Struct LabeledAnomaly 

pub struct LabeledAnomaly {

    pub anomaly_id: String,
    pub anomaly_type: AnomalyType,
    pub document_id: String,
    pub document_type: String,
    pub company_code: String,
    pub anomaly_date: NaiveDate,
    pub detection_timestamp: NaiveDateTime,
    pub confidence: f64,
    pub severity: u8,
    pub description: String,
    pub related_entities: Vec<String>,
    pub monetary_impact: Option<Decimal>,
    pub metadata: HashMap<String, String>,
    pub is_injected: bool,
    pub injection_strategy: Option<String>,
    pub cluster_id: Option<String>,
    pub original_document_hash: Option<String>,
    pub causal_reason: Option<AnomalyCausalReason>,
    pub structured_strategy: Option<InjectionStrategy>,
    pub parent_anomaly_id: Option<String>,
    pub child_anomaly_ids: Vec<String>,
    pub scenario_id: Option<String>,
    pub run_id: Option<String>,
    pub generation_seed: Option<u64>,
}

A labeled anomaly for supervised learning.

Fields

anomaly_id: String

Unique anomaly identifier.

anomaly_type: AnomalyType

Type of anomaly.

document_id: String

Document or entity that contains the anomaly.

document_type: String

Document type (JE, PO, Invoice, etc.).

company_code: String

Company code.

anomaly_date: NaiveDate

Date the anomaly occurred.

detection_timestamp: NaiveDateTime

Timestamp when detected/injected.

confidence: f64

Confidence score (0.0 - 1.0) for injected anomalies.

severity: u8

Severity (1-5).

description: String

Description of the anomaly.

related_entities: Vec<String>

Related entities (user IDs, account codes, etc.).

monetary_impact: Option<Decimal>

Monetary impact if applicable.

metadata: HashMap<String, String>

Additional metadata.

is_injected: bool

Whether this was injected (true) or naturally occurring (false).

injection_strategy: Option<String>

Injection strategy used (if injected) - legacy string field.

cluster_id: Option<String>

Cluster ID if part of an anomaly cluster.

original_document_hash: Option<String>

Hash of the original document before modification. Enables tracking what the document looked like pre-injection.

causal_reason: Option<AnomalyCausalReason>

Causal reason explaining why this anomaly was injected. Provides “why” tracking for each anomaly.

structured_strategy: Option<InjectionStrategy>

Structured injection strategy with parameters. More detailed than the legacy string-based injection_strategy field.

parent_anomaly_id: Option<String>

Parent anomaly ID if this was derived from another anomaly. Enables anomaly transformation chains.

child_anomaly_ids: Vec<String>

Child anomaly IDs that were derived from this anomaly.

scenario_id: Option<String>

Scenario ID if this anomaly is part of a multi-step scenario.

run_id: Option<String>

Generation run ID that produced this anomaly. Enables tracing anomalies back to their generation run.

generation_seed: Option<u64>

Seed used for RNG during generation. Enables reproducibility.

Implementations

impl LabeledAnomaly

pub fn new( anomaly_id: String, anomaly_type: AnomalyType, document_id: String, document_type: String, company_code: String, anomaly_date: NaiveDate, ) -> Self

Creates a new labeled anomaly.

pub fn with_description(self, description: &str) -> Self

Sets the description.

pub fn with_monetary_impact(self, impact: Decimal) -> Self

Sets the monetary impact.

pub fn with_related_entity(self, entity: &str) -> Self

Adds a related entity.

pub fn with_metadata(self, key: &str, value: &str) -> Self

Adds metadata.

pub fn with_injection_strategy(self, strategy: &str) -> Self

Sets the injection strategy (legacy string).

pub fn with_cluster(self, cluster_id: &str) -> Self

Sets the cluster ID.

pub fn with_original_document_hash(self, hash: &str) -> Self

Sets the original document hash for provenance tracking.

pub fn with_causal_reason(self, reason: AnomalyCausalReason) -> Self

Sets the causal reason for this anomaly.

pub fn with_structured_strategy(self, strategy: InjectionStrategy) -> Self

Sets the structured injection strategy.

pub fn with_parent_anomaly(self, parent_id: &str) -> Self

Sets the parent anomaly ID (for anomaly derivation chains).

pub fn with_child_anomaly(self, child_id: &str) -> Self

Adds a child anomaly ID.

pub fn with_scenario(self, scenario_id: &str) -> Self

Sets the scenario ID for multi-step scenario tracking.

pub fn with_run_id(self, run_id: &str) -> Self

Sets the generation run ID.

pub fn with_generation_seed(self, seed: u64) -> Self

Sets the generation seed for reproducibility.

pub fn with_provenance( self, run_id: Option<&str>, seed: Option<u64>, causal_reason: Option<AnomalyCausalReason>, ) -> Self

Sets multiple provenance fields at once for convenience.

pub fn to_features(&self) -> Vec<f64>

Converts to a feature vector for ML.

Returns a vector of 15 features:

• 6 features: Category one-hot encoding (Fraud, Error, ProcessIssue, Statistical, Relational, Custom)
• 1 feature: Severity (normalized 0-1)
• 1 feature: Confidence
• 1 feature: Has monetary impact (0/1)
• 1 feature: Monetary impact (log-scaled)
• 1 feature: Is intentional (0/1)
• 1 feature: Number of related entities
• 1 feature: Is part of cluster (0/1)
• 1 feature: Is part of scenario (0/1)
• 1 feature: Has parent anomaly (0/1) - indicates derivation

pub fn feature_count() -> usize

Returns the number of features in the feature vector.

pub fn feature_names() -> Vec<&'static str>

Returns feature names for documentation/ML metadata.

Trait Implementations

impl Clone for LabeledAnomaly

fn clone(&self) -> LabeledAnomaly

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LabeledAnomaly

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for LabeledAnomaly

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for LabeledAnomaly

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

where __S: Serializer,

Serialize this value into the given Serde serializer.