'p_eb' not working for MAP

[Sebastian-Belkner]

Must have to do with the migration to lenspyx. QE looks alright. Other estimators don't seem to be affected.
Likely shape mismatch or similar in the lensgclm() function. This must be fixed before going public.

Once this is fixed, we probably want to upgrade testing and test all estimator keys..

[Sebastian-Belkner]

This had nothing to do with the migration; there had never been a working p_eb estimator for MAP. There is a solution on the fullsky now and we might implemented it in the near future.