• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

wwu-mmll / confound_corrected_cpm / 35718134144
88%
main: 88%

Build:
Build:
LAST BUILD BRANCH: develop
DEFAULT BRANCH: main
Ran 22 Sep 2026 11:02AM UTC
Jobs 1
Files 37
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

22 Sep 2026 10:47AM UTC coverage: 87.267% (-0.1%) from 87.395%
35718134144

push

github

NilsWinter
Suppress meaningless increments; drop the deprecation shims for a clean 0.7.0

Two things Nils asked about.

**increment for Pearson r was still being computed.** The plan called for
suppressing it and I had not implemented it -- calculate_final_cv_results
subtracted full minus covariates across the whole metrics axis indiscriminately.
`increment` is a difference of two metrics, which only means something where
differences of that quantity are standard. It now covers explained variance, the
error metrics (as error reduction), accuracy, balanced accuracy and ROC AUC, and
is NaN for Pearson r -- comparing correlations needs Fisher z or Steiger's test,
not subtraction -- and for F1, a harmonic mean whose difference has no
established interpretation. The plan listed only accuracy and roc_auc for
classification; balanced accuracy is kept because it is a mean of two
proportions, so its difference is as standard as accuracy's.

The report rendered these as the literal string "nan". Table cells that are NaN
by design -- a suppressed increment, or a model a run does not define -- now
render as an em dash. Verified zero occurrences of "nan" in the rendered text.

**No deprecation shims.** The package has one author and a few students, the
results format breaks regardless (Models.residuals is gone from every CSV), and
the two shims were not equal in risk: edge_statistic was a pure rename, but
calculate_residuals *changed meaning*, so shimming it meant code that ran fine
and silently produced different numbers -- the exact failure mode this codebase
keeps getting bitten by, against which a DeprecationWarning is weak. So both are
removed, with errors naming the replacement, and 'pearson'/'spearman' carry over
untouched since they meant the same thing under the old name.

Version 0.5.0 -> 0.7.0 (0.6.0 was Part I's designation; it was never released
separately).

Suite 339 -> 344.

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

18 of 19 new or added lines in 4 files covered. (94.74%)

62 existing lines in 5 files now uncovered.

2481 of 2843 relevant lines covered (87.27%)

0.87 hits per line

Uncovered Changes

Lines Coverage ∆ File
1
93.4
1.4% src/cccpm/reporting/reporting_utils.py

Coverage Regressions

Lines Coverage ∆ File
38
59.31
0.0% src/cccpm/reporting/plots/plots.py
11
95.18
-0.88% src/cccpm/cpm_analysis.py
9
93.01
0.27% src/cccpm/edge_selection.py
3
97.58
-0.34% src/cccpm/models/linear_model.py
1
97.14
0.0% src/cccpm/constants.py
Jobs
ID Job ID Ran Files Coverage
1 35718134144.1 22 Sep 2026 11:02AM UTC 37
87.27
GitHub Action Run
Source Files on build 35718134144
  • Tree
  • List 37
  • Changed 10
  • Source Changed 10
  • Coverage Changed 8
Coverage ∆ File Lines Relevant Covered Missed Hits/Line
  • Back to Repo
  • Github Actions Build #35718134144
  • 4c1ffb49 on github
  • Prev Build on develop (#35615129980)
  • Next Build on develop (#35736986668)
  • Delete
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc