Code improvements

[veenstrajelmer]

Documentation:

• update docstrings in all exposed functions in modules section
• add extra examples
• filter/remove example scripts
• fix python code coloring in embedded ipynb
• fix code quality badge (& in url goes wrong): Including a README.md with links that contain &s, "translates them" to & omnilib/sphinx-mdinclude#19 >> Certain characters in inline code incorrectly parsed (e.g., &) lepture/mistune#372 and Certain characters in markdown code incorrectly parsed (e.g., &) lepture/mistune#394

hatyansettings:

• move validation of timestep argument for prediction() to HatyanSettings?
• clean up HatyanSettings related code, maybe validate arguments in a more charming way
• also validate const_list and times with hatyan.HatyanSettings, const_list is now validated with check_requestedconsts multiple times (u/f/analysis functions), times is validated in hatyan.prediction()
• maybe make kwargs in HatyanSettings to reduce number of input arguments. Still raise exception if not all kwargs were popped after init. complexity of this class is also way too high, so maybe first redesign it.

resample_timeseries:

• hatyan.resample_timeseries(): round dataframe to 'S' before doing anything, start new datetimeindex on 00 minute
• make resample_timeseries() work with times slice instead of three keywords? Or deprecate. At least covert timestep_min to timestep and do freq validation there

metadata:

• Consider updating metadata "origin" attribute of components object: when slotgem is overwritten, with component splitting, with merging of component sets, in case of analysis-perperiod vs. atonce and maybe other settings. Not possible to derive these details from existing/old components file, so is it useful?
• add vertref/station from metadata to components/timeseries plot (if available), also requires comparing these attrs between input arguments
• write_dia does not allow vertref/station arguments anymore, add deprecationerror

Code improvements:

• Improve code quality #57
• Increase performance with numba (Lru_cache also helps significantly), maybe add pandas performance dependencies
• Align foreman/schureman, avoids code with same functions. Not easy, but aligning shallowrelations should be possible. Or move from schureman to simplified foreman?
• Improve analysis prediction varnames? (eg v, u, f)
• Check if there are any try/except combi's that are not error specific. Raise predefined errors like ValueError or custom errors like like hatyan.analyis_prediction.MatrixConditionTooHigh and catch them properly (not a general 'except:' and no general 'raise Exception()')
• Numbering of extremes/HWLW: store phasediff-compute-script (numbering_extremes.py) properly

Improve configfiles:

• Bug? : 19Ycomp4Ydia failing stations: ABDN and BAALHK, might be due to gaps in measurement timeseries
• Bug? : rounding error for 2022 and 2023 prediction
• Update all acceptancetests, remove what is possible