Crate pii 

PII detection and anonymization interfaces.

This crate provides a deterministic pipeline for detecting and redacting personally identifiable information (PII). It is designed for CPU-only execution, explicit auditability, and controlled degradation when certain language features (lemma, POS, NER) are unavailable.

Typical usage:

• Build an Analyzer with an NlpEngine, recognizers, and policy.
• Run analyze to get detections with stable byte offsets.
• Pass detections to the Anonymizer with an operator configuration.

The pipeline is modular: you can swap in a custom NlpEngine, add custom recognizers, or augment the system with Candle-based NER when needed.

Specification: docs/rfc-1200-pii.md.

Re-exports

pub use analyzer::Analyzer;

pub use anonymize::AnonymizeConfig;

pub use anonymize::Anonymizer;

pub use anonymize::Operator;

pub use capabilities::Capabilities;

pub use config::PolicyConfig;

pub use error::PiiError;

pub use error::PiiResult;

pub use presets::default_recognizers;

pub use profile::ContextTerms;

pub use profile::LanguageProfile;

pub use profile::LanguageRegistry;

pub use types::AnalyzeResult;

pub use types::AnonymizeResult;

pub use types::AnonymizedItem;

pub use types::Detection;

pub use types::DetectionExplanation;

pub use types::EntityType;

pub use types::Language;

pub use types::NerSpan;

pub use types::NlpArtifacts;

pub use types::Token;

pub use types::LANGUAGE_DE;

pub use types::LANGUAGE_EN;

pub use types::LANGUAGE_ES;

Modules

analyzer

Analyzer pipeline wiring for PII detection.

anonymize

Anonymization operators and helpers.

capabilities

Capability flags produced by an NLP engine.

config

Policy configuration for entity filtering and thresholds.

context

Context-aware score boosting for ambiguous detections.

decision

Candidate resolution and deterministic overlap handling.

error

Error types for the PII library.

nlp

NLP engine traits and a simple reference implementation.

presets

Built-in recognizer presets for common PII types.

profile

Language profiles and context term configuration.

recognizers

Recognizer traits and implementations.

types

Core data types for PII detection and anonymization.

Constants

SPEC_PATH

Relative path to the RFC specification document.