swiftpy-gcd 0.0.2

Swift's Grand Central Dispatch (GCD) implemented in Python

swift.py

A Python implementation of Apple's Grand Central Dispatch (GCD) pattern, bringing efficient concurrent and parallel execution to Python applications (for no good reason).

Installation

# Using uv (recommended - faster)
uv pip install swiftpy-gcd

# Using pip
pip install swiftpy-gcd

Usage

from swift import (
    DispatchQueue,
    DispatchQueueAttributes,
    DispatchWorkItem
)

# Create a concurrent queue for I/O operations
io_queue = DispatchQueue(
    label="com.example.io",
    attributes=DispatchQueueAttributes.CONCURRENT
)

# Async execution
io_queue.async_(lambda: print("Async work"))

# With a work item
work = DispatchWorkItem(block=lambda: print("Work item"))
io_queue.async_(work)

# Delayed execution
io_queue.async_after(5.0, lambda: print("Delayed work"))

# CPU-bound parallel processing
cpu_queue = DispatchQueue(
    label="com.example.cpu",
    attributes=DispatchQueueAttributes.CONCURRENT_WITH_MULTIPROCESSING
)

# This will run in true parallel on multiple cores
cpu_queue.async_(lambda: process_large_dataset())

# Synchronous execution
result = cpu_queue.sync(lambda: compute_something())

# Serial queue for ordered operations
serial_queue = DispatchQueue(label="com.example.serial")
serial_queue.async_(lambda: print("First"))
serial_queue.async_(lambda: print("Second"))  # Guaranteed to run after First

Features

• Dispatch Queues: Both serial and concurrent execution
• Multiprocessing Support: True parallel execution for CPU-bound tasks
• Work Items: Cancellable tasks with completion handlers
• Delayed Execution: Schedule work to run after a delay
• Thread Safety: Built-in synchronization mechanisms
• Global Queues: Pre-configured queues for common scenarios
  • Main queue for UI operations
  • Global concurrent queue for I/O
  • Global CPU queue for parallel processing

Requirements

• Python 3.8+
• No external dependencies

Performance

• Efficient thread/process pool management
• Smart handling of CPU vs I/O bound tasks
• Low overhead task scheduling
• Automatic worker scaling based on system capabilities

License

MIT License