get_sunlight_coverage never completes at 2023-05-10 02:00

[gerritholl]

Describe the bug

When checking the sunlight coverage for a certain area with a certain area extent, get_sunlight_coverage never completes.

This stopped our operational production at 2023-05-10 02:00 UTC.

To Reproduce

(needs pytroll-schedule or it will fail differently)

import trollflow2.plugins
import datetime
import pyresample
ar = pyresample.create_area_def("test", 4326, area_extent=[-34.91577, 27.5636854, 54.8726391, 72.45789], shape=(1250, 2500))
trollflow2.plugins._get_sunlight_coverage(ar, datetime.datetime(2023, 5, 10, 2, 0))

Expected behavior

I expect _get_sunlight_coverage to complete.

Actual results

No output, but a KeyboardInterrupt reveals:

^CTraceback (most recent call last):
  File "/data/gholl/checkouts/protocode/mwe/trollflow2-area-issue.py", line 7, in <module>
    trollflow2.plugins._get_sunlight_coverage(ar, datetime.datetime(2023, 5, 10, 2, 0))
  File "/data/gholl/mambaforge/envs/py311/lib/python3.11/site-packages/trollflow2/plugins/__init__.py", line 703, in _get_sunlight_coverage
    daylight = cut_area_poly.intersection(poly)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 713, in intersection
    return self._bool_oper(other, -1)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 691, in _bool_oper
    inter, edge2 = edge1.get_next_intersection(narcs2, inter)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 496, in get_next_intersection
    inter = self.intersection(arc)
            ^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 476, in intersection
    for i in self.intersections(other_arc):
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 457, in intersections
    return (SCoordinate(lon, lat),
            ^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 131, in __init__
    lon, lat = _check_lon_lat(lon, lat)
               ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 119, in _check_lon_lat
    _check_lat_validity(lat)
  File "/data/gholl/checkouts/pyresample/pyresample/spherical.py", line 110, in _check_lat_validity
    if np.any(np.logical_or(lat > np.pi / 2, lat < -np.pi / 2)):
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<__array_function__ internals>", line 200, in any
  File "/data/gholl/mambaforge/envs/py311/lib/python3.11/site-packages/numpy/core/fromnumeric.py", line 2333, in any
    @array_function_dispatch(_any_dispatcher)

KeyboardInterrupt

Environment Info:

• OS: openSUSE 15.3
• Trollflow2 Version: 0.14.0
• Satpy Version: 0.40.1
• Pyresample Version: 1.24.0

Additional context

This particular combination of polygons triggers pytroll/pyresample#329, but I don't know why only then.

If I change the datetime to one second later or earlier, the problem does not occur.