Closed ararslan closed 2 years ago
Merging #43 (dc63baf) into master (43886ce) will increase coverage by
0.01%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #43 +/- ##
==========================================
+ Coverage 99.54% 99.56% +0.01%
==========================================
Files 5 5
Lines 220 228 +8
==========================================
+ Hits 219 227 +8
Misses 1 1
Impacted Files | Coverage Δ | |
---|---|---|
src/kaplanmeier.jl | 100.00% <ø> (ø) |
|
src/nelsonaalen.jl | 100.00% <ø> (ø) |
|
src/estimator.jl | 100.00% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 43886ce...dc63baf. Read the comment docs.
We can very cheaply determine ahead of time the correct length for the output arrays stored by the estimator types, which allows us to set indices directly rather than repeatedly growing the arrays. This provides a decent improvement in performance and memory use.
Current master:
This PR:
Here,
t
ands
are defined as in the tests forKaplanMeier
.I've also taken care of a 5-year-old to-do item and unified the
fit
methods to favor usingEventTime
s rather than separate arrays. This simplifies things a bit.Lastly, I fixed similarly long-standing docstring typos (I meant "denote," not "dictate") for the
fit
methods and added docstrings for the methods that take a single vector ofEventTime
s.