slices-lexicon

Rust implementation of AT Protocol lexicon validation.

Overview

This validation engine can be used in any project that needs AT Protocol lexicon validation. It provides high-performance, spec-compliant validation of AT Protocol lexicon documents and data records. It can also be compiled to WebAssembly for use in JavaScript/TypeScript environments.

Architecture

This package serves as the core validation engine and is typically consumed by higher-level packages:

• @slices/lexicon - TypeScript/Deno package with ergonomic APIs
• @slices/cli - Deno command-line tool for lexicon/appview management
• lexicon-intellisense - VS Code extension for lexicon development

Features

• High-performance lexicon validation implemented in Rust
• Comprehensive error reporting with detailed context
• Full AT Protocol lexicon specification compliance
• Support for all lexicon types: records, queries, procedures, subscriptions
• Optional WebAssembly compilation for JavaScript/TypeScript environments
• Zero JavaScript runtime dependencies when using WebAssembly

Direct Usage

Rust Library

Add to your Cargo.toml:

[dependencies]
slices-lexicon = "0.2"

Basic validation:

use slices_lexicon::{validate, validate_record, is_valid_nsid};
use serde_json::json;

// Define lexicon documents
let lexicons = vec![
    json!({
        "id": "com.example.post",
        "lexicon": 1,
        "defs": {
            "main": {
                "type": "record",
                "key": "tid",
                "record": {
                    "type": "object",
                    "required": ["text"],
                    "properties": {
                        "text": { "type": "string", "maxLength": 300 }
                    }
                }
            }
        }
    })
];

// Validate lexicon documents
match validate(lexicons.clone()) {
    Ok(()) => println!("All lexicons are valid"),
    Err(errors) => {
        for (lexicon_id, error_list) in errors {
            println!("Errors in {}: {:?}", lexicon_id, error_list);
        }
    }
}

// Validate a data record against the schema
let record = json!({ "text": "Hello, world!" });
validate_record(lexicons, "com.example.post", record)?;

// Validate NSID format
let is_valid = is_valid_nsid("com.example.post");
println!("NSID valid: {}", is_valid);

WebAssembly

Build the WASM module:

wasm-pack build --target web --features wasm

Use in JavaScript environments:

import init, { WasmLexiconValidator } from "./pkg/slices_lexicon.js";

await init();

// Validate lexicons
const lexicons = [{
  id: "com.example.post",
  lexicon: 1,
  defs: {
    main: {
      type: "record",
      key: "tid",
      record: {
        type: "object",
        required: ["text"],
        properties: {
          text: { type: "string", maxLength: 300 },
        },
      },
    },
  },
}];

const validator = new WasmLexiconValidator(JSON.stringify(lexicons));
const errorsJson = validator.validate_lexicons();
const errors = JSON.parse(errorsJson);

if (Object.keys(errors).length > 0) {
  console.log("Validation errors:", errors);
} else {
  console.log("All lexicons valid");
}

validator.free(); // Clean up WASM resources

JavaScript/TypeScript Usage

If you're using JavaScript or TypeScript, use the higher-level packages instead of consuming this library directly:

• TypeScript/JavaScript: Use @slices/lexicon for ergonomic APIs with automatic resource management
• VS Code Development: Install the lexicon-intellisense extension
• CLI Tools: Use the Slices CLI for lexicon management tasks

Supported Types

• Objects with required/optional fields
• Arrays with min/max length constraints
• Strings with format validation (datetime, URI, DID, handle, etc.)
• Integers with range constraints
• Booleans with const values
• Bytes (base64 encoded)
• Unions (open and closed)
• References (local and cross-lexicon)
• CID links
• Tokens (unit types for type discrimination)
• Unknown types

String Formats

• datetime - RFC3339/ISO8601 datetime
• uri - Generic URI
• at-uri - AT Protocol URI
• did - Decentralized Identifier
• handle - AT Protocol handle
• at-identifier - DID or handle
• nsid - Name Spaced Identifier
• cid - Content Identifier
• language - BCP47 language tag
• tid - Timestamp-based Identifier
• record-key - Record key format

License

MIT