Make DatetimeIndex timezone naive for compatibility with xarray/numpy engine

darts/tests/test_timeseries.py

@@ -1773,6 +1773,55 @@ def test_time_col_convert_datetime_strings(self):
         self.assertEqual(ts.time_index.dtype, "datetime64[ns]")
         self.assertEqual(ts.time_index.name, "Time")
 
+    def test_time_col_with_tz(self):
+        # numpy and xarray don't support "timezone aware" pd.DatetimeIndex
+        # the BUGFIX removes timezone information without conversion
+
+        time_range_MS = pd.date_range(
+            start="20180501", end="20200301", freq="MS", tz="CET"
+        )
+        values = np.random.uniform(low=-10, high=10, size=len(time_range_MS))
+        # pd.DataFrame loses the tz information unless it is contained in its index
+        # (other columns are silently converted to UTC, with tz attribute set to None)
+        df = pd.DataFrame(data=values, index=time_range_MS)
+        ts = TimeSeries.from_dataframe(df=df)
+        self.assertEqual(list(ts.time_index), list(time_range_MS.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_MS))
+        self.assertEqual(ts.time_index.tz, None)
+
+        serie = pd.Series(data=values, index=time_range_MS)
+        ts = TimeSeries.from_series(pd_series=serie)
+        self.assertEqual(list(ts.time_index), list(time_range_MS.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_MS))
+        self.assertEqual(ts.time_index.tz, None)
+
+        ts = TimeSeries.from_times_and_values(times=time_range_MS, values=values)
+        self.assertEqual(list(ts.time_index), list(time_range_MS.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_MS))
+        self.assertEqual(ts.time_index.tz, None)
+
+        time_range_H = pd.date_range(
+            start="20200518", end="20200521", freq="H", tz="CET"
+        )
+        values = np.random.uniform(low=-10, high=10, size=len(time_range_H))
+
+        df = pd.DataFrame(data=values, index=time_range_H)
+        ts = TimeSeries.from_dataframe(df=df)
+        self.assertEqual(list(ts.time_index), list(time_range_H.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_H))
+        self.assertEqual(ts.time_index.tz, None)
+
+        serie = pd.Series(data=values, index=time_range_H)
+        ts = TimeSeries.from_series(pd_series=serie)
+        self.assertEqual(list(ts.time_index), list(time_range_H.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_H))
+        self.assertEqual(ts.time_index.tz, None)
+
+        ts = TimeSeries.from_times_and_values(times=time_range_H, values=values)
+        self.assertEqual(list(ts.time_index), list(time_range_H.tz_localize(None)))
+        self.assertEqual(list(ts.time_index.tz_localize("CET")), list(time_range_H))
+        self.assertEqual(ts.time_index.tz, None)
+
     def test_time_col_convert_garbage(self):
         expected = [
             "2312312asdfdw",

darts/timeseries.py

@@ -654,7 +654,17 @@ def from_dataframe(
                 "a DatetimeIndex, or with a RangeIndex.",
                 logger,
             )
-            time_index = df.index
+            # BUGFIX : force time-index to be timezone naive as xarray doesn't support it
+            # pandas.DataFrame loses the tz information if it's not its index
+            if isinstance(df.index, pd.DatetimeIndex) and df.index.tz is not None:
+                print(
+                    "The provided DatetimeIndex was associated with a timezone, which is currently not supported "
+                    "by xarray. To avoid unexpected behaviour, the tz information was removed, don't forget to "
+                    f"call `ts.time_index.tz_localize({df.index.tz})` when exporting the results."
+                )
+                time_index = df.index.tz_localize(None)
+            else:
+                time_index = df.index
 
         if not time_index.name:
             time_index.name = time_col if time_col else DIMS[0]
@@ -831,6 +841,14 @@ def from_series(
         TimeSeries
             A univariate and deterministic TimeSeries constructed from the inputs.
         """
+        # BUGFIX : force time-index to be timezone naive as xarray doesn't support it
+        if isinstance(pd_series, pd.DatetimeIndex) and pd_series.dt.tz is not None:
+            print(
+                "The `times` argument was associated with a timezone, which is currently not supported "
+                "by xarray. To avoid unexpected behaviour, the tz information was removed, don't forget to "
+                f"call `ts.time_index.tz_localize({pd_series.dt.tz})` when exporting the results."
+            )
+            pd_series = pd_series.dt.tz_localize(None)
 
         df = pd.DataFrame(pd_series)
         return cls.from_dataframe(
@@ -924,6 +942,15 @@ def from_times_and_values(
             "TimeSeries.from_values() if you want to use an automatic RangeIndex.",
         )
 
+        # BUGFIX : force time-index to be timezone naive as xarray doesn't support it
+        if isinstance(times, pd.DatetimeIndex) and times.tz is not None:
+            print(
+                "The `times` argument was associated with a timezone, which is currently not supported "
+                "by xarray. To avoid unexpected behaviour, the tz information was removed, don't forget to "
+                f"call `ts.time_index.tz_localize({times.tz})` when exporting the results."
+            )
+            times = times.tz_localize(None)
+
         times_name = DIMS[0] if not times.name else times.name
 
         # avoid copying if data is already np.ndarray: