search

brittanyblouin / ANCRTAdjust

An R package to adjust routine HIV testing data from antenatal care to reduce bias in estimating HIV prevalence trends
MIT License
2 stars 3 forks source link

Testing #17

Open seabbs opened 4 years ago

seabbs commented 4 years ago

As discussed in the draft paper this package is aiming to be a standardised toolbox for cleaning a specific data format. To give others users confidence all functionality needs to be thoroughly tested (both at a high level with examples and at a low unit test level).

Whilst there is currently a test.R script as it currently stands it isn't really testing anything (apart from checking for errors). To act as a tests expected output needs to be given (or interpretation of plots).

I would like to see unit tests using testthat and vdiffr for all low level functionality. I would also like to see a high level walkthrough going through what is actually happening on an example dataset.

These tests also really need to be linked up to a CI service link Travis and have an integrated code coverage report (like codecov.io).

Here is an example of some unit tests I have previously implemented to get you started. Happy to give more details if needed.

Linked to #16

As mentioned by @ellessene in their review.

For this # openjournals/joss-reviews/issues/1740 review

m-maheu-giroux commented 4 years ago

Thank you for this suggestion. You are correct to have pointed out this omission in our first submission. We have implemented the main tests and everything is running smoothly.

seabbs commented 4 years ago

Great - thanks for implementing.

Looking at current testing it is hard to have an idea of coverage.

Can you link up to a CI and code coverage tool - see the original issue for details on this.

ellessenne commented 4 years ago

I also see there are no documented examples when running R CMD check - you may want to add some to the documentation.