The FMA subset of RDFBones is quite large and complex. This makes RDFBones unwieldy. With the implementation of skeletal inventories, even more classes need to be imported. Most of the subset's complexity stems from classes referenced by the classes we need for RDFBones, i.e. from their semantic context.
I would like to raise the question if we actually need the semantic context. Our purpose is to reference FMA classes so that RDFBones data can be analysed on the semantic background of the FMA. Researchers doing this can import relevant parts of the FMA for their analyses so we do not have to provide it through RDFBones.
I have just played with the 'extract' function of ROBOT while extracting some FMA classes. The MIREOT method extracts class annotations and class hierarchy but no restrictions. I would be little work to write a ROBOT script that extracts desired classes and merges them into a subset file. This might have the following advantages:
The FMA subset becomes considerably leaner
Unwanted superclass axioms are not extracted
The script can pull the identical subset from newer FMA versions for update purposes
The FMA subset of RDFBones is quite large and complex. This makes RDFBones unwieldy. With the implementation of skeletal inventories, even more classes need to be imported. Most of the subset's complexity stems from classes referenced by the classes we need for RDFBones, i.e. from their semantic context.
I would like to raise the question if we actually need the semantic context. Our purpose is to reference FMA classes so that RDFBones data can be analysed on the semantic background of the FMA. Researchers doing this can import relevant parts of the FMA for their analyses so we do not have to provide it through RDFBones.
I have just played with the 'extract' function of ROBOT while extracting some FMA classes. The MIREOT method extracts class annotations and class hierarchy but no restrictions. I would be little work to write a ROBOT script that extracts desired classes and merges them into a subset file. This might have the following advantages: