celery-salt 1.1.0

Event-driven architecture library for Python that extends Celery with event publishing/subscribing patterns

CelerySalt

A modern, event-driven architecture library for Python that extends Celery with schema-validated event publishing/subscribing patterns. Built with Pydantic for type safety and import-time schema registration.

🚀 v1.0.0 - First release! Pydantic-based event schemas, import-time registration, RPC support, and protocol compatibility with tchu-tchu.

Features

• ✨ Pydantic-Based Schemas - Type-safe event definitions with automatic validation
• 🚀 Import-Time Registration - Schemas registered at import time for early error detection
• 📡 Broadcast Events - Fire-and-forget pub/sub messaging (one event → many subscribers)
• ⚡ RPC Support - Synchronous request/response with response/error schema validation
• 🔄 Protocol Compatible - Works with existing tchu-tchu applications
• 🎯 Framework Agnostic - Core library works with any Python app (Django optional)
• 🛡️ Schema Registry - Centralized schema management (in-memory, PostgreSQL, Cloud)
• 🔁 Celery Integration - Full Celery features: retries, time limits, rate limiting, monitoring
• 📦 Simple API - Decorator-based: @event and @subscribe

Quick Start

Installation

pip install celery-salt

Broadcast Example

from celerysalt import event, subscribe

# Define event schema
@event("user.signup.completed")
class UserSignupCompleted:
    user_id: int
    email: str
    company_id: int
    signup_source: str = "web"

# Publish event
UserSignupCompleted.publish(
    user_id=123,
    email="[e-mail]",
    company_id=1,
    signup_source="web"
)

# Subscribe to event
@subscribe("user.signup.completed")
def send_welcome_email(data: UserSignupCompleted):
    print(f"Sending welcome email to {data.email}")

RPC Example

from celerysalt import event, subscribe, RPCError

# Define RPC request/response schemas
@event("rpc.calculator.add", mode="rpc")
class CalculatorAddRequest:
    a: float
    b: float

@event.response("rpc.calculator.add")
class CalculatorAddResponse:
    result: float
    operation: str = "add"

@event.error("rpc.calculator.add")
class CalculatorAddError:
    error_code: str
    error_message: str

# Handler
@subscribe("rpc.calculator.add")
def handle_add(data: CalculatorAddRequest) -> CalculatorAddResponse:
    return CalculatorAddResponse(result=data.a + data.b, operation="add")

# Client call
response = CalculatorAddRequest.call(a=10, b=5, timeout=10)
print(f"Result: {response.result}")  # 15.0

Architecture

Publisher → RabbitMQ Exchange (tchu_events) → Subscribers

• Exchange: tchu_events (topic exchange, protocol compatible)
• Routing: Topic-based with wildcard support (user.*, #)
• Serialization: JSON with Pydantic validation
• Result Backend: Redis (required for RPC)

Documentation

• Examples - Working examples for broadcast and RPC
• Design Document - Full architecture and design
• Implementation Context - Development context

Requirements

• Python 3.10+
• Celery 5.3+
• RabbitMQ (message broker)
• Redis (optional, required for RPC)

Installation

# Basic installation
pip install celery-salt

# With Django support
pip install celery-salt[django]

# With all extras
pip install celery-salt[all]

Examples

See the examples directory for complete working examples:

• Basic Broadcast - Pub/sub messaging
• Basic RPC - Request/response pattern

Run examples with Docker Compose:

cd examples
docker-compose up -d  # Starts RabbitMQ and Redis

Key Concepts

Event Schemas

Schemas are defined using Pydantic models and registered at import time:

@event("user.created")
class UserCreated:
    user_id: int
    email: str
    created_at: datetime

Publishing Events

# Broadcast (fire-and-forget)
UserCreated.publish(user_id=123, email="[e-mail]", created_at=datetime.now())

# RPC (synchronous)
response = CalculatorAddRequest.call(a=10, b=5, timeout=10)

Subscribing to Events

@subscribe("user.created")
def handle_user_created(data: UserCreated):
    # Process event
    pass

RPC Response Validation

@event.response("rpc.calculator.add")
class CalculatorAddResponse:
    result: float

@event.error("rpc.calculator.add")
class CalculatorAddError:
    error_code: str
    error_message: str

Protocol Compatibility

CelerySalt maintains protocol compatibility with tchu-tchu:

• Same exchange name: tchu_events
• Same message format: _tchu_meta field
• Same routing key conventions

This allows gradual migration: apps using celery-salt can communicate with apps still using tchu-tchu.

Development

# Clone repository
git clone https://github.com/sigularusrex/celery-salt.git
cd celery-salt

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Run examples
cd examples
docker-compose up -d
celery -A basic_broadcast.subscriber worker --loglevel=info

License

MIT License - see LICENSE file for details.

Contributing

Contributions welcome! Please open an issue or pull request on GitHub.

Links

• GitHub: https://github.com/sigularusrex/celery-salt
• Documentation: https://github.com/sigularusrex/celery-salt
• Issues: https://github.com/sigularusrex/celery-salt/issues