tokmd-analysis-types

Analysis receipt contracts for tokmd.

Overview

This is a Tier 0 crate defining pure data structures for analysis receipts. It contains no I/O or business logic - only type definitions and serialization.

Installation

[dependencies]
tokmd-analysis-types = "1.3"

Key Types

Core Receipt

pub struct AnalysisReceipt {
    pub schema_version: u32,
    pub archetype: Option<Archetype>,
    pub topics: Option<TopicClouds>,
    pub entropy: Option<EntropyReport>,
    pub derived: Option<DerivedReport>,
    pub git: Option<GitReport>,
    // ... and more optional sections
}

Analysis Result Types

Type	Purpose
`Archetype`	Project kind detection (CLI, library, web app)
`TopicClouds`	Semantic topic extraction with TF-IDF scores
`EntropyReport`	High-entropy file detection
`PredictiveChurnReport`	Git-based change trend prediction
`CorporateFingerprint`	Author domain statistics
`LicenseReport`	SPDX license detection
`DerivedReport`	Core metrics (density, distribution, COCOMO)
`AssetReport`	Non-code file categorization
`GitReport`	Hotspots, bus factor, freshness, coupling
`ImportReport`	Module dependency graph
`DuplicateReport`	Content duplication detection
`FunReport`	Eco-label and novelty outputs

Schema Version

pub const ANALYSIS_SCHEMA_VERSION: u32 = 8;

v4 added cognitive complexity, nesting depth, and function-level details. v5 added Halstead metrics, maintainability index, complexity histogram, technical debt ratio, duplication density, and code age distribution. v6 added API surface enricher. v7 added coupling normalization (Jaccard/Lift), commit intent classification, and near-duplicate detection.

Design Principles

All analysis sections are Option<T> to support preset-based inclusion
Uses BTreeMap for deterministic key ordering
No I/O operations - pure data definitions

License

MIT OR Apache-2.0

tokmd-analysis-types 1.7.2