Curly

Curly is a minimal wrapper around libcurl, made to download data and images using pure C threads and libcurl, and release some pressure on Python GIL.

The goal is to prevent micro lag on the Kivy UI running in the main thread by preventing the GIL to be locked in another threads, or even prevent completely Python threads and so GIL switch between threads.

Features:

• Asynchronously download HTTP URL (no GIL)
• Basic caching support (no GIL)
• Preload image (GIL required for now)

Todo:

• make it work under msvc (http://forum.blackvoxel.com/index.php?topic=84.0)
• unittests
• Android support
• iOS support

Requirements

• libcurl
• SDL2
• SDL2_image

Installation

There is no release yet, so you must install it by sources.

Basic usage

You have 2 ways to execute a download:

• use request() to download a resource at a specific URL
• use download_image() to download a resource and preload the image, if the request succedded and the image format is known

The download will be done in background, and the threads will emit a result in a queue. You need to process as much as possible the queue to have your callback called within your application:

• use install() to install a Kivy clock scheduler that will process the result every tick (you can uninstall() it at any times)
• or call yourself process() when you need too.

Request an URL and get the data

from curly import curl
curl.install()

def on_complete(result):
    result.raise_for_status()
    print("Data: {!r}".format(result.data))

curl.request("https://kivy.org", on_complete)

Download an image

TIP: check for the AsyncCurlImage in the examples/image_browser.py to have an idea about how to use it within an Image widget.

from curly import curl
curl.install()

def on_complete(result):
    result.raise_for_status()
    print("Image is: {!r}", result.image)
    print("Texture is: {!r}", result.image.texture)

url = "https://dummyimage.com/600x400/000/fff"
curl.download_image(url, on_complete)