Remove redundant test on trunc_probs()
#970
Conversation
these aspects are already covered by the other test
| probs_2 <- probs_1 | ||
| probs_2[3] <- NA_real_ | ||
|
|
||
| expect_equal(parsnip:::trunc_probs(probs_1, 0), probs_1) |
There was a problem hiding this comment.
this covers "leave probs untouched" and "check the full vector", both of which are covered by
parsnip/tests/testthat/test-survival-censoring-weights.R
Lines 18 to 22 in ab42409
| probs_2[3] <- NA_real_ | ||
|
|
||
| expect_equal(parsnip:::trunc_probs(probs_1, 0), probs_1) | ||
| expect_equal(parsnip:::trunc_probs(probs_2, 0), probs_2) |
There was a problem hiding this comment.
this one adds an NA which is also part of the following, the whole vector is just covered over multiple expectations
parsnip/tests/testthat/test-survival-censoring-weights.R
Lines 13 to 16 in ab42409
| expect_equal( | ||
| parsnip:::trunc_probs(probs_1, 0.1), | ||
| ifelse(probs_1 < 0.05 / 2, 0.05 / 2, probs_1) | ||
| ) |
There was a problem hiding this comment.
this checks for what I called "data derived truncation" in
parsnip/tests/testthat/test-survival-censoring-weights.R
Lines 8 to 11 in ab42409
| expect_equal(min(parsnip:::trunc_probs(probs_2, 0.1), na.rm = TRUE), 0.05 / 2) | ||
| expect_equal(is.na(parsnip:::trunc_probs(probs_2, 0.1)),is.na(probs_2)) | ||
| }) |
There was a problem hiding this comment.
This one checks that there is an NA in the right place and the data-derived truncation value kicks in. This is also checked in
parsnip/tests/testthat/test-survival-censoring-weights.R
Lines 13 to 16 in ab42409
|
This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue. |
I think the test removed in this PR is mainly to test aspects which are already covered by the other test on
trunc_probs(). I'll comment with links :)