daniel1noble / BACE / 25331923558
91%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 15
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 71.444% (+4.0%) from 67.466%
25331923558

push

github

daniel1noble 
Phase C: targeted model_functions.R internals tests

Adds tests/testthat/test-model_functions.R with 7 test_thats / 30
assertions, targeting the largest uncovered chunk in the package
(model_functions.R, was 35% covered with 655 uncovered LOC).

Coverage delta after this commit (covr local with NOT_CRAN=true):
  model_functions.R   34.8% -> 51.6%   (+169 covered LOC)
  bace_imp.R          86.0% -> 87.6%
  Overall             67.8% -> 71.6%

Tests grouped by what they exercise:

  bace_imp(ovr_categorical = FALSE) on a K=3 categorical trait:
    The multinomial-probit categorical path
    (.pred_cat / .pred_cat_iter / .pred_cat_forward /
    .pred_cat_forward_iter) was completely unreachable from existing
    tests because every other test used the OVR fallback (J binary
    threshold fits) which routes through threshold helpers. This
    test forces ovr_categorical = FALSE on a real K=3 factor and
    verifies imputed values are valid factor levels.

  bace_imp on a K=4 ordinal trait:
    Exercises .pred_threshold_forward_iter at K > binary --
    previous tests only ran K=2 binary thresholds.

  .cat_mode unit tests:
    Most-frequent / NA-handling edge cases. Direct unit tests
    (no MCMCglmm fit needed) so they're fast.

  .impute_levels(sample = TRUE) path:
    test-level_alignment.R covers the sample = FALSE / argmax path;
    this fills in the sample = TRUE branch on a name-aligned
    probability matrix.

  .make_prior across types and gelman options:
    gelman = 0 path on each of gaussian / poisson / categorical /
    threshold; gelman = 2 path on categorical (which needs fixform +
    data per the type-specific validation); n_rand = 3 input-
    validation error.

  .align_levels_to_probs identity + observed-level paths:
    test-level_alignment.R covers the warning/fallback; this covers
    the no-op (n_cols == declared) and observed-equals-declared
    paths.

R CMD check: 0 errors / 0 warnings / 1 timestamp NOTE (network
issue verifying ... (continued)

2622 of 3670 relevant lines covered (71.44%)

54.27 hits per line

Jobs
ID Job ID Ran Files Coverage
1 25331923558.1 15
71.44
 GitHub Action Run
Source Files on build 25331923558