Ontology-Registry

A robust, thread-safe Rust library for managing the lifecycle of biological ontologies.

ontology-registry automates the process of resolving, downloading, and caching ontology files (JSON, OBO, OWL). It acts as a centralized local registry, ensuring that your applications always have access to the data they need without redundant network requests or race conditions.

---

✨ Features

• 🚀 Smart Caching: Automatically downloads ontologies from the OBO Foundry and persists them locally. Subsequent requests load instantly from the disk.
• 🔄 Version Resolution: resolving Version::Latest automatically queries BioRegistry.io to find the most recent semantic version.
• 🛡️ Thread-Safe & Atomic: Built for concurrency. It uses Mutex locks and atomic file writes (downloading to .tmp first) to ensure you never read a corrupted or partially downloaded file.
• 🔌 Modular Architecture: The logic is split into MetadataProviding, OntologyProviding, and Registration traits, allowing you to swap out backends if needed.
• 📂 Multiple Formats: First-class support for .json, .obo, and .owl formats.

📦 Installation

Add this to your Cargo.toml:

cargo add ontology-registry

Quick Start

use ontology_registry::blocking::bio_registry_metadata_provider::BioRegistryMetadataProvider;
use ontology_registry::blocking::file_system_ontology_registry::FileSystemOntologyRegistry;
use ontology_registry::blocking::obolib_ontology_provider::OboLibraryProvider;
use ontology_registry::enums::{FileType, Version, SupportedOntology};
use ontology_registry::traits::OntologyRegistration;
use std::path::PathBuf;
use std::io::Read;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // 1. Setup the registry with standard providers
    // In a real app, use a persistent path like "~/.cache/ontologies"
    let cache_dir = PathBuf::from("./local_ontology_cache");

    let registry = FileSystemOntologyRegistry::new(
        cache_dir,
        BioRegistryMetadataProvider::default(), // Resolves versions via BioRegistry.io
        OboLibraryProvider::default(),          // Downloads content from OBO Library
    );

    // 2. Register (Download & Cache)
    // This resolves 'Latest' to a specific date (e.g., "2024-01-01")
    let mut reader = registry.register(
        SupportedOntology::MONDO, // This can also just be a string "mondo"
        Version::Latest,
        FileType::Obo
    )?;

    // 3. Read the content
    let mut content = String::new();
    reader.read_to_string(&mut content)?;

    println!("Successfully loaded Mondo Ontology ({} bytes)", content.len());
    Ok(())
}

Pinning a Specific Version

If you need reproducibility, you can request a specific version string.

use ontology_registry::enums::{FileType, Version};
use ontology_registry::blocking::file_system_ontology_registry::FileSystemOntologyRegistry;
use ontology_registry::traits::OntologyRegistration;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let registry = FileSystemOntologyRegistry::default();

    let version = Version::Declared("2023-01-01".to_string());

    let reader = registry.register(
        "go",
        version,
        FileType::Owl
    )?;
}

📂 Supported Formats

Enum Variant	Extension	Description
FileType::Json	.json	OBO Graph JSON format
FileType::Obo	.obo	Standard OBO flat file
FileType::Owl	.owl	Web Ontology Language format