Added test cases in curado.jl

[Jay-sanjay]

Added test cases in curado.jl for issue #85

[codecov[bot]]

Codecov Report

Merging #319 (9e66da0) into main (547b8fa) will not change coverage. Report is 1 commits behind head on main. The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main     #319   +/-   ##
=======================================
  Coverage   87.76%   87.76%           
=======================================
  Files          76       76           
  Lines        1920     1920           
=======================================
  Hits         1685     1685           
  Misses        235      235           

Files Changed	Coverage Δ
...c/encoding_implementations/combination_encoding.jl	100.00% <100.00%> (ø)
src/encoding_implementations/gaussian_cdf.jl	100.00% <100.00%> (ø)
src/encoding_implementations/ordinal_pattern.jl	86.79% <100.00%> (ø)
...rc/encoding_implementations/rectangular_binning.jl	94.44% <100.00%> (ø)
...lementations/relative_first_difference_encoding.jl	95.83% <100.00%> (ø)
...encoding_implementations/relative_mean_encoding.jl	94.44% <100.00%> (ø)

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more

[Jay-sanjay]

Hi all , as far as the codecov report the test-cases I added are not worth enough to increase the code coverage. I wanted to ask what can I do so that my tests creates the difference.

[kahaaga]

Hi, @Jay-sanjay! Thanks for your work on this PR.

The test coverage for Curado on the main branch is actually already at 100% (see https://app.codecov.io/gh/JuliaDynamics/ComplexityMeasures.jl/tree/main/src%2Finformation_measure_definitions). That's why you don't see any difference in the coverage for this PR.

The methods that are tested for every entropy definition is information and information_maximum. These are implemented specifically for the Curado entropy in the src/information_measure_definitions/curado.jl file (the information(::Curado, ...) and information_maximum(::Curado) methods).

As you can see in the test/infomeasures/infomeasure_types/curado.jl file, both these methods are already tested with a Curado instance as input. Therefore, the new tests in this PR don't actually contribute any fundamentally new test cases (with respect to methods being tested, since coverage of these methods are at 100% already).

What your contribution adds, however, is coverage for more input parameters, which is not (and cannot) be caught by CodeCov. Specifically,

• Your first block of tests check that we get the correct answer for a differently sized probability distribution, and more input parameters not already being tested for.
• Your second block is essentially identical to the existing tests (so a bit redundant), except that it is tested that normalization works as expected for Probabilities (which is also tested elsewhere).

I'm fine with merging these new tests, since they do increase analytical test coverage, and especially since it is aimed to address #85. Agree/disagree/comments, @Datseris?

[Jay-sanjay]

@kahaaga sir can you please tell from where can I know that which ones are the potential ones , where the testcase's needs to be added and have less code coverage...?

[kahaaga]

@Jay-sanjay I added a comment in #85 specifying how to find the latest coverage reports.

There's quite a few minor fixes to do scattered around the library. Some require quite a bit of technical insight into the methods, while others are easier (for example, verifying that errors are thrown for certain input parameters).

Please ignore the lack of test coverage for the multiscale.jl file - we are currently rethinking the multiscale code, so it doesn't have or need any tests at the moment.

[Datseris]

Sure, more tests is always better, as long as they are meaningful, even if they don't increase coverage.

[Jay-sanjay]

Thank you sir :)