Serializer fails on NaT (not a time in pandas)

[duetosymmetry]

Minimal example demonstrating the problem:

import pandas as pd
from wolframclient.serializers import export

# this has type pandas.Timestamp
good_time = pd.to_datetime('2026-02-25T17:30:16+00:00')
# this rather has type pandas.api.typing.NaTType
bad_time = pd.NaT

export(good_time)
# returns b'DateObject[{2026, 2, 25, 17, 30, 16.}, "Instant", "Gregorian", "UTC"]'

export(bad_time)
# raises the below ValueError

Here is the resulting exception:

----> 1 export(bad_time)

File ~/miniconda3/envs/test_wolframclient/lib/python3.14/site-packages/wolframclient/serializers/__init__.py:68, in export(data, stream, target_format, **options)
     63 if not target_format in available_formats:
     64     raise ValueError(
     65         "Invalid export format %s. Choices are: %s"
     66         % (target_format, ", ".join(available_formats.keys()))
     67     )
---> 68 return available_formats[target_format](**options).export(data, stream=stream)

File ~/miniconda3/envs/test_wolframclient/lib/python3.14/site-packages/wolframclient/serializers/base.py:35, in FormatSerializer.export(self, data, stream)
     32         stream.write(token)
     33     return stream
---> 35 return concatenate_bytes(self.generate_bytes(data))

File ~/miniconda3/envs/test_wolframclient/lib/python3.14/site-packages/wolframclient/serializers/wl.py:17, in yield_with_separators(iterable, first, last, separator)
     15     if i:
     16         yield separator
---> 17     for sub in arg:
     18         yield sub
     19 yield last

File ~/miniconda3/envs/test_wolframclient/lib/python3.14/site-packages/wolframclient/serializers/wl.py:17, in yield_with_separators(iterable, first, last, separator)
     15     if i:
     16         yield separator
---> 17     for sub in arg:
     18         yield sub
     19 yield last

File ~/miniconda3/envs/test_wolframclient/lib/python3.14/site-packages/wolframclient/serializers/wl.py:58, in WLSerializer.serialize_int(self, number)
     57 def serialize_int(self, number):
---> 58     yield b"%i" % number

ValueError: cannot convert float NaN to integer

This is different from how the exporter handles NaN (not a number):

import numpy as np

export(np.nan)
# returns b'Indeterminate'

Expected behavior: The serializer should detect NaT and treat it like NaN, returning Indeterminate.

EDITED to add: This affects both target_format='wl' and target_format='wxf'