feat: conversion of ROOT histograms to PlottableHistogram

[pieterdavid]

Tested for TH1F by comparing as follows:

import ROOT
h = ROOT.TH1F("h1", "h1", 50, -2.5, 2.5)
h.FillRandom("gaus", 10000)
h.Draw()
from matplotlib import pyplot as plt
fig, ax = plt.subplots()
import mplhep
mplhep.histplot(h)
plt.show()

The conversion to numpy exploits the fact that TH1 inherits from TArray (which can directly be converted to a numpy array) for storing the counts, and the sums of squared weights is stored as another TArray (so most could be replaced by equivalent PyROOT pythonizations in the future).
For more-dimensional histograms I put order="F" in the reshape such that .values()[i,j] equals .GetBinContent(i+1, j+1), but didn't test beyond that. Overflow bins are dropped.

CC @andrzejnovak

[henryiii]

This is awesome, thanks! Bug me if I don't review it by the end of the week.

[henryiii]

PS: Sourcery (c34d809) recommends using "guard" style, which has the short-circuit branch come first. It's easier to read with less action at a distance, and avoids excessive indentation.

[andrzejnovak]

I was just wondering if this was already in, would be great to have by PyHEP :)

[henryiii]

Looks good! A couple of optional ideas:

• We could either make type aliases to improve readability for the ROOT types (TAxis = Any), or maybe even make them Protocols (TAxis would have GetNbins, etc). But it's really not any different from "Any".
• It's optional and beyond the strict protocol, but maybe we could promote fName to .name?

[pieterdavid]

Thanks for the review!

• We could either make type aliases to improve readability for the ROOT types (TAxis = Any), or maybe even make them Protocols (TAxis would have GetNbins, etc). But it's really not any different from "Any".

I'm not familiar enough with the type annotations, protocols etc. yet to tackle that, sorry, but feel free to add it.

• It's optional and beyond the strict protocol, but maybe we could promote fName to .name?

Good idea, done

[henryiii]

I've added a ROOT test system (you can even run it locally with nox, though it will be a bit slow until wntrblm/nox#444, and it will be a little harder to maintain unless wntrblm/nox#447 gets addressed). Could you add a test converting a ROOT histogram to the mostly-empty test file I've added?

[sourcery-ai]

Sourcery Code Quality Report

✅  Merging this PR will increase code quality in the affected files by 2.63%.

Quality metrics	Before	After	Change
Complexity	2.84 ⭐	2.39 ⭐	-0.45 👍
Method Length	30.91 ⭐	28.12 ⭐	-2.79 👍
Working memory	6.00 ⭐	5.46 ⭐	-0.54 👍
Quality	82.81% ⭐	85.44% ⭐	2.63% 👍

Other metrics	Before	After	Change
Lines	237	378	141

Changed files	Quality Before	Quality After	Quality Change
noxfile.py	89.48% ⭐	89.97% ⭐	0.49% 👍
src/uhi/numpy_plottable.py	80.09% ⭐	84.49% ⭐	4.40% 👍

Here are some functions in these files that still need a tune-up:

File	Function	Complexity	Length	Working Memory	Quality	Recommendation
src/uhi/numpy_plottable.py	ensure_plottable_histogram	15 🙂	192 😞	12 😞	41.34% 😞	Try splitting into smaller methods. Extract out complex expressions

Legend and Explanation

The emojis denote the absolute quality of the code:

• ⭐ excellent
• 🙂 good
• 😞 poor
• ⛔ very poor

The 👍 and 👎 indicate whether the quality has improved or gotten worse with this pull request.

---

Please see our documentation here for details on how these metrics are calculated.

We are actively working on this report - lots more documentation and extra metrics to come!

Help us improve this quality report!

[henryiii]

Beautiful, thank you!