Corpus Library

The corpus library is a shared Rust crate for loading and managing the regulation corpus. It abstracts over multiple source types (local filesystem and GitHub repositories) and handles YAML parsing, registry management, and schema validation.

Overview

Language: Rust
Location: packages/corpus/
Type: Library crate (used by other packages)

What it does

The corpus library provides a single API for loading law files regardless of where they are stored. It reads the corpus-registry.yaml manifest, authenticates with remote sources, fetches YAML files, and parses them into typed Rust structures.

Other packages use it:

The editor-api uses it to serve law files to the frontend
The admin uses it to proxy corpus data to the dashboard
The engine uses the parsed output for execution

Key modules

Module	Purpose
`registry.rs`	`CorpusRegistry` - loads `corpus-registry.yaml`, merges local overrides
`source_map.rs`	`SourceMap` - maps law IDs to parsed regulation YAML
`models.rs`	`Source`, `SourceType` (Local/GitHub), `RegistryManifest`
`github.rs`	`GitHubFetcher` - fetches YAML via GitHub API (feature-gated)
`validation.rs`	Schema validation against the JSON schema
`auth.rs`	Token management for private repositories
`client.rs`	`CorpusClient` - higher-level read/write access used by editor-api
`backend.rs`	The `RepoBackend` trait and the factory that picks an implementation per source
`github_api_backend.rs`	API-only backend: no local clone, reads through the Contents API, writes buffered in memory and flushed as one PUT or DELETE per file
`implements_index.rs`	Reads the committed implements-index, the precomputed artefact that replaces a full-corpus scan for IoC resolution
`timing.rs`	Request-scoped phase timings, surfaced as a `Server-Timing` header and as tracing spans
`annotation_schema.rs`, `dto.rs`, `config.rs`, `error.rs`	Annotation schema, transfer types, deployment config, and error types

Usage

use regelrecht_corpus::CorpusRegistry; use std::path::Path; // load(manifest_path, local_override_path): the second argument merges a // corpus-registry.local.yaml override when present. let registry = CorpusRegistry::load(Path::new("corpus-registry.yaml"), None)?; // load_all_sources_async includes GitHub sources. Pass the auth file when // private repositories need a token, or None when all sources are public. let source_map = registry .load_all_sources_async(Some(Path::new("corpus-auth.yaml"))) .await?; let law = source_map.get_law("wet_op_de_zorgtoeslag");

The github feature flag enables remote fetching from GitHub repositories. Without it, only local filesystem sources are available.

Overview

What it does

Key modules

Usage

Further reading

RegelRecht

Links

Contact

Part of