perf: Various improvements to chisquare

[FreezyLemon]

The current code always allocates a new f_exp: Vec<f64> just to iterate through it later. This is not necessary:

• if f_exp is Some(..), we can iterate through the slice itself
• if f_exp is None, we don't need to iterate anything. we can just set a fixed value for e and use that.

This has the added benefit of being no_std-compatible.

The amount of iterations through f_obs can also be reduced in some cases. If we need to f_obs.iter().zip(f_exp) anyways, there's no reason to do f_obs.iter().sum() beforehand. We can just calculate the sum "on the way", so to speak. The same goes for f_exp.

EDIT: I used approx::relative_eq since nothing comparable exists in crate::prec yet (on master). When #325 is merged, this should be updated to use the prec equivalent of relative_eq.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.99%. Comparing base (526e85e) to head (0af8efb).
Report is 8 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #336      +/-   ##
==========================================
- Coverage   95.01%   94.99%   -0.02%     
==========================================
  Files          60       61       +1     
  Lines       13529    13615      +86     
==========================================
+ Hits        12854    12934      +80     
- Misses        675      681       +6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[YeungOnion]

whoops! used the web editor for the merge and missed the closing brace. Will fix and retry.

[YeungOnion]

oh man, I boofed it. I think I'll need you to fix it. Missing semicolon closing src/stats_tests/chisquare.rs:90

[FreezyLemon]

No worries