Use precomputed mean/std to compute skewness and kurtosis

[maxnoe]

master:

334 µs ± 1.48 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

this:

55.7 µs ± 549 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)

[dneise]

I like it

[watsonjj]

These are simple mathematical operations and has couple of temporary arrays....you could check what further performance you get with numba's njit, and maybe with (fastmath=True)

[maxnoe]

I tried with njit gets us another 10 %.

[codecov]

Codecov Report

Merging #1349 into master will decrease coverage by 0.06%.
The diff coverage is 88.00%.

@@            Coverage Diff             @@
##           master    #1349      +/-   ##
==========================================
- Coverage   91.10%   91.04%   -0.07%     
==========================================
  Files         183      184       +1     
  Lines       12548    12609      +61     
==========================================
+ Hits        11432    11480      +48     
- Misses       1116     1129      +13     

Impacted Files	Coverage Δ
ctapipe/__init__.py	100.00% <ø> (ø)
ctapipe/calib/camera/calibrator.py	100.00% <ø> (ø)
ctapipe/calib/camera/gainselection.py	96.55% <0.00%> (ø)
...tapipe/coordinates/tests/test_engineering_frame.py	100.00% <ø> (ø)
ctapipe/coordinates/tests/test_roundtrip.py	100.00% <ø> (ø)
ctapipe/core/__init__.py	100.00% <ø> (ø)
ctapipe/core/logging.py	100.00% <ø> (ø)
ctapipe/core/support.py	100.00% <ø> (ø)
ctapipe/core/tool.py	89.93% <ø> (ø)
ctapipe/image/__init__.py	100.00% <ø> (ø)
... and 247 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 23bf356...aed86a1. Read the comment docs.

[maxnoe]

Added more tests and njit, since it gave a 10 % improvement. Pls. review again.

[kosack]

can you run "black", or better yet try "pre-commit install" and see if it works properly ("pre-commit run" to manually run it). Mainly just to be pep 257-complant:

"For consistency, always use """triple double quotes""" around docstrings."

Otherwise looks fine! Nice to get a bit more speed without much effort.

[maxnoe]

@kosack done

[maxnoe]

pytest cov seems to be unable to work with the njit functions. This is unfortunate but also not really surprising in hindsight.

This is way diff coverage is so low