Maintain original feature and target names after shuffling

[Plutone11011]

Fixing Issue #375

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 35.54%. Comparing base (d910389) to head (8d9cf61).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #377      +/-   ##
==========================================
+ Coverage   35.42%   35.54%   +0.11%     
==========================================
  Files          96       96              
  Lines        6379     6381       +2     
==========================================
+ Hits         2260     2268       +8     
+ Misses       4119     4113       -6     

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

[relf]

Thanks for the fix.

[relf]

In an example, it would be better to have assertions with explicit values rather than println!() which requires the user to run the example.

As the purpose of the function is rather straightforward and as checking names in an example of shuffle() seems a bit contrived, here I would simply go with a test in mod.rs instead of an example. What do you think?

[Plutone11011]

I agree, it does seem contrived to check for names in shuffle, but since this is the purpose of this PR, do you think I should still add an assertion for feature and target names in the test in mod.rs?

[relf]

Yes, my point was to remove the example and add a test of shuffle() in mod.rs testing also you've fixed the bug, something like:

    #[test]
    fn test_dataset_shuffle() {
        let mut rng = SmallRng::seed_from_u64(42);
        let f_names = vec!["1", "2", "3"];
        let t_names = vec!["one"];
        let dataset = Dataset::new(
            array![[1., 2., 3.], [4., 5., 6.], [7., 8., 9.]],
            array![0., 1., 3.],
        )
        .with_feature_names(f_names.clone())
        .with_target_names(t_names.clone());

        let shuffled = dataset.shuffle(&mut rng);

        assert_abs_diff_ne!(dataset.records(), shuffled.records());
        assert_abs_diff_ne!(dataset.targets(), shuffled.targets());
        assert_eq!(f_names, shuffled.feature_names());
        assert_eq!(t_names, shuffled.target_names());
    }

[relf]

Thanks!