Expand description
Implementation crate for cargo-bench-history; do not depend on this directly.
The stored data model for a benchmark run: a run reduced to a set of benchmark
results (each a stable identity plus its measured metrics), the run context that
situates it in time and against a commit, and the comparability rules that
partition runs into independently comparable series. Split out of
cargo-bench-history so this I/O-free data model is cheap to
mutation-test in isolation.
Every type is re-exported flat from the crate root, so consumers write
cbh_model::DiscriminantSet rather than reaching into a submodule.