Closed Datseris closed 11 months ago
Hey, excellent! Yup, I encountered this some time ago, and I fixed it with a similar idea to this next_id
way, but with much hackier way, and didn't have the time to improve the code to submit a PR.
This is very important, as you said. Glad it'll be fixed now! Thanks for the work :)
Merging #90 (1a99990) into main (cd0e857) will increase coverage by
5.19%
. Report is 4 commits behind head on main. The diff coverage is92.03%
.
@@ Coverage Diff @@
## main #90 +/- ##
==========================================
+ Coverage 68.61% 73.81% +5.19%
==========================================
Files 22 21 -1
Lines 1128 1199 +71
==========================================
+ Hits 774 885 +111
+ Misses 354 314 -40
Files Changed | Coverage Δ | |
---|---|---|
ext/plotting.jl | 0.00% <0.00%> (ø) |
|
src/Attractors.jl | 100.00% <ø> (ø) |
|
src/basins/basins_utilities.jl | 100.00% <ø> (ø) |
|
src/basins/fractality_of_basins.jl | 96.62% <ø> (-0.08%) |
:arrow_down: |
src/continuation/continuation_grouping.jl | 97.72% <ø> (ø) |
|
src/mapping/attractor_mapping_featurizing.jl | 62.29% <ø> (ø) |
|
src/mapping/attractor_mapping_recurrences.jl | 90.24% <ø> (+8.78%) |
:arrow_up: |
src/mapping/grouping/cluster_config.jl | 66.19% <ø> (ø) |
|
src/tipping/tipping_probabilities.jl | 100.00% <ø> (ø) |
|
src/continuation/continuation_recurrences.jl | 79.31% <50.00%> (-2.23%) |
:arrow_down: |
... and 3 more |
... and 2 files with indirect coverage changes
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
This re-works the way matching works in
RAFM
to account for a crucial mistake. Previously, if an attractor dissapeared, and a new one appeared later, then they would both be assigned the same ID even though they would have never been compared with each other.This new continuation matching allows for two options:
@kalelR this should be highly relevant for you if you use
RAFM
in scientific work.@kalelR or @awage please review.