add support for pandas>=2.2.0

With pandas 2.2.0 we get different havengetallen values, because all culmination hours seem to be shifted. This code will help to debug it further. All assertions will succeed with pandas<=2.1.0, but some will fail with pandas>=2.2.0.
### HAVENGETALLEN
import kenmerkendewaarden as kw
import hatyan
hatyan.close("all")
import numpy as np
import logging
logging.basicConfig(format='%(message)s')
logging.getLogger("kenmerkendewaarden").setLevel(level="INFO")
import datetime as dt
import pandas as pd

# dir_meas = r"c:\Users\veenstra\Downloads\data_temp"
dir_meas = r"p:\11210325-005-kenmerkende-waarden\work\measurements_wl_18700101_20240101"
dir_meas = r"c:\Users\veenstra\Downloads"

df_ext = kw.read_measurements(dir_output=dir_meas, station="HOEKVHLD", extremes=True)
df_ext_12 = hatyan.calc_HWLW12345to12(df_ext)
tstop_str = "2010" # "2010" "2020"
df_ext_12 = df_ext_12.loc["2010":tstop_str]

# df_havengetallen = kw.calc_havengetallen(df_ext=df_ext_12)
current_station = df_ext_12.attrs["station"]
# TODO: check culm_addtime and HWLWno+4 offsets. culm_addtime could also be 2 days or 2days +1h GMT-MET correction. 20 minutes seems odd since moonculm is about tidal wave from ocean
# culm_addtime is a 2d and 2u20min correction, this shifts the x-axis of aardappelgrafiek
# HW is 2 days after culmination (so 4x25min difference between length of avg moonculm and length of 2 days)
# 1 hour (GMT to MET, alternatively we can also account for timezone differences elsewhere)
# 20 minutes (0 to 5 meridian)
culm_addtime = 4*dt.timedelta(hours=12,minutes=25) + dt.timedelta(hours=1) - dt.timedelta(minutes=20)

# TODO: move calc_HWLW_moonculm_combi() to top since it is the same for all stations
# TODO: we added tz_localize on 29-5-2024 (https://github.com/Deltares-research/kenmerkendewaarden/issues/30)
# this means we pass a UTC+1 timeseries as if it were a UTC timeseries
if df_ext.index.tz is not None:
    df_ext_12 = df_ext_12.tz_localize(None)

# kw.havengetallen.get_moonculm_idxHWLWno()
data_pd_moonculm = hatyan.astrog.astrog_culminations(tFirst=df_ext_12.index.min()-dt.timedelta(days=3),tLast=df_ext_12.index.max()) # in UTC, which is important since data_pd_HWLW['culm_hr']=range(12) hourvalues should be in UTC since that relates to the relation dateline/sun
data_pd_moonculm['datetime'] = data_pd_moonculm['datetime'].dt.tz_convert('UTC') #convert to UTC (is already)
data_pd_moonculm['datetime'] = data_pd_moonculm['datetime'].dt.tz_localize(None) #remove timezone
data_pd_moonculm = data_pd_moonculm.set_index('datetime',drop=False)
data_pd_moonculm['values'] = data_pd_moonculm['type'] #dummy values for TA in hatyan.calc_HWLWnumbering()
data_pd_moonculm['HWLWcode'] = 1 #all HW values since one every ~12h25m
data_pd_moonculm = hatyan.calc_HWLWnumbering(data_pd_moonculm,doHWLWcheck=False) #TODO: currently w.r.t. cadzd, is that an issue? With DELFZL the matched culmination is incorrect (since far away), but that might not be a big issue
moonculm_idxHWLWno = data_pd_moonculm.set_index('HWLWno')

df_ext_12_culm = kw.havengetallen.calc_HWLW_moonculm_combi(data_pd_HWLW_12=df_ext_12, culm_addtime=culm_addtime) #culm_addtime=None provides the same gemgetijkromme now delay is not used for scaling anymore
# moonculm_idxHWLWno = kw.havengetallen.get_moonculm_idxHWLWno(tstart=df_ext_12.index.min()-dt.timedelta(days=3),tstop=df_ext_12.index.max())
# moonculm_idxHWLWno.index = moonculm_idxHWLWno.index + 4 #correlate HWLW to moonculmination 2 days before. TODO: check this offset in relation to culm_addtime.

# data_pd_HWLW_idxHWLWno = hatyan.calc_HWLWnumbering(df_ext_12)
# data_pd_HWLW_idxHWLWno['times'] = data_pd_HWLW_idxHWLWno.index
# data_pd_HWLW_idxHWLWno = data_pd_HWLW_idxHWLWno.set_index('HWLWno',drop=False)

# HW_bool = data_pd_HWLW_idxHWLWno['HWLWcode']==1
# data_pd_HWLW_idxHWLWno.loc[HW_bool,'getijperiod'] = data_pd_HWLW_idxHWLWno.loc[HW_bool,'times'].iloc[1:].values - data_pd_HWLW_idxHWLWno.loc[HW_bool,'times'].iloc[:-1] #this works properly since index is HWLW
# data_pd_HWLW_idxHWLWno.loc[HW_bool,'duurdaling'] = data_pd_HWLW_idxHWLWno.loc[~HW_bool,'times'] - data_pd_HWLW_idxHWLWno.loc[HW_bool,'times']
# data_pd_HWLW_idxHWLWno['culm_time'] = moonculm_idxHWLWno['datetime'] #couple HWLW to moonculminations two days earlier (this works since index is HWLWno)
# data_pd_HWLW_idxHWLWno['culm_hr'] = (data_pd_HWLW_idxHWLWno['culm_time'].round('h').dt.hour)%12
# data_pd_HWLW_idxHWLWno['HWLW_delay'] = data_pd_HWLW_idxHWLWno['times'] - data_pd_HWLW_idxHWLWno['culm_time']
# if culm_addtime is not None:
#     data_pd_HWLW_idxHWLWno['HWLW_delay'] -= culm_addtime
# df_ext_12_culm = data_pd_HWLW_idxHWLWno.set_index('times')

