Fix NumPy 2.4+ / pandas 3 compatibility

[cfuselli]

Why

Recent CI started failing after dependency refreshes (NumPy 2.4+, pandas 3 / pyarrow strings), with errors like:

• AttributeError: module 'numpy' has no attribute 'in1d'
• TypeError: Expected string or tuple of strings, got <class 'pandas.arrays.ArrowStringArray'>
• event display tests failing due to scalar vs 1-element array time handling.

What changed

• Replaced np.in1d with np.isin in straxen code and tests where applicable.
• Fixed _event_display to consistently use scalar event[...] fields (instead of events[...] array fields) for time ranges and titles.
• Added runtime compatibility layer in straxen/_compat.py and invoked it from straxen/__init__.py:
  • alias np.in1d to np.isin when missing (NumPy >= 2.4)
  • patch strax.to_str_tuple / strax.utils.to_str_tuple to accept pandas ExtensionArray (e.g. ArrowStringArray)
• Updated SCADA tests for newer pandas behavior:
  • timezone assertion compares string representation (zoneinfo vs pytz)
  • positional indexing now uses .iloc.

Notes

• Compatibility patches were moved out of inline __init__ logic into straxen/_compat.py to keep imports clean and satisfy pre-commit (flake8 E402).

[coveralls]

coverage: 88.965% (+0.01%) from 88.954%
when pulling 8dca461 on fix-numpy
into 310bd23 on master.