fix pyarrow issues (#468)

src/traffic/algorithms/openap.py

@@ -32,8 +32,7 @@ def phases(self, twindow: int = 60) -> "Flight":
 
         fp = FlightPhase()
         fp.set_trajectory(
-            (self.data.timestamp.values - np.datetime64("1970-01-01"))
-            / np.timedelta64(1, "s"),
+            (self.data.timestamp.astype(int) // 1_000_000_000).values,
             altitude,
             groundspeed,
             vertical_rate,

src/traffic/core/flight.py

@@ -116,17 +116,17 @@ def _split(
     # This method helps splitting a flight into several.
     if data.shape[0] < 2:
         return
-    diff = data.timestamp.diff().values
+    diff = data.timestamp.diff()
     if unit is None:
         delta = pd.Timedelta(value).to_timedelta64()
     else:
         delta = np.timedelta64(value, unit)
     # There seems to be a change with numpy >= 1.18
     # max() now may return NaN, therefore the following fix
-    max_ = np.nanmax(diff)
+    max_ = diff.max()
     if max_ > delta:
         # np.nanargmax seems bugged with timestamps
-        argmax = np.where(diff == max_)[0][0]
+        argmax = diff.argmax()
         yield from _split(data.iloc[:argmax], value, unit)
         yield from _split(data.iloc[argmax:], value, unit)
     else:
@@ -1490,7 +1490,7 @@ def split(
 
         - in the NumPy style: ``Flight.split(10, 'm')`` (see
           ``np.timedelta64``);
-        - in the pandas style: ``Flight.split('10T')`` (see ``pd.Timedelta``)
+        - in the pandas style: ``Flight.split('10 min')`` (see ``pd.Timedelta``)
 
         If the `condition` parameter is set, the flight is split between two
         segments only if `condition(f1, f2)` is verified.

src/traffic/core/time.py

@@ -24,7 +24,7 @@
 _log = logging.getLogger(__name__)
 
 
-def to_datetime(time: timelike) -> datetime:
+def to_datetime(time: timelike) -> pd.Timestamp:
     """Facility to convert anything to a datetime.
 
     This function will soon be replaced by pd.to_datetime.
@@ -41,10 +41,10 @@ def to_datetime(time: timelike) -> datetime:
 
     if isinstance(time, str):
         time = pd.Timestamp(time, tz="utc")
-    if isinstance(time, pd.Timestamp):
-        time = time.to_pydatetime()
+    if isinstance(time, datetime):
+        time = pd.Timestamp(time)
     if isinstance(time, Real):
-        time = datetime.fromtimestamp(float(time), timezone.utc)
+        time = pd.Timestamp(float(time), tz="utc", unit="s")
     if time.tzinfo is None:  # coverage: ignore
         _log.warning(
             "This timestamp is tz-naive. Things may not work as expected. "

src/traffic/data/adsb/opensky.py

@@ -150,11 +150,6 @@ def format_history(
         if column_name in df.columns:
             df[column_name] = df[column_name].astype(float)
 
-    if "onground" in df.columns and df.onground.dtype != bool:
-        df.onground = df.onground == "true"
-        df.alert = df.alert == "true"
-        df.spi = df.spi == "true"
-
     # better (to me) formalism about columns
     df = df.rename(
         columns={