Fix problems related to missing pol argument when using calc_R_from…

…_A, also did some minor cleanup of example notebook (#116)

Things that were done in this PR:
* it is now forced to pass `pol` to `calc_R_from_A` when passing `f_GHz`, which is the way we commonly use this function. One can also pass `a` and `b` parameters instead of `f_GHz` and `pol`. There are now also tests that check this behavior.
* The WAA function `waa_leijnse_2008_from_A_obs` and `waa_pastorek_2021_from_A_obs` which use `calc_R_from_A` internally, have been adjusted to require a `pol` argument.
* Docstrings of the functions have been adjusted and also improved in some parts
* tests have been adjusted accordingly and new tests added
* black formating in files that were touched in this PR
* updated notebook on basic example workflow
  *  use `pol` correctly (which is no enforce anyway, see above)
  * removed some unnecessary cells
  * improves some markdown text
  * changed the way resampling to 5-min and 1-h is done, which we do now after concatenating the list of CML `Datasets` to one large `Dataset` because resampling is much faster like this
  * made plotting of interpolated maps a bit nicer (in my opinion)
  * made plotting of time series more consistent (same figure width, removing unnecessary labels, better colors in some places)
  * use `cml['A'].values[cml.A < 0] = 0` instead of `cml['A'] = cml.A.where(cml.A > 0, 0)`, see #117
  * added interpolation of short gaps in `trsl`, up to 5 minutes, which improves comparison with reference for high rain rates (I think we had this interpolation in the example notebook some time ago, but maybe it was removed to replace it with the more sophisticated blackout gap filling, which is not there yet..)