spinicist
commented
1 year ago For the majority of commands the output file has -command appended to the end. The exceptions I can find right now are:
precond- user has to specify the output name. VERY happy to change this to fit the scheme.sense-calib- appends-sense, which conflicts with the actualsensecommand. Again happy to change this.
Can you spot any others? Turns out I added quite a lot of commands 😅
An alternative to consider is to use the -o name.h5 option that most commands support and specify a set filename to use, such as -o temp_out.h5? That could make cleaning up easier. I'm leaning towards making sure the output names are consistent regardless though.
As to programmatic API discovery, it is trivial to print a list of the commands in JSON or similar. Sadly, it is harder to then get a list of options/flags for each command due to the design of the argument parsing library I used. I think the latter could be done - but I suspect it would be more work than working out how to parse the help text!
emilljungberg
Starting to work on a python interface now, similar to the BART structure. What we need to make this work is a list of what the suffix is for the files generated by each command in order to read in the output files into python again. Is there an easy way to figure this out? Would also need to know which flags for each command that triggers a specific output. I think building a list or dictionary of some kind for this would be easier than having separate functions for each tool that needs to be maintained.