Unexpected timezones in case of different timezones in components and times

When passing components and times with different timezones to hatyan.prediction(), the prediction will have the timezone of the components instead of the times. This is unexpected:

import pandas as pd
import hatyan
comp = pd.DataFrame({"A": [1, 0.5, 0.2],
                     "phi_deg": [10,15,20]}, 
                    index=["M2","M4","S2"])
comp.attrs["nodalfactors"] = True
comp.attrs["fu_alltimes"] = True
comp.attrs["xfac"] = False
comp.attrs["source"] = "schureman"
comp.attrs["tzone"] = "UTC"
dtindex = pd.date_range("2020-01-01 00:00 +01:00","2020-01-02 00:00 +01:00", freq="10min")
pred = hatyan.prediction(comp, times=dtindex)
print(pred)

Gives:

                             values
2019-12-31 23:00:00+00:00 -0.397264
2019-12-31 23:10:00+00:00 -0.406804
2019-12-31 23:20:00+00:00 -0.422670
2019-12-31 23:30:00+00:00 -0.443680
2019-12-31 23:40:00+00:00 -0.468382
                            ...
2020-01-01 22:20:00+00:00 -0.542722
2020-01-01 22:30:00+00:00 -0.498343
2020-01-01 22:40:00+00:00 -0.456521
2020-01-01 22:50:00+00:00 -0.418966
2020-01-01 23:00:00+00:00 -0.387146
[145 rows x 1 columns]

When using:

import pandas as pd
import hatyan
comp = pd.DataFrame({"A": [1, 0.5, 0.2],
                     "phi_deg": [10,15,20]}, 
                    index=["M2","M4","S2"])
comp.attrs["nodalfactors"] = True
comp.attrs["fu_alltimes"] = True
comp.attrs["xfac"] = False
comp.attrs["source"] = "schureman"
comp.attrs["tzone"] = "UTC+01:00"
dtindex = pd.date_range("2020-01-01 00:00","2020-01-02 00:00", freq="10min")
pred = hatyan.prediction(comp, times=dtindex)
print(pred)

Gives:

                             values
2020-01-01 00:00:00+01:00 -0.522025
2020-01-01 00:10:00+01:00 -0.547194
2020-01-01 00:20:00+01:00 -0.568637
2020-01-01 00:30:00+01:00 -0.584397
2020-01-01 00:40:00+01:00 -0.592610
                            ...
2020-01-01 23:20:00+01:00 -0.345106
2020-01-01 23:30:00+01:00 -0.336258
2020-01-01 23:40:00+01:00 -0.335843
2020-01-01 23:50:00+01:00 -0.343647
2020-01-02 00:00:00+01:00 -0.359096
[145 rows x 1 columns]

The usecase in test_prediction_comp_and_times_different_timezones() even interprets the naive tstart/tstop in the tzone of the components:

import hatyan
file_dia = 'c:\\DATA\\hatyan\\tests\\data_unitsystemtests\\VLISSGN_obs1.txt'
df_meas = hatyan.read_dia(filename=file_dia)
comp = hatyan.analysis(ts=df_meas, const_list='year')

times_pred = slice("2019-01-01","2019-02-01", "10min")
ts_pred = hatyan.prediction(comp=comp, times=times_pred)
print(ts_pred)

Gives:

                             values
2019-01-01 00:00:00+01:00  0.989719
2019-01-01 00:10:00+01:00  0.878398
2019-01-01 00:20:00+01:00  0.758829
2019-01-01 00:30:00+01:00  0.631713
2019-01-01 00:40:00+01:00  0.498225
                            ...
2019-01-31 23:20:00+01:00  1.706555
2019-01-31 23:30:00+01:00  1.724520
2019-01-31 23:40:00+01:00  1.722927
2019-01-31 23:50:00+01:00  1.704116
2019-02-01 00:00:00+01:00  1.670706
[4465 rows x 1 columns]

Todo:

[x] change the code so that the timezone (and extent) of times is maintained in the prediction dataframe
[x] still allow for tznaive times with tzaware comp, otherwise all RWS scripts will fail. So the second two testcases will result in the same results as in this issue description. However, give a warning that this happens.
[x] add/update test for mixed timezones (with code above)
[x] run testbank, some timezone tests might fail
[x] run example scripts, especially ddl script
[x] update whatsnew
[x] maybe reduce complexity of prediction_singpleperiod()
[x] check pytest warnings if we get the newly created warning numerous times

Deltares / hatyan

Unexpected timezones in case of different timezones in components and times #334