fix: Step sizes for fixed parameters in minuit should be 0.0

[kratsg]

Description

This resolves a comment made in #1051 where step sizes were not being zero'd out correctly.

Suggestion: For the minuit optimizer, the initial step size is set in minuit_optimizer._get_minimizer. For a fixed Gaussian-constrained nuisance parameter with default bounds [-5, 5], this is 0.01 with the default 1000 steps. This will be the parameter uncertainty reported by minuit after the fit. Given that it does not make sense to have an uncertainty for a fixed parameter, it might be nice to set the entries of step_sizes corresponding to fixed parameters to 0 instead.

Originally posted by @alexander-held in #1051 (comment)

Checklist Before Requesting Reviewer

• Tests are passing
• "WIP" removed from the title of the pull request
• Selected an Assignee for the PR to be responsible for the log summary

Before Merging

For the PR Assignees:

• Summarize commit messages into a comprehensive review of the PR

* Step sizes are set to 0.0 for fixed parameters in a fit for minuit
   - The step size acts as the uncertainty for fixed parameters
* Add test for step sizes for fixed parameters

[codecov]

Codecov Report

Merging #1054 into master will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1054   +/-   ##
=======================================
  Coverage   96.78%   96.78%           
=======================================
  Files          59       59           
  Lines        3388     3389    +1     
  Branches      489      489           
=======================================
+ Hits         3279     3280    +1     
  Misses         68       68           
  Partials       41       41           

Flag	Coverage Δ
#unittests	96.78% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/pyhf/optimize/opt_minuit.py	100.00% <100.00%> (ø)

---

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 0f094a4...c567cd0. Read the comment docs.

[matthewfeickert]

LGTM. Thanks @kratsg. @alexander-held let us know if you have any comments here.

[alexander-held]

Thanks for including this! I rebased it into #1051 to test and it worked as expected.