Assay improvements - issues #369, #370

[Zigur]

Work in progress, for now, issue #369 is completed, #370 still todo.

[coveralls]

Coverage increased (+4.2%) to 71.121% when pulling 47c366e on assay_improvements into 1c946bf on develop.

[Zigur]

This is now ready to be reviewed and merged. @proccaserra have a look at the outputted ISA tables using https://github.com/Zigur/isa-create-notebook/blob/dev/notebooks/Investigation-from-datascriptor-config.ipynb for instance. Check that there are no duplicates in the node names (I'll add some tests on that too, before merging).

[proccaserra]

Line1056 in create/model.py causes an exception "TypeError: expected string or bytes-like object" when the assay definition used is annotated with ontology terms (to reproduce switch flag for MS (annotated) to true, when testing with the notebook).

then at line 2334 in the same module, some required keys are missing "'FreeInductionDecayDataFile" (and possibly "ImageDataFile") -
TODO: review assay definitions and test all current layouts)

[Zigur]

This is mostly completed, I'd say. I have included support for all data file types, both raw and derived, for completeness:

isa-api/isatools/create/model.py

Lines 2349 to 2355 in 746fde3

	assert isa_class in {
	# expand this set if needed
	RawDataFile, RawSpectralDataFile, ArrayDataFile, FreeInductionDecayDataFile,
	DerivedDataFile, DerivedSpectralDataFile, DerivedArrayDataFile,
	ProteinAssignmentFile, PeptideAssignmentFile, DerivedArrayDataMatrixFile,
	PostTranslationalModificationAssignmentFile, AcquisitionParameterDataFile
	}

I have added two prefixes to the node names to clarify the notations:

• AT is prepended to the assay type id
• S is prepended to the start node (or sample node) index
• R is prepended to the replicate node index

Some examples:

• extract name: AT1-S9-Extract-R1
• labelled extract name: AT1-S9-LE-R1
• process name: AT1-S9-mass-spectrometry-Acquisition-R1
• raw data file name: AT1-S9-raw-spectral-data-file-R1

@proccaserra have a look at this notebook for reviewing a scenario with the ontology annotated MS assay and the NMR assay

[Zigur]

This seems to work also for assays with chained protocols now (see this notebook, at the moment the only functional one, others to be updated soon).

[proccaserra]

looks good to me now.