Struct HealthSummary 

pub struct HealthSummary {
    pub stale_entities: Vec<StaleEntity>,
    pub missing_fields: Vec<HealthReport>,
    pub orphan_count: usize,
    pub stub_count: usize,
    pub warnings: Vec<WarningHint>,
    pub dangling_links: Option<Vec<DanglingLink>>,
    pub findings: Option<Vec<IntegrityFinding>>,
    pub tag_distribution: Option<Vec<TagDistribution>>,
    pub tag_distribution_folded: Option<Vec<FoldedTag>>,
    pub untagged_entities: Option<UntaggedStats>,
}

Aggregated health report for the whole graph.

Fields

stale_entities: Vec<StaleEntity>

missing_fields: Vec<HealthReport>

orphan_count: usize

stub_count: usize

warnings: Vec<WarningHint>

Typed non-fatal issues visible to every caller of Engine::health(). Populated in two layers: Engine.load_warnings contributes drift warnings surfaced during mem load / reload / attach (SuspiciousNestedPrefix, future load-time checks); the MCP handler additionally appends request-scoped warnings (unknown include keys, clamped limit) on top of whatever the engine merged. Empty on the happy path.

dangling_links: Option<Vec<DanglingLink>>

Inline wiki-links in entity section bodies that resolve to stub targets (no on-disk markdown file). Populated only when the caller opts in via include=["dangling_links"]; None otherwise, so absence-of-key means “not requested” and presence-of-empty-array means “requested, zero findings”. Scan is handler-driven (same pattern as warnings above), so non-MCP callers of Engine::health() always see None unless they invoke health::collect_dangling_links directly.

findings: Option<Vec<IntegrityFinding>>

Integrity findings ({ id, axis, code, detail }) over the conformance axis — and, under include=["integrity"], the consistency axis too. Populated only when the caller opts in via include=["conformance"] / include=["integrity"]; None otherwise (same handler-driven pattern as dangling_links: absence means “not requested”, an empty array means “requested, fully integral”).

tag_distribution: Option<Vec<TagDistribution>>

Tag distribution (count per distinct tag, case-sensitive) over non-stub entities. Populated only when the caller opts in via include=["tags"]. Case-variant drift is surfaced via the sibling field [tag_distribution_folded].

tag_distribution_folded: Option<Vec<FoldedTag>>

Case-drift audit sidecar: entries where two or more casings of the same canonical tag (lowercase) both appear in authored tags. Only entries with variants.len() > 1 are returned — the default read of tag_distribution stays untouched. Populated alongside tag_distribution.

untagged_entities: Option<UntaggedStats>

Count of non-stub entities whose tags metadata is missing, empty, or resolves to zero effective tags after splitting on , and trimming. Populated alongside tag_distribution when include=["tags"].

Trait Implementations

impl Clone for HealthSummary

fn clone(&self) -> HealthSummary

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HealthSummary

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

Formats the value using the given formatter.

impl Serialize for HealthSummary

fn serialize<__S>( &self, __serializer: __S, ) -> Result<<__S as Serializer>::Ok, <__S as Serializer>::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.