40 add support for different timezones (#98)

* add support for timezone in input time extents

* added timezone testcase

* updated history.rst

HISTORY.rst

@@ -7,6 +7,7 @@ UNRELEASED
 * avoid duplicated periods in dataframe returned by `ddlpy.measurements_amount()` in https://github.com/Deltares/ddlpy/pull/93
 * allow for different retrieval frequencies (including None) in `ddlpy.measurements()` in https://github.com/Deltares/ddlpy/pull/95
 * only catch "Geen waarnemingen aanwezig!" error message and raise all others (for instance for a too large request) in https://github.com/Deltares/ddlpy/pull/97
+* support for timezones in start_date/end_date in https://github.com/Deltares/ddlpy/pull/98
 
 0.4.0 (2024-04-08)
 ------------------

ddlpy/ddlpy.py

@@ -105,12 +105,21 @@ def _check_convert_dates(start_date, end_date, return_str=True):
     start_date = pd.Timestamp(start_date)
     end_date = pd.Timestamp(end_date)
 
+    # check if timezones are the same
+    assert start_date.tz == end_date.tz
+
+    # set UTC timezone if tz is None
+    if start_date.tz is None:
+        start_date = pytz.UTC.localize(start_date)
+    if end_date.tz is None:
+        end_date = pytz.UTC.localize(end_date)
+
     if start_date > end_date:
-        raise ValueError("start_date is larger than end_date")
+        raise ValueError(f"start_date {start_date} is larger than end_date {end_date}")
 
     if return_str:
-        start_date_str = pytz.UTC.localize(start_date).isoformat(timespec='milliseconds')
-        end_date_str = pytz.UTC.localize(end_date).isoformat(timespec='milliseconds')
+        start_date_str = start_date.isoformat(timespec='milliseconds')
+        end_date_str = end_date.isoformat(timespec='milliseconds')
         return start_date_str, end_date_str
     else:
         return start_date, end_date
@@ -388,7 +397,7 @@ def measurements(location, start_date, end_date, freq=dateutil.rrule.MONTHLY, cl
 
     if clean_df:
         measurements = _clean_dataframe(measurements)
-
+    
     return measurements
 
 

tests/test_ddlpy.py

@@ -7,6 +7,7 @@
 import pytest
 import ddlpy
 import dateutil
+import numpy as np
 
 
 @pytest.fixture
@@ -170,6 +171,31 @@ def test_measurements_duplicated(measurements):
     assert isinstance(meas_clean.index, pd.DatetimeIndex)
 
 
+def test_measurements_timezone_behaviour(location):
+    start_date = "2015-01-01 00:00:00 +01:00"
+    end_date = "2015-01-03 00:00:00 +01:00"
+    measurements = ddlpy.measurements(location, start_date=start_date, end_date=end_date)
+    assert str(measurements.index[0].tz) == 'UTC+01:00'
+    assert measurements.index[0] == pd.Timestamp(start_date)
+    assert measurements.index[-1] == pd.Timestamp(end_date)
+
+    data_amount_dag = ddlpy.measurements_amount(location, start_date=start_date, end_date=end_date, period="Dag")
+    # when retrieving with tzone +01:00 we expect 1 value on 2015-01-03
+    assert np.allclose(data_amount_dag["AantalMetingen"].values, [144,144,1])
+
+
+    start_date = "2015-01-01"
+    end_date = "2015-01-03"
+    measurements = ddlpy.measurements(location, start_date=start_date, end_date=end_date)
+    assert str(measurements.index[0].tz) == 'UTC+01:00'
+    assert measurements.index[0] == pd.Timestamp(start_date).tz_localize("UTC").tz_convert('UTC+01:00')
+    assert measurements.index[-1] == pd.Timestamp(end_date).tz_localize("UTC").tz_convert('UTC+01:00')
+
+    data_amount_dag = ddlpy.measurements_amount(location, start_date=start_date, end_date=end_date, period="Dag")
+    # when retrieving with tzone +00:00 we expect 7 values on 2015-01-03
+    assert np.allclose(data_amount_dag["AantalMetingen"].values, [138,144,7])
+
+
 def test_nodataerror(location):
     """
     Test whether a request that returns no data is indeed properly catched