The file allows for testing of the `plotArrow()` function. Two included tests check for general functionality on DIABLO objects as well as inability to function on `(mint).(s)plsda` objects

fix: `plotVar()` can now utilise `block.pls` and `block.spls` objects which have a single Y feature (ie. (s)PLS1)

fix: predict function has updated error messages for when feature sets are different or in different order

refactor: cleaned output of `test_that` tests to increase ease of debugging

fix: resolved issue preventing `perf()` from properly accounting for near-zero-variance features

feat: added `plot.graph` parameter to network to allow user control over whether to plot the network or just return the output. Either way, the returned object is the same

fix: if `validation = "loo"` in `tune.block.splsda()`, the parameter `n` would not be set. moved the lines which define `n` to outside of the  `if(validation= "Mfold")` statement so its run in all cases. Additionally, increased the specificity of some warnings and errors in this function

refactor: removed depreciated `plot.tune.splsda()`. Reduces package clutter and document repetition

tests: increased suffix length in `plotLoadings()` label-length check

test commit

test

* Attempted fix for issue #262

fix: adjusting action `setup-r` version call to `v2` rather than `master`

* Attempted fix for Issue #262

fix: adjusting another action `setup-r` version call to `v2` rather than `master`

* Fix for Issue #264

fix: updating `set-output` commands in `Get R/Bioc versions` action

* Enhancement for Issue #260

feat: added 'legend.title' parameter (takes a string) for `circosPlot()`. Allows for control over legend title

* Enhancement for Issue #260

docs: updated documentatation with `legend.title` parameter

* Fix for Issue #266

fix: added fail safe for when `Inf` or `-Inf` are found in transformed `newdata` data frame.

Changes them to `NaN` which can be safely handled by downstream functions

* Fix for Issue #266

tests: added test to maintain coverage

fix: improved nzv feature handling for block contexts, particularly via `auroc()`

Filtration applied more consistently via `Check.entry.wrapper.mint.block()` . Additional failsafe added here for zero variance features. `predict()` also now checks to see if filtration has been applied to prevent it applying filtering twice.


tests: adjusted new test to ensure it passes

fix: changed the depreciated `size` parameter to `linewidth` and changed `aes_string()` to `aes()`

fix: additional `aes_string()` call changed to `aes()`

fix: changing `aes_string()` to `aes()` introduced error. Adjusted `geom_path` calls so X and Y vectors are appropriately set

refactor: removed some commented lines

tests: added test for `plotIndiv.mint.plsda()` with `ellipse = TRUE`

docs: added my name to author list and updated maintainer

refactor: adjusted order of processes to increase efficiency in `repeat_cv_j()` within `tune.spca()`

enhance: added safety catch for NAs provided to `tune.spca()`. It replaces NAs with 0 and notifies user of the change

test: added new test file for `tune.spca()`

`perf()` now only runs T-test when `nrepeat` >= 3. Notifies user on unreliable results if `nrepeat` < 3. New `test` file for `perf()` functions to maintain coverage

Homogenised way in which `tune.mint.splsda()` and `perf.mint.splsda()` calculate BER, such that now both use weighted average of BER across studies. Added unit tests to ensure this homogenity

test: removed some `expect_equal()` states from `test-tune.spca()` as they were producing inconsistent values locally vs via GitHub Actions

build: added explicit global definition for `setNames()`. Resolves NOTE in R-CMD-Check

build: removed `middle` parameter from Al Abadi's reference in DESCRIPTION

The file allows for testing of the `plotArrow()` function. Two included tests check for general functionality on DIABLO objects as well as inability to function on `(mint).(s)plsda` objects

…into issue-150