Generate statistics csv

[veenstrajelmer]

• concise statistics per station for RWS, eg dataframe/csv with: used settings, statistics of used timeseries (interval, #missings), vertical reference
• is minimal print output ok or is csv or so required?
• analysis statistics like confidence interval and Signal to Noise ratio. See T-Tide or U-Tide (suggested in Suggestions for features and changes #81)
• consider only generating statistics if hatyan is called via main, would prevent if from being present in scripts

Statistics idea:

• metadata is now mostly available, so hatyan process is aware of station, vertref and others
• create HatyanSettings.statistics = pd.DataFrame()
• prevent overwriting of HatyanSettings, or at least of HatyanStatistics
• statistics would then be a separate class, with add_statistic that checks whether column already exists for current_station. Otherwise a column like ana_ncomp will be duplicated when doing analysis_perperiod='Y'
• hatyan_settings.statistics.loc[current_station,'ana_len_ts'] = len(ts)
• or add statistics to global defaults: Set hatyan global defaults #133

More notes:

• separate function to do analysis of timeseries (#nans etc), or does that not cover?
• analysis quality
• standard deviation of phase of four years of analysis results (like in pfkw.txt). Information about this variation over years is now also available in the components plot (when including the comp_allperiods)