ETL: From NeXML to ISAtab Project Update

[Elliott Hauser]

Refresher:  I'm building a bridge between two important data formats, NeXML and ISAtab, to promote minimum information standards, sharing, and reuse of phylo data.  The meat of the conversion is a set of XSLT files, with (currently) two options for users to run the transforms: a bash script and a tiny java program. 

Completed:

• Provided basic UI in two flavors:
• Simple bash shell script ready for test: user provides path to xml, path to xslt, path to Saxon, and optional output file and voila transformation.  Product of this transformation is not ISAtab (yet; see below)
• Simple java program: ditto, except no Saxon needed:  JAXP handles conversion.
• Defined two draft ISA configurations.  Philippe Rocca-Serra is also working on one, and we meet tomorrow to discuss
• Meeting with Rutger Vos & Philippe made clear that most logical mapping into Assays is via nexml's tb:analysisstep input & output attributes. Still not trivial to map all elements and attributes correctly (see next steps

Problems

• As I mentioned early on, not being able to validate with current ISA configurations added time to the schedule and messed up my old way of breaking the work down.  Hence this week devoted to ISA config instead of finishing transforms.

Next Steps:

• Complete draft ISAconfig
• submit for review:
• mentors
• community lists
• Ideally, submit also a sample transformation so people have an example.
• treebase-record.xml --> ISAtab txt files

If we can get these two steps done this week we'll be in good shape, ready to implement community feedback while moving on towards other clear goals.

Next next steps:

• State MIAPA in terms of treebase NeXML (will need mucho community input here)
• Better UI (possibly GUI but I'm not sure that is beneficial) & tweaking possibilities during the conversion process.  e.g. prompting for missing MIAPA info.