Alexander Held

Wildcards in CLI workspace pruning

1

# Description It would be useful to support wildcards in the CLI when pruning workspaces. ## Is your feature request related to a problem? Please describe. Pruning similarly named parameters...

TypeError when plotting 2d histograms with flow="show"

2

Plotting a 2d histogram with `flow="show"` results in errors like this: > TypeError: Dimensions of C (12, 12) should be one smaller than X(11) and Y(11) while using shading='flat' see...

**Describe the bug** When wrapping an existing histogram object by `hist.Hist()`, I expected that I can set metadata (such as axis labels) in the `hist.Hist()` call. This currently does not...

### Summary `model.config.suggested_bounds()` usually is a list of tuples, but the list elements become lists themselves when using a custom parameter configuration. That doesn't break any functionality I am aware...

Documentation for staterror modifier

4

# Question The staterror documentation https://scikit-hep.org/pyhf/likelihood.html#mc-statistical-uncertainty-staterror shows a modifier example: ```json { "name": "mod_name", "type": "staterror", "data": [0.1] } ``` but does not clarify what exactly it is that should...

Ideas for navigation in `browse`

1

I imagine that an important use case for `uproot-browser browse` might be quick checks of a produced ROOT file, where the user already knows fairly well what they are looking...

The `visualize.templates` functionality previously only covered `NormPlusShape` systematics, for which template histograms exist. This generalizes the functionality to also support `Normalization` systematics by scaling the nominal histograms by the relevant...

### Discussed in https://github.com/scikit-hep/cabinetry/discussions/472 Originally posted by **Tomoya-Iizawa** April 19, 2024 I think cabinetry.visualize.templates produces plots for shape uncertainties and not for normalization uncertainties. Is there any way to produce...

Support for Python 3.12

1

I do not expect any changes needed in `cabinetry` to support Python 3.12. Currently not all backends of `pyhf` support that version, so official support in `cabinetry` should follow `pyhf`:...

Add back TensorFlow backend tests

1

#468 removed TensorFlow temporarily from the tests using all available `pyhf` backends. This is due to `tensorflow` / `tensorflow-probability` diverging and addressed in https://github.com/scikit-hep/pyhf/pull/2452. Add back this backend to the...