Add option to remove integrator drift

[odstrcilt]

Flux loop signals and magnetics probe signals can have nonnegligible errors caused by the analog integrator drift. Assuming that the drift is linear in time, this will set signals back to zero after the end of the discharge.

[orso82]

@odstrcilt perhaps you forgot to commit the remove_integrator_drift function?

[odstrcilt]

@orso82 do you have any guess what is causing the regression test to fail? It does not seem to be related to any changes done by me.

[smithsp]

@odstrcilt It looks like you ended up with a branch of omas that depends on a new DD from IMAS, where the EC launchers have changed structure. @orso82 Does this imply that we need to address the launchers on omas now?

[orso82]

Yes @smithsp we made changes on a OMFIT branch, but we still need to address this on the OMAS side. Mostly it's about the regression tests and machine mappings.

[smithsp]

Need to merge with master.

[smithsp]

Still need to merge with master branch.

[smithsp]

@odstrcilt When @orso82 originally looked into fetching NSTX data using omas, he had gone to great lengths to compare what was fetched to the existing k-files. How do these changes affect that comparison and validation?

[orso82]

@odstrcilt in case it's needed, you can look at the https://github.com/gafusion/OMFIT-source/blob/unstable/modules/EFITtime/SCRIPTS/NSTX/update_keqdsk_from_omas.py script, and in particular use the OMFITkeqdsk.compare_omas_constraints function.

[smithsp]

Ping @odstrcilt to compare to previous k-files.

[odstrcilt]

I have tested shot 133964 at 285ms

Here is compared OMAS case with and without drift substraction

Here is OMAS with drift subtraction compared to EFIT01 kfile

Here is OMAS without drift subtraction compared to EFIT01 kfile

I will try to plot also timetraces...

[odstrcilt]

@orso82 can we merge it?

[orso82]

A