Non-consistent calculation of concentration parameters

[jsitarek]

concentration parameters for MAGIC are computed from non cleaned image:

magic-cta-pipe/magicctapipe/scripts/lst1_magic/magic_calib_to_dl1.py

Line 291 in 01f273d

conc_params = concentration_parameters(camera_geom, image, hillas_params)

same is done for MCs (both MAGIC and LST-1):

magic-cta-pipe/magicctapipe/scripts/lst1_magic/lst1_magic_mc_dl0_to_dl1.py

Lines 328 to 330 in 01f273d

	conc_params = concentration_parameters(
	camera_geoms[tel_id], image, hillas_params
	)

however for LST this is done after selecting only image pixels:
https://github.com/cta-observatory/cta-lstchain/blob/e4a5b7a36dfd643acf47b9006f51fe882b6dbc5c/lstchain/reco/r0_to_dl1.py#L164

the difference is not huge, but especially if there are multiple islands the concentration parameters can behave in complicated way.

Thankfully the concentration parameters are not used in the standard chain.

There are two possible ways to fix it:

1. change both for MAGIC data and for MC to have it the same as for LST-1 data
2. change the behavior for LST-1 MCs to reproduce what is done in the data (which would mean that the parameter would have slightly different meaning for different telescopes)

[jsitarek]

this is the difference between the values of the parameters for the data between LST data (computed from the cleaned image in blue and from full image in orange:

(computed by @nzywucka )