Jute Parser & Rust Code Generator

A Rust implementation of the Apache ZooKeeper Jute IDL, providing a compiler-grade solution for protocol schema parsing and code generation.

This crate is designed for compiler-grade correctness, not just simple code generation, ensuring your generated code is based on a fully validated schema.

✨ Features

• 📜 Parsing: Full Jute grammar support, including module, class, primitive types, vector<T>, map<K, V>, and cross-module type references.
• 🔗 Cross-File Resolution: Robust handling of relative file includes (include "./path.jute";).
• 🧠 Dependency Validation:
  • Dependency graph resolution.
  • Detects unknown types and ambiguous references.
  • Detects circular dependencies.
• 🦀 Rust Code Generation:
  • Idiomatic Rust output.
  • Read/Write based serialization.
  • No runtime reflection.
• 🔄 Serialization & Deserialization: Provides high-performance, strongly-typed serialization logic.
• 🧪 Strong Test Coverage: Comprehensive tests for the parser, resolver, and serialization round-trips.

🚀 Quick Start

1. Example Jute Schema

Create a schema file, e.g., schema/serialization.jute:

module serialization {
    class Person {
        int id;
        string name;
        boolean active;
    }

    class Company {
        int id;
        Person owner;
        vector<Person> employees;
    }
}

2. Generate Rust Code

Use the JuteGenerator to process your schema and generate source files:

use jute::JuteGenerator;

JuteGenerator::new()
    .add_src_file("schema/serialization.jute") // Input schema
    .add_out_dir("generated") // Output directory src (curr crate) will be used as root 
    .generate()
    .unwrap();

3. Using Generated Code

The generated code is not automatically wired into your crate. You must include it manually:

// src/lib.rs or src/main.rs
pub mod serialization {
    include!("generated/serialization/mod.rs");
}

4. 🔄 Serialization & Deserialization

All generated types implement two core methods for I/O-based serialization:

fn serialize<W: std::io::Write>(&self, out: &mut W) -> Result<(), JuteError>;
fn deserialize<R: std::io::Read>(input: &mut R) -> Result<Self, JuteError>;

Example Round-Trip:

use crate::serialization::{Person, Company};

let person = Person {
    id: 1,
    name: "Alice".into(),
    active: true,
};

let mut buffer = Vec::new();
// Serialize
person.serialize(&mut buffer).unwrap();

// Deserialize
let decoded = Person::deserialize(&mut &buffer[..]).unwrap();

assert_eq!(decoded.id, 1);
// ... more assertions

A complete end-to-end example demonstrating how to use this crate is available here:View the full E2E example

📂 Include Semantics

Includes are resolved relative to the file that contains the include directive:

include "./common.jute";
include "../shared/types.jute";

The resolver ensures all paths are:

• Canonicalized
• Deduplicated
• Cycle-checked

❌ Error Handling

All failures return a unified, descriptive error type:

use jute::errors::JuteError;

Errors are descriptive and contextual, including file paths and module names. They cover:

• Syntax errors in the schema
• Unknown or ambiguous types
• Circular dependencies
• IO failures
• Invalid schema definitions

🏗 Project Structure

The project follows a classic compiler architecture:

src/
├── compiler/
│   ├── ast_printer.rs // utility
│   ├── ast.rs              
│   ├── dependency_resolver.rs
│   ├── lexer.rs 
│   ├── parser.rs  
│   ├── state_machine.rs  
│   ├── token.rs   
│   ├── mod.rs     
├── code_generator/
│   └── rust/    
│   |   └── utilities.rs
│   |   └── writer.rs
│   |   └── mod.rs
│   └── mod.rs    
├── errors.rs
└── lib.rs

🎯 Design Goals

This crate prioritizes stability and correctness:

• Correctness > Convenience
• Explicit over implicit
• No hidden build steps
• Deterministic code generation

It is intended for:

• Zookepeer adapters
• Protocol tooling
• Schema-driven serialization

🧠 Inspiration

• Apache ZooKeeper Jute
• Protocol Buffers (IDL-style)

Missing Implementation

Following features are currently not supported

• Containerized recursive schema
• Can be enhanced with async versions of serialization.

📜 License

This project is licensed under the MIT License.

🤝 Contributing

Contributions are welcome! Focus areas include:

• Adding new Jute features
• Improving error diagnostics
• Extending code generators (e.g., for other languages)
• Improving test coverage
• Implementing missing features
• benchmarking