31101344291
80%

Ran 06 Aug 2026 12:44PM UTC

Jobs 1

Files 179

Run time 1min

Badge

Embed ▾

Committed 06 Aug 2026 12:25PM UTC coverage: 77.808% (+0.2%) from 77.602%

Build # 31101344291

Build Type

push

github

Committed by

hmgomes

Commit Message

docs(drift): explain what a concept length does and does not guarantee

`num_instances` places the drifts; it does not ration instances. Around an
AbruptDrift it is exact, because the switch is a step. Around a
GradualDrift the concepts overlap, and the default logistic ramp reaches
only 0.12 at the start of the window and 0.88 at the end, approaching 0
and 1 asymptotically -- so the older concept keeps appearing past the
nominal end and a concept declared as 500 instances can be drawn from more
often than that.

The drift *window* is exactly where the definition says. The *provenance*
of an individual instance near it is probabilistic, which matters for
anyone building a labelled benchmark.

Documented on Concept and GradualDrift, and shown in the notebook with the
ramp evaluated at the window boundaries.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Coverage Stats

8134 of 10454 relevant lines covered (77.81%)

0.78 hits per line

Coverage Regressions

Lines	Coverage	∆	File
30	89.12	2.79%	src/capymoa/stream/drift.py
1	99.65	0.05%	tests/test_stream.py

Jobs

ID	Job ID	Ran	Files	Coverage
1	31101344291.1	06 Aug 2026 12:44PM UTC	179	77.81	GitHub Action Run

adaptive-machine-learning / CapyMOA / 31101344291
80%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 31101344291

adaptive-machine-learning / CapyMOA / 31101344291 80%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 31101344291

adaptive-machine-learning / CapyMOA / 31101344291
80%

README BADGES
x