-
Notifications
You must be signed in to change notification settings - Fork 52
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Devtools::check() working on Windows! (+ add / fix some tensor stuff) #314
Conversation
- improve some utils and argument checking - tests still reporting 3F, 4W, 258P
save snapshot so can return after devtools::check() ruins the repo...
need to worry about writing better tensor docs in the future...
finnicky - allow specification of multiple observations of categorical label across time points, with potentially new classes introduced that were not present in previous time points
UPDATE: For the parallel workers, I think the partial import that devtools does during In the Github tests, is there a way to make sure Sorry my terrible function documenting is producing a chain of warnings, but heres the output from the completed
Let me know if this works for you too. If so I'll try work on refactoring some of the parallel code in the future when I have time, and once we are sure all the old tests pass as intended on all environments. |
Hey, just some stuff on new tensor functions here.
tplsda
function and unit testsmagick
to the renv.lock, I saw you added this to DESCRIPTION so I installed locally and sync'ed the lockfileLocal testing and building status
Test status is unchanged. Just more passing tests as more unit tests have been introduced. Might open a separate PR in the near future to try and refactor some of the parallel code so that it runs, although keen to get your opinion!
Thanks for adding
magick
into DESCRIPTION! With this, and installing pandoc locally into PATH,devtools::check
can move forward a bit more now on Windows before crashing again...devtools::check()
error:Which is complaining about the following example code in the vignette:
When I briefly checked the test failures, the error seemed to be related to the fact that the workers in SnowParam are unable to find the installed mixOmics library (because I think
devtools
just does some partial import when we invoke the various commands) and as a result the workers just crash.Surely this must be a common testing / documenting problem for libraries that have parallel code, so I'm hoping to have a look sometime in the future and maybe there's a somewhat straightforward fix.
I'm also keen to refactor as much parallel code as possible using BiocParallel...
For instance in
perf.diablo.R
there's still a cluster being created manually (which is also causing a test failure, although I think it's for the same reason as SnowParam() in the biocparallel code)which I am hoping can be rewritten using BiocParallel and snowparam() and multicoreparam() but probably need to sort out this BiocParallel issue first!