Struct ReleaseBumpStats 

pub struct ReleaseBumpStats {
    pub merged: usize,
    pub recorded: usize,
    pub pr_opened: usize,
    pub automerge_enabled: usize,
    pub merged_directly: usize,
    pub needs_attention: usize,
}

What the post-merge release-bump step did, across every merged run.

merged is the denominator for Self::coverage_rate: a repository that never uses the release-bump step (no auto-tag.yml, see kata:agents:rust:*) should read as “0 of N merged runs recorded a bump”, not vanish from the report the way it would if recorded were the denominator instead — silently excluding those runs would hide the coverage gap itself.

Fields

merged: usize

Merged runs, the denominator for Self::coverage_rate.

recorded: usize

Merged runs that carry a crate::run::ReleaseBump record at all.

pr_opened: usize

Recorded bumps that opened a release pull request.

automerge_enabled: usize

Recorded bumps whose automerge_enabled is true at the point the bump step finished. This is the final state, not “was automerge enabled at some point” — bump::surface_problem flips it back to false when GitHub later rejects automerge, so a bump that briefly enabled it and then had it rejected counts here as not enabled.

merged_directly: usize

Recorded bumps magi merged directly because GitHub refused automerge on an already-green pull request. Counted apart from automerge_enabled on purpose: a merged_directly bump needed no human even though automerge itself never took, so folding it into (or leaving it out of) the automerge count would misstate either number.

needs_attention: usize

Recorded bumps that ended with crate::run::RunState::needs_attention true — a human has something to do.

Implementations

impl ReleaseBumpStats

pub fn coverage_rate(&self) -> f64

Share of merged runs that recorded a release bump at all.

pub fn automerge_rate(&self) -> f64

Share of opened release PRs that ended with automerge enabled.

pub fn attention_rate(&self) -> f64

Share of recorded bumps that needed a human.

pub fn clean(&self) -> usize

Recorded bumps that finished with nothing for a human to do.

Deliberately the difference recorded - needs_attention, not a separate counter kept in step with automerge_enabled: a bump that GitHub refused automerge on but that magi merged directly (merged_directly) is clean — nobody had to act — even though automerge_enabled is false for it. Counting clean bumps as automerge_enabled alone would leave merged_directly cases in neither the clean nor the attention bucket, and the two would stop summing to recorded.

Trait Implementations

impl Clone for ReleaseBumpStats

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ReleaseBumpStats

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ReleaseBumpStats

fn default() -> Self

Returns the “default value” for a type.