[alok-108]

What was wrong?

Issue #1105 requested implementing get_extended_sample_count from the Ethereum Fulu specification draft (peer-sampling.md). The codebase lacked the hypergeometric cumulative distribution functions and sample-count calculation routines for peer sampling.

How was it fixed?

• Implemented math_comb (binomial coefficient combinations n choose k) matching Python's math.comb.
• Implemented hypergeom_cdf computing hypergeometric distribution CDF matching consensus-specs reference.
• Implemented get_extended_sample_count matching the Fulu peer sampling specification.
• Added comprehensive unit tests covering:
  • Exact spec test vectors across failure ranges 0..=64
  • Monotonicity checks
  • Edge cases and panic bounds checks

Note: As noted in discussions, upstream Ethereum consensus-specs PR #4393 removed peer sampling in favor of custody sampling.

Closes #1105

To-Do

• I have read CONTRIBUTING.md.
• This PR title follows conventional commits.

[perfogic]

Same for this one, peer-sampling is removed from consensus-specs since ethereum/consensus-specs#4393

[alok-108]

Thanks for pointing this out @perfogic!

I verified that upstream consensus-specs#4393 indeed dropped specs/fulu/peer-sampling.md from the Fulu scope in favor of custody sampling (as also tracked in #1478 and #1552).

The implementation here (math_comb, hypergeom_cdf, and get_extended_sample_count) was built against the reference in #1105 and passes all original spec test vectors and property tests.

Given that peer sampling is no longer part of the official Fulu consensus specification:

• If Ream is strictly adhering to post-#4393 Fulu specifications (custody sampling), we can close this PR and Issue Implement get_extended_sample_count from the Fulu Spec #1105 as superseded by upstream spec changes.
• Alternatively, if any of the underlying hypergeometric / combinatorial math utilities (math_comb, hypergeom_cdf) are useful to retain in a general math/stats helper module for custody scoring or analysis, I can gladly refactor and move them.

@shariqnaiyer @perfogic please let me know your preferred direction and I can close or adapt accordingly!

[KolbyML]

I don't think we should be using shorthands, can you make the code more readable and reduce the comments

[alok-108]

Thanks for the feedback @KolbyML! I've addressed this in commit 2f714f5:

• Eliminated shorthands: Renamed math_comb to combinations and hypergeom_cdf to hypergeometric_cdf. Replaced single-letter and abbreviated variable names with self-documenting ones (total_items, chosen_items, result, index, population_size, sample_size, success_states, denominator, numerator, cumulative_probability, etc.).
• Aligned parameter ordering: Reordered hypergeometric_cdf parameters to match the canonical Python consensus-specs order: (allowed_successes, success_states, population_size, sample_size).
• Reduced comment density: Lowered comment ratio, removing redundant inline comments in tests while preserving essential specification references and doc comments.
• All 6 unit tests, cargo fmt, and cargo clippy pass cleanly.