LinearInfiniteRTransform behaviour has memory about grid HOT 3 CLOSED

Thank you @marco-2023. This also affects ExpRTransform, and PowerRtransform and is related to the fix in issue #143. This is a feature rather than a bug.

The dilemma of setting the b-parameter in the way that fixes your problem is if the user wants to use the functions inverse, deriv or transform again with the same b-parameter. It is written in the code document 1, 2 that the b parameter is set from the first grid that it is used but isn't very clear that it is permanent

The solution to make it more clear for the users is to either:

1. Require the user to define the b parameter in the initializer __init__ so that it is apparent. Currently, it is automatic.
2. Write out a warning if the maximum of the grid being used is different from the b parameter. This adds an additional O(N) cost, but for radial grids N is small, so it's fine.

from grid.

Thank you very much @Ali-Tehrani for clearing my doubts. I will close this issue then (as it is not an issue at all).

from grid.

@marco-2023 and @Ali-Tehrani the main solution to this issue is, as Ali notes, just documentation/examples. I think option #2 is probably the way to go, and then clearify the documentation that the b parameter is permanent.

from grid.