eventide 0.0.0a8

Fast and simple queue workers

Eventide

A fast, simple, and extensible queue worker framework for Python.

Overview

Eventide is a modern, lightweight framework for building robust queue-based worker systems in Python. It provides a clean, modular, and provider-agnostic architecture for consuming and processing messages from a variety of queue backends (SQS, Cloudflare, and more to come).

Key Features:

• Multiprocess architecture for high throughput and resilience
• Provider-agnostic queue abstraction with built-in and custom queue support
• Declarative, decorator-based message handler registration
• Robust retry/backoff logic
• Graceful startup and shutdown with signal handling
• Type-safe configuration using Pydantic models
• Extensible handler matching and routing

Architecture

Eventide orchestrates the lifecycle of your queue worker system:

• Main Process: Manages configuration, queue instantiation, worker process lifecycle, and graceful shutdown.
• Queue: Continuously pull messages from external queues (SQS, Cloudflare, etc.) into internal buffers.
• Worker Processes: Each worker consumes messages from the buffer and routes them to user-defined handlers that run within the Worker's process.
• Handlers: User functions decorated with @app.handler that process messages matching specific patterns.

All configuration is done via Pydantic models, ensuring type safety and validation.

Installation

pip install eventide

# With SQS support:
pip install eventide[sqs]

# With Cloudflare Queues support:
pip install eventide[cloudflare]

Quick Start

from eventide import Eventide, EventideConfig, Message, SQSQueueConfig

# Instantiate the eventide app
app = Eventide(
    config=EventideConfig(
        queue=SQSQueueConfig(
            region="us-east-1",
            url="https://sqs.us-east-1.amazonaws.com/123456789012/my-queue",
        ),
    ),
)

# Define a handler
@app.handler("body.type == 'greeting'")
def handle_greeting(message: Message) -> None:
    print(f"Received greeting: {message.body.get('content')}")

Configuration

All configuration is via Pydantic models:

• EventideConfig: Main config (queue, concurrency, handler paths, timeouts, retry policies, etc)
• SQSQueueConfig, CloudflareQueueConfig, ...: Provider-specific queue configs

Example:

from eventide import EventideConfig, SQSQueueConfig

config = EventideConfig(
    queue=SQSQueueConfig(
        region="us-east-1",
        url="https://sqs.us-east-1.amazonaws.com/123456789012/my-queue",
    ),
    concurrency=4,
    timeout=30.0,  # Handler timeout (seconds)
    retry_for=[ValueError],  # Retry for specific exceptions
    retry_limit=5,  # Max retries per message
    retry_min_backoff=1.0,  # Min backoff in seconds
    retry_max_backoff=60.0, # Max backoff in seconds
    handler_paths=["myapp/handlers"],  # Auto-discover handler modules
)

Message Handlers

Handlers are registered using the @app.handler decorator. You can match on message attributes, set retry/backoff policies, and more.

from eventide import Eventide, EventideConfig

app = Eventide(EventideConfig(...))

@app.handler("body.type == 'email'", retry_limit=3, retry_for=[ValueError])
def process_email(message):
    print(f"Processing email: {message.body}")

Advanced matching (multiple matchers, logical operators):

from eventide import Eventide, EventideConfig

app = Eventide(EventideConfig(...))

@app.handler(
    "body.type == 'notification'",
    "body.priority == 'high'",
    operator=all,  # or any
)
def process_notification(_message):
    ...

Advanced Usage

• Graceful Shutdown: Eventide handles SIGINT/SIGTERM for clean shutdown.
• Retries & Backoff: Handlers can specify retry policies and backoff intervals.
• Extensible Matching: Handler matcher logic can be customized for advanced routing.

SQS Queue

from eventide import SQSQueueConfig

sqs_config = SQSQueueConfig(
    region="us-east-1",
    url="https://sqs.us-east-1.amazonaws.com/123456789012/my-queue",
    visibility_timeout=30,  # seconds
    max_number_of_messages=10,  # max messages to fetch
    buffer_size=100,  # internal buffer size
)

Cloudflare Queue

from eventide import CloudflareQueueConfig

cf_config = CloudflareQueueConfig(
    account_id="my-account-id",
    queue_id="my-queue-id",
    buffer_size=100,  # internal buffer size
    batch_size=10,  # max messages to fetch
    visibility_timeout_ms=30000,  # milliseconds
)

Message Routing with JMESPath

Eventide uses JMESPath expressions to route messages to the appropriate handlers. This provides a powerful and flexible way to match messages based on their content.

What is JMESPath?

JMESPath is a query language for JSON that allows you to extract and transform elements from a JSON document. In Eventide, it's used to match messages to handlers based on their content.

Examples of JMESPath Expressions

# Match messages with a specific type
"body.type == 'email'"

# Match messages with a specific attribute value
"body.customer_id == '12345'"

# Match messages with a specific attribute in an array
"contains(body.tags, 'urgent')"

# Match messages with a numeric comparison
"body.priority > 5"

# Match messages with a specific structure
"body.user.verified == true"

# Complex condition with multiple operators
"body.type == 'order' && body.total > 100"

Combining Multiple Expressions

You can combine multiple JMESPath expressions with logical operators:

from eventide import Eventide, EventideConfig

app = Eventide(EventideConfig(...))

# Match messages that satisfy ALL conditions
@app.handler(
    "body.type == 'notification'",
    "body.priority == 'high'",
    operator=all
)
def priority_notifications_handler(_message):
  pass

# Match messages that satisfy ANY condition
@app.handler(
    "body.type == 'email'",
    "body.type == 'sms'",
    operator=any
)
def email_or_sms_handler(_message):
  pass

This approach gives you fine-grained control over which messages are routed to which handlers, allowing for clean separation of concerns in your application.

Practical Example: Order Processing System

Here's a complete example of using Eventide to build an order processing system:

# app.py
from eventide import Eventide, EventideConfig, SQSQueueConfig

app = Eventide(
    config=EventideConfig(
        queue=SQSQueueConfig(
            region="us-east-1",
            url="https://sqs.us-east-1.amazonaws.com/123456789012/orders-queue",
            # Increase visibility timeout for longer processing tasks
            visibility_timeout=120,
        ),
        # Use multiple workers for better throughput
        concurrency=4,
    ),
)

# Define handlers for different message types
@app.handler("body.type == 'new_order'")
def process_new_order(message):
    order = message.body.get('order', {})
    order_id = order.get('id')
    print(f"Processing new order: {order_id}")
    # Your order processing logic here
    return True

@app.handler("body.type == 'payment_confirmed'")
def process_payment(message):
    order_id = message.body.get('order_id')
    amount = message.body.get('amount')
    print(f"Payment of ${amount} confirmed for order: {order_id}")
    # Update order status, trigger shipping, etc.
    return True

@app.handler(
    "body.type == 'order_status_update'",
    "body.status == 'shipped'"
)
def handle_shipped_order(message):
    order_id = message.body.get('order_id')
    tracking_number = message.body.get('tracking_number')
    print(f"Order {order_id} shipped with tracking number: {tracking_number}")
    # Send confirmation email to customer, update database, etc.
    return True

To run this application:

# Install dependencies
pip install eventide[sqs]

# Run the application
eventide run -a app:app

This example demonstrates how to:

1. Define multiple handlers for different types of messages
2. Use JMESPath expressions to route messages to the appropriate handlers
3. Configure the application with the appropriate queue settings
4. Run multiple workers for better throughput

Roadmap

• Lifecycle hooks
• Comprehensive test suite
• Message scheduling (cron and one-off)

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.