Skip to content

CLAUDE.md

Latest commit

聽

History

History
115 lines (86 loc) 路 3.09 KB

CLAUDE.md

File metadata and controls

115 lines (86 loc) 路 3.09 KB
Raw

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

This is Rusty Celery, a Rust implementation of the Celery distributed task queue system. The project provides both producer and consumer capabilities, supporting AMQP and Redis brokers. Key components include:

  • Core App (src/app/mod.rs): The main Celery struct and CeleryBuilder for creating apps
  • Brokers (src/broker/): AMQP and Redis broker implementations
  • Tasks (src/task/): Task definitions, signatures, async results, and execution
  • Protocol (src/protocol/): Message serialization and Celery protocol implementation
  • Beat (src/beat/): Periodic task scheduling (cron-like functionality)
  • Codegen: Procedural macros for the #[celery::task] attribute

Development Commands

Build and Test

# Build the project
cargo build
# or
make build

# Run tests (excludes broker integration tests)
make test

# Run all tests including broker integration tests
make run-all-tests

# Run specific broker tests
make broker-tests

Code Quality

# Format code
make format

# Lint (format check + clippy)
make lint

# Individual checks
make check-fmt
make check-clippy

Documentation

# Build documentation
make build-docs

Examples

# Run Rust celery app as consumer
cargo run --example celery_app consume

# Run Rust celery app as producer
cargo run --example celery_app produce [task_name]

# Available tasks: add, buggy_task, long_running_task, bound_task

# Run Beat scheduler
cargo run --example beat_app

Architecture Notes

Task Definition

Tasks are defined using the #[celery::task] procedural macro which generates a Task struct and implementation. The macro supports options like name, max_retries, time_limit, bind, etc.

App Creation

Apps are created using the celery::app! macro which configures brokers, registers tasks, and sets up routing rules.

Broker Architecture

  • AMQP Broker: Uses lapin crate for RabbitMQ/AMQP connections
  • Redis Broker: Uses redis crate with connection manager
  • Brokers implement the Broker trait for sending/receiving messages

Message Protocol

  • Supports multiple content types: JSON (default), MessagePack, YAML, Pickle
  • Compatible with Python Celery message format
  • Handles task routing, delivery, and acknowledgments

Concurrency Model

Built on Tokio async runtime with:

  • Async task execution
  • Stream-based message consumption
  • Configurable prefetch counts for concurrency control

Python Interoperability

The project maintains compatibility with Python Celery:

  • Same message format and protocol
  • Can consume tasks produced by Python Celery
  • Can produce tasks for Python Celery workers
  • Example Python app in examples/celery_app.py

Environment Setup

Requires AMQP_ADDR environment variable for broker connection:

export AMQP_ADDR="amqp://localhost:5672//"

For Redis:

export REDIS_ADDR="redis://localhost:6379/"

Use provided broker setup script:

./scripts/brokers/amqp.sh