Random sampler analysis class

[wedeling]

Made the computation of the confidence intervals more efficient, by vectorizing the for loop over the number of bootstrap samples.

[lgtm-com[bot]]

This pull request introduces 1 alert when merging fcff1c3fdf2b1c0e74c2f93acadd7a3e61a9a4a8 into 6ab8e9ecd3c93dd2d06c9e45f248afe982f3233c - view on LGTM.com

new alerts:

• 1 for Unused local variable

[orbitfold]

oh you will need to change test_mc_analysis.py, the [0] indexing just needs to be removed i think. since this changes how results are returned?

[wedeling]

It shouldn't matter I think. I'll have a look

[orbitfold]

>       assert(s1_conf['x1']['low'] == pytest.approx(0.14387035, 0.01))
E       assert 0.19341946180114922 == 0.14387035 ± 1.4e-03
E        +  where 0.14387035 ± 1.4e-03 = <function approx at 0x7f12acd2fa60>(0.14387035, 0.01)
E        +    where <function approx at 0x7f12acd2fa60> = pytest.approx

Is there some deeper issue here or just because of the changed order of the calls to the random number generator?

[lgtm-com[bot]]

This pull request introduces 1 alert when merging 8fa541a9cd01867582466268353bd90d8d029a48 into 6ab8e9ecd3c93dd2d06c9e45f248afe982f3233c - view on LGTM.com

new alerts:

• 1 for Unused local variable

[DavidPCoster]

@wedeling -- should this be fixed and merged? Or discarded? Work on this started more than 3 years ago ...

[wedeling]

@DavidPCoster Not sure, it's got quite some work in it, but's it's been 3 years, there might be conflicts with the current version. Also, the HDF5 collator probably deserves it's own PR.