add APS POLAR geometry

[prjemian]

• close add APS POLAR geometry #337

[prjemian]

Aack! Common CI failures (Py3.11) today at this line:

hklpy/hkl/tests/test_save_restore_UB.py

Line 30 in d295995

return databroker.temp().v2

Exception trace (fragment):

2024-08-07T14:50:13.8430400Z _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
2024-08-07T14:50:13.8430784Z 
2024-08-07T14:50:13.8431047Z descriptor = <slot wrapper '__call__' of 'type' objects>
2024-08-07T14:50:13.8431682Z obj = <property object at 0x7fde9025bd80>
2024-08-07T14:50:13.8431976Z 
2024-08-07T14:50:13.8432327Z     def _descriptor_get(descriptor, obj):
2024-08-07T14:50:13.8432770Z         if isclass(descriptor):
2024-08-07T14:50:13.8433213Z             return descriptor
2024-08-07T14:50:13.8433721Z         get = getattr(type(descriptor), '__get__', _sentinel)
2024-08-07T14:50:13.8434205Z         if get is _sentinel:
2024-08-07T14:50:13.8434673Z             return descriptor
2024-08-07T14:50:13.8435081Z >       return get(descriptor, obj, type(obj))
2024-08-07T14:50:13.8435789Z E       TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
2024-08-07T14:50:13.8436350Z 
2024-08-07T14:50:13.8436728Z ../../../micromamba/envs/hklpy-test-py-3.11/lib/python3.11/inspect.py:2432: TypeError
2024-08-07T14:50:13.8437451Z ___________________ ERROR at setup of test_no_primary_stream ___________________
2024-08-07T14:50:13.8437879Z 
2024-08-07T14:50:13.8438007Z     @pytest.fixture
2024-08-07T14:50:13.8438406Z     def cat():
2024-08-07T14:50:13.8438763Z >       return databroker.temp().v2
2024-08-07T14:50:13.8439026Z 
2024-08-07T14:50:13.8439173Z hkl/tests/test_save_restore_UB.py:30: 

Occurs in many places:

=========================== short test summary info ============================
FAILED hkl/tests/test_util.py::test_RE_documents[None] - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
FAILED hkl/tests/test_util.py::test_RE_documents[short] - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
FAILED hkl/tests/test_util.py::test_RE_documents[long] - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_fourc_orientation_save - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_fourc_run_orientation_info - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_list_orientation_runs - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_no_primary_stream - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_missing_energy_key - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_restore_orientation - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_save_restore_UB.py::test_restore_sample - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
ERROR hkl/tests/test_util.py::test_issue215 - TypeError: descriptor '__call__' for 'type' objects doesn't apply to a 'property' object
============= 3 failed, 481 passed, 1 warning, 8 errors in 36.04s ==============

This is a new failure. Something different upstream?

[prjemian]

note:

2024-08-07T14:48:52.7701679Z   databroker                        2.0.0b45      pyhd8ed1ab_0             conda-forge

Just might be the common factor. These tests should be revised:

hkl/tests/test_save_restore_UB.py:    return databroker.temp().v2
hkl/tests/test_util.py:    yield databroker.temp().v2
hkl/tests/test_util.py:    cat = databroker.temp().v2

And some docs will need attention, as well.

docs/source/examples/notebooks/geo_e4cv.ipynb:    "db = databroker.temp().v1\n",
docs/source/examples/notebooks/geo_e6c.ipynb:    "db = databroker.temp().v1\n",
docs/source/examples/notebooks/geo_k4cv.ipynb:    "db = databroker.temp().v1\n",
docs/source/examples/notebooks/tst_UB_in_descriptor_document.ipynb:    "Use a local, temporary, file-based databroker.  It will reset after each restart of the notebook.  Prepare to define the diffractometers needed here plus some items from the ophyd simulators."
docs/source/examples/notebooks/tst_UB_in_descriptor_document.ipynb:    "cat = databroker.temp().v2\n",
docs/source/examples/notebooks/tst_UB_in_stream.ipynb:    "cat = databroker.temp().v2\n",

[prjemian]

Also, upstream package dask is not ready for numpy v2

../../../../../../micromamba/envs/hklpy-test-py-3.11/lib/python3.11/site-packages/dask/array/routines.py:1569: in <module>
    @implements(np.round, np.round_)
../../../micromamba/envs/hklpy-test-py-3.11/lib/python3.11/site-packages/numpy/__init__.py:397: in __getattr__
    raise AttributeError(
E   AttributeError: `np.round_` was removed in the NumPy 2.0 release. Use `np.round` instead.

[prjemian]

2024-08-07T15:19:34.3789947Z hkl/tests/test_util.py:11: in <module>
2024-08-07T15:19:34.3790495Z     from tiled.utils import safe_json_dump
2024-08-07T15:19:34.3791122Z E   ModuleNotFoundError: No module named 'tiled'

and

(bluesky_2024_2) prjemian@arf:~/.../Bluesky/hklpy$ git grep tiled
hkl/tests/test_util.py:from tiled.utils import safe_json_dump
hkl/tests/tools.py:from tiled.utils import safe_json_dump
pyproject.toml:  "tiled",

tiled is not needed by the hklpy source code, only the safe_json_dump() function for testing. Can it be vendored?

[prjemian]

@padraic-shafer @mrakitin - This is ready for review!

[padraic-shafer]

Looks reasonable. I'm wondering about the unusual version spec for dependencies though.

[padraic-shafer]

Why not <2 ?

[prjemian]

In Python version comparisons: 2.0.0b46 < 2, we need to be below anything with a 2 in the major position, even pre-releases such as 2.0.0b46.

[padraic-shafer]

Ahh, pre-releases. I thought that would not be used without explicitly requesting --allow-prerelease.

I suppose it comes by default because 2.0 is only available as a pre-release. That behavior will presumably go away when 2.0 gets released.

[prjemian]

It's on conda-forge. Might be that pip respects pre-releases. Certainly, conda has stepped firmly into that pothole.

[padraic-shafer]

Why not <2 ?