Create netcdf overview function

[veenstrajelmer]

• slice time before plotting, is faster (and saves memory) for timeseries extending beyond tstart/tstop
• extend the statistics csv: filename, platform_name, platform_id, t0, tend, nr of datapoints, mean(dt) in min, mode(dt) in min, completeness in %, min. wl, max. wl, min(dt) in min, max(dt) in min, lon, lat, source. Also all values for timeseries without nans?
• add xyn writing, ssh_catalog_toxynfile() already exists in dfm_tools but we should first derive the catalog from the netcdf files
• think about coarsening, if plotting three months, is hourly/daily scatter still enough to plot continuous line? Maybe add optimize=True argument. Maybe just first try it.
• less printing
• derive stats from original dataset (before time slicing)
• improve figure: smaller text, more x-ticks, remove xlabel
• increase float precision of ssh_catalog_toxynfile()?
• filter out nans from timestamps
• add None option for time_min/time_max/yearstep
• add overview figure to example notebook
• optionally create kml if very simple: https://stackoverflow.com/questions/71139672/how-to-generate-kml-file-with-colored-markers >> leave out for now
• fix alphabetic ordering of stations in figures
• update whats-new.md
• filenames as station names in plot
• clear output folder before starting
• add to all getijverandering download scripts (incl time_min/time_max/yearstep), remove matlab output and script (after comparison)
• add testcase for dfmt.ssh_netcdf_overview() (assert file presence)