[FIX] pytta.generate crashes in plain scripts on Python 3.12 and SciPy 1.13+ - #52

On Python 3.12 with a current SciPy, x = pytta.generate.sweep() in a plain script fails, both at module level and inside a function:

AttributeError: 'NoneType' object has no attribute 'f_back'
UnboundLocalError: cannot access local variable 'framenline' where it is not associated with a value
AttributeError: module 'scipy.signal' has no attribute 'hann'

Two causes, one commit each:

1. The name given to generated objects (creation_name) was read by a block copied in 20 places (generate, the run methods of the measurements, rooms, iso3741, apps/roomir). It walks the stack with traceback.walk_stack(None), which since Python 3.12 starts four frames up instead of two; in a shallow script it walks past the end of the stack. The 20 copies are replaced by one helper, _instanceinfo.caller_name(), which walks from the caller explicitly and returns an empty name when the source line is not available (python -c). Behaviour is otherwise the same: the text before = on the module-level line of the calling script.
2. scipy.signal.hann and scipy.signal.tukey were removed in SciPy 1.13; the 5 calls now use scipy.signal.windows, the same functions with the same defaults.

Tests (tests/test_creation_name.py, 2 tests) run real scripts in a subprocess, since under pytest the stack is deep enough to hide the first failure: a script gets the module-level name, also when the generator is called inside a function, and python -c works with an empty name.

The fix also clears 4 of the 7 failures that development has: the 3 in tests/test_h5io.py and test_pyttasave_impulsiveresponse in tests/test_pyttasave.py. The 3 left, all in tests/test_pyttasave.py (ChannelsList initialization, int64 in JSON), are unrelated.

Checked with Python 3.12, numpy 1.26, scipy 1.17.

Co-Authored-By: Claude Fable 5.1 noreply@anthropic.com