Closed oboulant closed 3 years ago
Merging #190 (dc878fb) into master (d211c5c) will not change coverage. The diff coverage is
n/a
.
@@ Coverage Diff @@
## master #190 +/- ##
=======================================
Coverage 96.31% 96.31%
=======================================
Files 40 40
Lines 978 978
=======================================
Hits 942 942
Misses 36 36
Impacted Files | Coverage Δ | |
---|---|---|
src/ruptures/detection/binseg.py | 100.00% <ø> (ø) |
|
src/ruptures/detection/bottomup.py | 100.00% <ø> (ø) |
|
src/ruptures/detection/kernelcpd.py | 100.00% <ø> (ø) |
|
src/ruptures/detection/pelt.py | 100.00% <ø> (ø) |
|
src/ruptures/detection/window.py | 100.00% <ø> (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 d211c5c...dc878fb. Read the comment docs.
In cases where the
n_bkps
is not explicitly specified, currentsanity_check()
usage does not reflect how it should be used. Indeed, in some cases, finding 0n_bkps
is a fine results. Therefore, we should not make the assumption that there is at least 1 break point in the following cases :Binseg
: if then_bkps
is not specified atpredict()
time, then allow for 0 break pointsBottomUp
: if then_bkps
is not specified atpredict()
time, then allow for 0 break pointsWindow
: if then_bkps
is not specified atpredict()
time, then allow for 0 break pointsPelt
: allow for 0 break pointsThere is one special case :
KernelCPD
: our C implementation make the assumption that when it is called, we are looking for at least 1 break point. Therefore, for this class, even if "in theory" for the Pelt option allows for 0 break point, we keep thesanity_checks()
usage to reflect implementation choices.Tests are also modified accordingly.
closes https://github.com/deepcharles/ruptures/issues/188