Closed lrnv closed 2 months ago
Attention: Patch coverage is 96.15385%
with 1 lines
in your changes are missing coverage. Please review.
:exclamation: No coverage uploaded for pull request base (
main@647aa41
). Click here to learn what that means.:exclamation: Current head f9f7309 differs from pull request most recent head 7a8367f. Consider uploading reports for the commit 7a8367f to get more accurate results
Files | Patch % | Lines |
---|---|---|
src/NPNSEstimator.jl | 94.44% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
@rimhajal This simplifies the code but also changes the structure of the output of the pohar perme fitting when there are covariates, in particular when there is several covariates.
Maybe you could tell me what you think of the output of :
f1 = fit(PoharPerme, @formula(Surv(time,status)~sex), colrec, frpop)
f2 = fit(PoharPerme, @formula(Surv(time,status)~(sex+stage)), colrec, frpop)
Edit: Wait a bit I'm trying other output formats, I will ping you when i am done
that way people can tell what dataframe refers to which group, this is good!
julia> x.age 0x0000000000000002
julia> x.keys 16-element Vector{@NamedTuple{sex::Symbol, stage::Int64}}:
(sex = :male, stage = 1)
(sex = :female, stage = 3)
(sex = :female, stage = 99)
(sex = :male, stage = 99)
(sex = :male, stage = 2)
(sex = :male, stage = 3)
(sex = :female, stage = 1)
(sex = :female, stage = 2)
there are some weird outputs though
@rimhajal Now the docs are adapted so this should also be good to go.
Hey I tried to do a bit of simplifications of the fit() code, I hope its more readable like that.