Struct ClassificationResult 

pub struct ClassificationResult {
    pub category: String,
    pub subcategory: Option<String>,
    pub top_level: Option<TopLevelCategory>,
    pub confidence: f64,
    pub method: ClassificationMethod,
    pub ticket_id: Option<String>,
    pub complexity: Option<u8>,
}

Output of any tier: a category verdict plus provenance.

Why: every classifier tier needs to return the same shape so the cascade orchestrator can compare tiers’ verdicts and decide which to accept. What: bundles the verdict (category, subcategory, top_level), confidence, source tier (method), extracted ticket id, and an optional LLM-only complexity score. Test: covered by unclassified_defaults_complexity_to_none and every classifier test that asserts the verdict shape.

The hierarchy is:

• top_level — one of the canonical TopLevelCategory variants (resolved from category via the crate::classify::taxonomy::TaxonomyRegistry).
• category — the subcategory name (e.g. "feature", "security"). Kept as a free-form string for backward compatibility with the DB schema.
• subcategory — an even-more-specific leaf label (e.g. "sql-injection").

Fields

category: String

Subcategory name (e.g. "feature", "bugfix", "security").

Despite the field name, this is the subcategory in the two-level taxonomy — the registered TopLevelCategory parent is reported in top_level. The field name is preserved for DB-schema compatibility.

subcategory: Option<String>

Optional leaf label (e.g. "sql-injection", "cleanup").

top_level: Option<TopLevelCategory>

Resolved top-level category (None if category is unregistered).

confidence: f64

Confidence in this verdict (0.0–1.0).

method: ClassificationMethod

Which tier produced this verdict.

ticket_id: Option<String>

Optional extracted ticket id (e.g. "PROJ-123").

complexity: Option<u8>

Optional commit complexity score on a 1–5 scale.

None means the commit was not scored — only the LLM tier produces a complexity score; rule/regex/fuzzy tiers always leave this None. The scale is: 1 = trivial, 2 = simple, 3 = moderate, 4 = complex, 5 = highly complex.

Implementations

impl ClassificationResult

pub fn unclassified() -> Self

Construct an “unclassified” result used as a default when no tier matches.

Why: callers prefer a deterministic verdict over a panic / Option when the cascade fails; this is the canonical safe default. What: returns category = "uncategorized", top_level = Unknown, confidence = 0.0, method = FuzzyMatch, and complexity = None. Test: covered by unclassified_defaults_complexity_to_none.

complexity defaults to None — unclassified commits are never complexity-scored.

Trait Implementations

impl Clone for ClassificationResult

fn clone(&self) -> ClassificationResult

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ClassificationResult

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ClassificationResult

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ClassificationResult

fn eq(&self, other: &ClassificationResult) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for ClassificationResult

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ClassificationResult