Repository navigation
Comparing changes
Open a pull request
base repository: ktsu-dev/Essentials
base: v2.10.2
head repository: ktsu-dev/Essentials
compare: v2.10.3
- 8 commits
- 8 files changed
- 3 contributors
Commits on Oct 7, 2026
-
Keep the binomial upper tail precise instead of rounding it to zero […
…patch] BinomialDistributionProvider fell back to IDistribution's 1 - Cdf for its survival function, so any upper tail below about 1e-16 came back as zero. Evaluate it directly as I(p; k + 1, n - k), as Poisson already does with its own complementary incomplete function. Fixes #50 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Vpup29Ycn8SGi7LvBjAq1c
Configuration menu - View commit details
-
Copy full SHA for 1314de8 - Browse repository at this point
Copy the full SHA 1314de8View commit details -
Keep the geometric CDF and quantile precise for a rare success [patch]
GeometricDistributionProvider worked with 1 - p directly, which rounds to exactly one below p of about 1e-16: the CDF was zero everywhere and the quantile's correction loop walked to int.MaxValue, about 22 s per call. Store log1p(-p) and build the CDF, survival function, mass and quantile on it through the shared ExpM1/Log1P, and bound the quantile's correction walk to two steps either way before falling back to bisecting the support. Fixes #51 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Vpup29Ycn8SGi7LvBjAq1c
Configuration menu - View commit details
-
Copy full SHA for ff302f4 - Browse repository at this point
Copy the full SHA ff302f4View commit details -
Cover the geometric quantile's bisection fallback directly
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Vpup29Ycn8SGi7LvBjAq1c
Configuration menu - View commit details
-
Copy full SHA for f1b6595 - Browse repository at this point
Copy the full SHA f1b6595View commit details -
Split the geometric mass function's nested conditional into statements
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Vpup29Ycn8SGi7LvBjAq1c
Configuration menu - View commit details
-
Copy full SHA for a0d19a1 - Browse repository at this point
Copy the full SHA a0d19a1View commit details -
Merge pull request #95 from ktsu-dev/fix/50-binomial-survival-function
Keep the binomial upper tail precise instead of rounding it to zero
Configuration menu - View commit details
-
Copy full SHA for 572588b - Browse repository at this point
Copy the full SHA 572588bView commit details -
[bot][skip ci] Update Metadata
Github Actions committedOct 7, 2026 Configuration menu - View commit details
-
Copy full SHA for ba7b59b - Browse repository at this point
Copy the full SHA ba7b59bView commit details -
Merge pull request #96 from ktsu-dev/fix/51-geometric-tiny-probability
Keep the geometric CDF and quantile precise for a rare success
Configuration menu - View commit details
-
Copy full SHA for db4b36a - Browse repository at this point
Copy the full SHA db4b36aView commit details -
[bot][skip ci] Update Metadata
Github Actions committedOct 7, 2026 Configuration menu - View commit details
-
Copy full SHA for 62720d1 - Browse repository at this point
Copy the full SHA 62720d1View commit details
This comparison is taking too long to generate.
Unfortunately it looks like we can’t render this comparison for you right now. It might be too big, or there might be something weird with your repository.
You can try running this command locally to see the comparison on your machine:
git diff v2.10.2...v2.10.3