Phase 3 test coverage: 3 proptest properties for DiscreteDistribution#132
Merged
Conversation
DiscreteDistribution::sample is hot-path code shared by every model (quality scores, fragment lengths, indel lengths, nucleotide selection). Blanketing its CDF binary search with randomized inputs is cheap insurance against off-by-one bugs that the 7 hand-rolled fixtures here can't surface. Adds `proptest = "1"` as a common dev-dep and three properties on DiscreteDistribution: - `proptest_sample_returns_value_from_input_set`: for any positive weight vector of length 2..=10 and any rand ∈ [0, 1], the sample output is in the input value set. Catches any future bisect bug that returns an out-of-range index. proptest's shrink space naturally exercises the rand=0.0 and rand=1.0 boundaries. - `proptest_empirical_frequency_approaches_weights`: 20,000 samples drawn through a deterministic NeatRng must produce a histogram within 3% of N per bin (well above the worst-case binomial standard deviation, comfortably below the threshold a systematic CDF bug would induce). Catches any off-by-one in cumulative_sum or the bisect predicate that would show up as a measurable skew. - `proptest_all_zero_weights_always_returns_first_value`: documents the actual (non-uniform!) behavior when all weights are zero. The `new()` constructor sets the CDF to `[1.0; n]`, which makes bisect-left return index 0 on every sample. Note that this is *not* a uniform fallback — the uniform fallback for quality models lives one layer up in QualityScoreModel::from_counts. Renaming this contract would now be a deliberate change, not a silent regression. cargo test --workspace: 200 common tests (was 197), 158 rneat tests, all passing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
DiscreteDistribution::sample is hot-path code shared by every model (quality scores, fragment lengths, indel lengths, nucleotide selection). Blanketing its CDF binary search with randomized inputs is cheap insurance against off-by-one bugs that the 7 hand-rolled fixtures here can't surface.
Adds
proptest = "1"as a common dev-dep and three properties on DiscreteDistribution:proptest_sample_returns_value_from_input_set: for any positive weight vector of length 2..=10 and any rand ∈ [0, 1], the sample output is in the input value set. Catches any future bisect bug that returns an out-of-range index. proptest's shrink space naturally exercises the rand=0.0 and rand=1.0 boundaries.proptest_empirical_frequency_approaches_weights: 20,000 samples drawn through a deterministic NeatRng must produce a histogram within 3% of N per bin (well above the worst-case binomial standard deviation, comfortably below the threshold a systematic CDF bug would induce). Catches any off-by-one in cumulative_sum or the bisect predicate that would show up as a measurable skew.proptest_all_zero_weights_always_returns_first_value: documents the actual (non-uniform!) behavior when all weights are zero. Thenew()constructor sets the CDF to[1.0; n], which makes bisect-left return index 0 on every sample. Note that this is not a uniform fallback — the uniform fallback for quality models lives one layer up in QualityScoreModel::from_counts. Renaming this contract would now be a deliberate change, not a silent regression.cargo test --workspace: 200 common tests (was 197), 158 rneat tests, all passing.
Closes #104