# TODO: this plot is shifted between pandas 2.1.0 and pandas 2.2.0
df_ext_12_culm["culm_hr"].plot()

if tstop_str=="2010":
    # df_ext_12_culm times/length seem identical between pandas versions
    assert df_ext_12_culm.index[0] == pd.Timestamp('2010-01-01 02:35:00')
    assert df_ext_12_culm.index[-1] == pd.Timestamp('2010-12-31 23:50:00')
    assert len(df_ext_12_culm) == 1411

    # moonculm_idxHWLWno seem identical between pandas versions
    assert len(moonculm_idxHWLWno) == 711
    assert (moonculm_idxHWLWno.index[:8].values == np.array([7053, 7054, 7055, 7056, 7057, 7058, 7059, 7060])).all() # do note `moonculm_idxHWLWno.index + 4`
    assert (moonculm_idxHWLWno.index[-8:].values == np.array([7756, 7757, 7758, 7759, 7760, 7761, 7762, 7763])).all() # do note `moonculm_idxHWLWno.index + 4`
    assert moonculm_idxHWLWno["datetime"].min() == pd.Timestamp('2009-12-29 09:37:07.049591019')
    assert moonculm_idxHWLWno["datetime"].max() == pd.Timestamp('2010-12-31 20:57:15.209495865')
    assert moonculm_idxHWLWno["datetime"].iloc[0] == pd.Timestamp('2009-12-29 09:37:07.049591019')
    assert moonculm_idxHWLWno["datetime"].iloc[-1] == pd.Timestamp('2010-12-31 20:57:15.209495865')

    # df_ext_12_culm culm_hr seems shifted between pandas version, also check the plot created above
    assert np.isclose(df_ext_12_culm["culm_hr"].min(), 0)
    assert np.isclose(df_ext_12_culm["culm_hr"].max(), 11)
    assert np.isclose(df_ext_12_culm["culm_hr"].iloc[0], 10)
    assert np.isclose(df_ext_12_culm["culm_hr"].iloc[-1], 7)
    count_hr = df_ext_12_culm.groupby("culm_hr").count()["values"].values
    assert (count_hr == np.array([116, 112, 116, 122, 118, 116, 124, 123, 116, 112, 110, 126])).all() # pandas 2.1.0
    # assert (count_hr == np.array([112, 116, 120, 122, 120, 120, 118, 121, 110, 112, 120, 120])).all() # pandas 2.2.0
    assert df_ext_12_culm["HWLW_delay"].min() == pd.Timedelta('0 days 00:09:26.723818778')
    assert df_ext_12_culm["HWLW_delay"].max() == pd.Timedelta('0 days 10:58:41.926870968')
    assert df_ext_12_culm["HWLW_delay"].iloc[0] == pd.Timedelta('0 days 02:06:52.757993453')
    assert df_ext_12_culm["HWLW_delay"].iloc[-1] == pd.Timedelta('0 days 02:21:37.475543290')
elif tstop_str=="2020":
    assert df_ext_12_culm.index[0] == pd.Timestamp('2010-01-01 02:35:00')
    assert df_ext_12_culm.index[-1] == pd.Timestamp('2020-12-31 22:59:00')
    assert len(df_ext_12_culm) == 15528
    assert np.isclose(df_ext_12_culm["culm_hr"].min(), 0)
    assert np.isclose(df_ext_12_culm["culm_hr"].max(), 11)
    assert np.isclose(df_ext_12_culm["culm_hr"].iloc[0], 10)
    assert np.isclose(df_ext_12_culm["culm_hr"].iloc[-1], 11)
    count_hr = df_ext_12_culm.groupby("culm_hr").count()["values"].values
    assert (count_hr == np.array([1238, 1288, 1292, 1298, 1304, 1288, 1352, 1290, 1332, 1290, 1300, 1256])).all() # pandas 2.1.0
    # assert (count_hr == np.array([1258, 1314, 1280, 1306, 1292, 1348, 1298, 1304, 1312, 1296, 1290, 1230])).all() # pandas 2.2.0
    assert df_ext_12_culm["HWLW_delay"].min() == pd.Timedelta('0 days 00:05:31.289305443')
    assert df_ext_12_culm["HWLW_delay"].max() == pd.Timedelta('0 days 12:17:03.623306535')
    assert df_ext_12_culm["HWLW_delay"].iloc[0] == pd.Timedelta('0 days 02:06:52.757993453')
    assert df_ext_12_culm["HWLW_delay"].iloc[-1] == pd.Timedelta('0 days 09:10:15.536207002')
Todo:
[x] fix the bug, test_gemiddeld_getijkromme_av_sp_np_corr and test_gemiddeld_getijkromme_av_sp_np_corr_boi should now be green with any pandas version
[x] check if tis is relevant for other kenmerkendewaarden (add tests and do more assertions)
[x] release pandas<2.2.0 version restriction in pyproject.toml
Deltares-research / kenmerkendewaarden

add support for pandas>=2.2.0 #53
