Creation of test-persistence.R

[a1danbryant]

This PR contains the addition of tests/test-persistence.R, a test file to support the R/persistence-class.R file.
Tests include:

• compatibility with persistence diagrams from TDA and ripserr (including version-specific code paths)
• correct handling of multiple input types (matrix, list of matrices, etc.)
• conversion between persistence objects and data.frame
• behavior of the get_pairs()

A couple questions I have include:

• Is there any additional edge cases we should include?
• Are there any preferred style tweaks or conventions I should follow?

Looking forward to your feedback!

[corybrunson]

@a1danbryant i believe the new test file should be located alongside the existing test files in inst/tinytest/. Once that's done, are you able to run the tests on your machine using the {tinytest} package?

It might eventually become part of test-persistence-class.R, though since both contain many tests they might be kept separate instead, since they cover somewhat different realms of behavior.

[astamm]

Thanks for this PR.

The CI however fails. Can you please take a look?

[corybrunson]

And, to be sure, these are mostly derived from the previous version of the "persistence" class from {plt}, correct? So it would not be surprising to find some incompatibilities with this iteration of the class.

[a1danbryant]

And, to be sure, these are mostly derived from the previous version of the "persistence" class from {plt}, correct? So it would not be surprising to find some incompatibilities with this iteration of the class.

Yes, this has been adapted from the "persistence" from {plt}.

The CI however fails. Can you please take a look?

Looking into this now.

[a1danbryant]

After making moving the test-persistence.R to the correct location I found the error you were receiving (as_persistence() not being found). I ran devtools::load_all and that fixed this error, however it modified the files: R/cpp11.R and src/cpp11.cpp. Does this seem correct?

Also, it seems running the tests through here causes the handling for the package version of {ripserr} to err. I'm looking into this error now.

[a1danbryant]

Okay! Everything is working smoothly on my end now, sorry for the hold up! Let me know if any issues persist.

[astamm]

Thanks!

It still does not pass the CI. Now because of:

Error in alphaComplexDiag(df3d, maxdimension = 1) : 
  could not find function "alphaComplexDiag"

[a1danbryant]

Sorry about that-- I must have ran the test with the {TDA} package loaded.
With the above additions, here is the response from running:

tinytest::test_all("C:/Users/bryan/phutil/")

Console output:

> tinytest::test_all("C:/Users/bryan/phutil/")
test-distances.R..............   24 tests OK 1.1s
test-persistence-class.R......   44 tests OK 0.6s
test-persistence.R............   35 tests OK 44ms
test-utils.R..................    4 tests OK 
! Birth values are expected to be smaller than death values.
test-utils.R..................    5 tests OK 
! Birth values are expected to be smaller than death values.
test-utils.R..................    8 tests OK 0.2s
All ok, 111 results (2.0s)
<\details>``` 

[a1danbryant]

Hey @astamm just shooting you a follow up about this.

[corybrunson]

@astamm i don't think we can add {ripserr} to Suggests and later add {phutil} as a dependency of {ripserr}, as here. Could we use its output without the package itself, e.g. via dput()?

[astamm]

We can skip tests relying on ripserr if not installed, which would avoid the suggested dependency.

[corybrunson]

Is that right? I was under the impression it would still need to be listed somewhere in 'DESCRIPTION'. But if it works, great!