Handle p=1 in NegativeBinomial::ln_pmf like Binomial

[teddytennant]

NegativeBinomial::ln_pmf multiplies k * ln(1-p). At p = 1 and k = 0 that is 0 * (-inf), so pmf(0) is NaN instead of 1.

That is the empty product (1-p)^0. Binomial already special-cases p == 1 as a point mass at n rather than evaluating the same 0/0. Do the same here: at p == 1, ln_pmf(0) = 0 and every other k is -inf.

test_pmf_p_one / test_ln_pmf_p_one cover integer and non-integer r. The old test_is_nan expectations for those points are gone.

CARGO_TARGET_DIR=... cargo test --lib distribution::negative_binomial : 19 passed.

Fixes #463

Summary by CodeRabbit

• Bug Fixes
  • Corrected negative binomial probability calculations when the probability parameter equals 1.
  • Zero-count results now return the correct finite value, while nonzero counts correctly indicate zero probability.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: adbbe123-4304-4c2e-a54a-f62516bedc4f

📥 Commits

Reviewing files that changed from the base of the PR and between 52248ee and 4b20665.

📒 Files selected for processing (1)

• src/distribution/negative_binomial.rs

Included review availability: Your plan provides up to 4 included reviews per hour; 3 remain after this review.

---

📝 Walkthrough

Walkthrough

The change handles p == 1 directly in ln_pmf. It returns point-mass values for zero and positive inputs. PMF and log-PMF tests cover multiple r values.

Changes

Negative binomial boundary handling

Layer / File(s)	Summary
Handle p = 1 in PMF calculations src/distribution/negative_binomial.rs	ln_pmf returns 0 for x == 0 and negative infinity for positive x when p == 1. PMF and log-PMF tests validate these values across additional r values.

Priority: ➖ Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Severity of issue fixed: Medium

Merge Risk: ⚪ Minimal · up to 4b206

This change makes the negative-binomial distribution return valid point-mass values at p=1 instead of NaN, with boundary behavior covered for PMF and log-PMF. It is ready to merge.

Suggested reviewers: day01

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: handling p = 1 in NegativeBinomial::ln_pmf, consistent with Binomial behavior.
Linked Issues check	✅ Passed	The changes address issue #463. NegativeBinomial::ln_pmf now returns 0 for k = 0 and negative infinity for other k when p = 1, which makes pmf(0) equal 1. Tests cover integer and non-integer r values.
Out of Scope Changes check	✅ Passed	The changes are limited to the p = 1 boundary case in NegativeBinomial::ln_pmf and related tests. No unrelated changes are present.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 5 functions across 1 files.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}