[python][UHI] Implement Slicing in TH1 and adapt the UHI backend and tests

[siliataider]

This Pull request:

Changes or fixes:

Previously, slicing was implemented at the pythonization level: when slicing a histogram, the values within the specified slice were kept and the ones outside were set to 0, without resizing the sliced axis (leading to inconsistencies..).
This PR implements the slicing logic in TH1 directly, adapting the pythonizations to make use of the new internal functions.

Other changes are made to enhance the interface and fully comply with the UHI specifications, improvements can be summarized here:

• The sliced axis is resized after slicing, which correctly adjusts the statistics.
• Integration operations now support full-range integration, non-flow integration, and ranged integration.
• Histograms can now be set with either a scalar, numpy.ndarray, or any ArrayLike instance.
• Broadcasting is now also supported (e.g., h[0:2, 0:2] = np.array([[42], [3]])).
• Enhance error handling.

Checklist:

• tested changes locally

[silverweed]

I know I wasn't a requested reviewer but I have a couple of comments :)

[github-actions]

Test Results

    19 files      19 suites   3d 15h 49m 12s ⏱️
 2 758 tests  2 758 ✅ 0 💤 0 ❌
50 908 runs  50 908 ✅ 0 💤 0 ❌

Results for commit abc7b47.

♻️ This comment has been updated with latest results.

[vepadulano]

Thank you a lot for this prompt effort to improve the implementation according to bleeding edge testing! I have a few quick remarks for a first iteration of the review

[vepadulano]

Thanks for pushing for this! Just a few quick pass-by comments

[silverweed]

A couple of last comments :)

[silverweed]

lgtm, thanks!

[siliataider]

Thanks a lot @vepadulano and @silverweed for the thorough rounds of review, I really appreciate it :)

Approved verbally

[smuzaffar]

@siliataider , CMSSW DQM bin-by-bin comparison tests are failing with error [a] ( see cms-sw/cmssw#49452 ) . We see this error in both of our ROOT master based IBs and Ibs with ROOT 6.36. I guess the change here is causing this failure. Is there something we should fix on our side or this change introduce a bug?

[a]

PR file successfully opened
Traceback (most recent call last):
  File "/cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/49419/49598/CMSSW_16_0_X_2025-11-21-1100/bin/el8_amd64_gcc13/compareHistograms.py", line 345, in <module>
    create_dif(args.base_file, args.pr_file, args.pr_number, args.test_number, cmssw_version, args.num_processes, args.output_dir)
  File "/cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/49419/49598/CMSSW_16_0_X_2025-11-21-1100/bin/el8_amd64_gcc13/compareHistograms.py", line 80, in create_dif
    compare(shared_paths, pr_flat_dict, base_flat_dict, paths_to_save_in_pr, paths_to_save_in_base)
  File "/cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/49419/49598/CMSSW_16_0_X_2025-11-21-1100/bin/el8_amd64_gcc13/compareHistograms.py", line 182, in compare
    pr_array = np.array(pr_item)
  File "/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02916/el8_amd64_gcc13/lcg/root/6.36.05-302290958d85790e78b48654c2f216ba/lib/ROOT/_pythonization/_uhi.py", line 367, in _getitem
    uhi_index = _compute_common_index(self, index)
  File "/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02916/el8_amd64_gcc13/lcg/root/6.36.05-302290958d85790e78b48654c2f216ba/lib/ROOT/_pythonization/_uhi.py", line 223, in _compute_common_index
    return [_compute_uhi_index(self, idx, axis, include_flow_bins) for axis, idx in enumerate(index)]
  File "/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02916/el8_amd64_gcc13/lcg/root/6.36.05-302290958d85790e78b48654c2f216ba/lib/ROOT/_pythonization/_uhi.py", line 223, in <listcomp>
    return [_compute_uhi_index(self, idx, axis, include_flow_bins) for axis, idx in enumerate(index)]
  File "/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02916/el8_amd64_gcc13/lcg/root/6.36.05-302290958d85790e78b48654c2f216ba/lib/ROOT/_pythonization/_uhi.py", line 192, in _compute_uhi_index
    return _process_index_for_axis(self, index, axis)
  File "/cvmfs/cms-ib.cern.ch/sw/x86_64/nweek-02916/el8_amd64_gcc13/lcg/root/6.36.05-302290958d85790e78b48654c2f216ba/lib/ROOT/_pythonization/_uhi.py", line 180, in _process_index_for_axis
    raise IndexError(f"Histogram index {index - 1} out of range for axis {axis}. Valid range: (0,{nbins})")
IndexError: Histogram index 30 out of range for axis 0. Valid range: (0,30)

[smuzaffar]

@siliataider , should I open a github issue if there is no quick answer/solution the the problem I mentioned above

[siliataider]

hi @smuzaffar , thanks for reporting this, if you don't need to access/compare the contents of the underflow/overflow bins, you can use pr_item.values() to get a NumPy array of your bin values and proceed with your comparison (instead of np.array(pr_item), that shouldn't raise any indexing errors, in the meantime could you open an issue with a small reproducer? I will look into it

[smuzaffar]

@siliataider , see #20511