Conversation
shawndove
force-pushed
the
outsource_uncertainty
branch
from
183b767 to
d3dcdd2
Compare
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #96 +/- ##
==========================================
- Coverage 75.74% 75.19% -0.55%
==========================================
Files 65 66 +1
Lines 3982 4051 +69
==========================================
+ Hits 3016 3046 +30
- Misses 966 1005 +39 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
This update moves the uncertainty calculation to a separate step of the workflow. Now all indicators except Hill diversity (which calculates internally using iNEXT) are calculated without CIs. The user can then add them using the add_ci() function, which defaults to bootstrapping the whole dataset using dubicube, but also provides the option to do it at the indicator level using the internal code. |
wlangera
left a comment
wlangera
left a comment
There was a problem hiding this comment.
Hi Shawn,
Thanks for the quick work. I doubt I will have time to do a complete review in January, because of my other deliverables.
Therefore, I already made some comments so you can work further on this if you want.
We should also have unit tests and perhaps a vignette. This vignette could also just be conceptual with links to the dubicube package.
Let me know what you think
… summaries in get_bootstrap_ci
…and updating documentation
pretty(NA) returns numeric(0), not logical(0). This resolves a test failure in R-devel where strict type checking caught the discrepancy.
…-specific bootstrapping - Species-level indicators (spec_occ, spec_range) now use group-specific logic. - Aggregate indicators use whole-cube logic. - Added critical guideline regarding manual ignore file management.
wlangera
left a comment
wlangera
left a comment
There was a problem hiding this comment.
I have updated dubicube.
Here is a review of the documentation.
| #' * `indicator`: Bootstrapping is done by resampling indicator | ||
| #' values. This is faster for large cubes but less robust. | ||
| #' | ||
| #' @param ci_type (Optional) Type of bootstrap confidence intervals to |
There was a problem hiding this comment.
"norm" is not a good default. Please use the order in dubicube:
type = c("perc", "bca", "norm", "basic") (perc can be the default)
https://b-cubed-eu.github.io/dubicube/reference/calculate_bootstrap_ci.html
| #' `dubicube::bootstrap_cube()`. (Default: `list()`) | ||
| #' @param ci_args (Optional) Named list of additional arguments passed to | ||
| #' `dubicube::calculate_bootstrap_ci()`. (Default: `list()`) | ||
| #' @param ... Additional arguments passed to the underlying functions. |
There was a problem hiding this comment.
What does this mean? Where can we find this underlying function? (It is not meantioned in the @params)
| #' @param ... Additional arguments passed to the underlying functions. | ||
| #' | ||
| #' @details | ||
| #' The function acts as a bridge to the `dubicube` package for statistical |
There was a problem hiding this comment.
| #' The function acts as a bridge to the `dubicube` package for statistical | |
| #' The function acts as a bridge to the \pkg{dubicube} package for statistical |
| #' | ||
| #' @details | ||
| #' The function acts as a bridge to the `dubicube` package for statistical | ||
| #' heavy lifting. |
There was a problem hiding this comment.
what does statistical heavy lifting mean? It is used to calculate bootstrap confidence intervals
2 participants
No description provided.