Use of deprecated `pandas` internal `DatetimeTZBlock` in for `Flight`

[spacewolfXfr]

When installing the latest version of traffic using conda, the Flight import fails due to the pandas library.

How to reproduce:

$ python
Python 3.12.13 | packaged by conda-forge | (main, Mar  5 2026, 16:50:00) [GCC 14.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import traffic
>>> traffic.__version__
'2.13'
>>> from traffic.core import Flight
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/mael/miniconda3/envs/ntraffic/lib/python3.12/site-packages/traffic/core/__init__.py", line 46, in <module>
    from .flight import Flight
  File "/home/mael/miniconda3/envs/ntraffic/lib/python3.12/site-packages/traffic/core/flight.py", line 41, in <module>
    from pandas.core.internals.blocks import DatetimeTZBlock
ImportError: cannot import name 'DatetimeTZBlock' from 'pandas.core.internals.blocks' (/home/mael/miniconda3/envs/ntraffic/lib/python3.12/site-packages/pandas/core/internals/blocks.py). Did you mean: 'DatetimeLikeBlock'?
>>> import pandas
>>> pandas.__version__
'3.0.1'

This is most likely since version 3.0.0 of pandas:

Removed Block, DatetimeTZBlock, ExtensionBlock, create_block_manager_from_blocks from pandas.core.internals and pandas.core.internals.api (GH 55139)

Based on this pandas issue, the fix would be to use pd.arrays.DatetimeArray (and in the meantime, force a pandas version to be less than 3.0.0)

[xoolive]

Thank you, it is fixed on HEAD actually
Otherwise if you are not ready to install from there; the other option is to keep pandas strictly <3.0

[spacewolfXfr]

Ok, sorry for that ! For now, I switched pandas do 2.3.3.