traits-futures 0.2.0

Patterns for reactive background tasks

Traits Futures

The traits_futures package demonstrates robust design patterns for reactive background jobs triggered from a TraitsUI application.

Motivating example

A customer has asked you to wrap their black-box optimization code in a GUI.

You build a simple TraitsUI GUI that allows the user to set inputs and options and then press the big green “Calculate” button. The requirements look something like this:

• The UI should remain usable and responsive while the background calculation is running.

• The UI should update (e.g., show a plot, or show results) in response to the background calculation completing normally.

• The background calculation should be cancellable.

• The UI should react appropriately if the background calculation raises an exception.

And there are some further ease-of-development requirements:

• As far as possible, the UI developer shouldn’t have to think about managing the background threads or re-dispatching incoming information from the background task(s). The UI developer should be able to simply listen to and react to suitable traits for information coming in from the background task.

• It should be possible to switch between using background threads and background process (and possibly also coroutines) with minimal effort.

Getting started

The ci helper package in the source repository aids in setting up a development environment and running tests and examples. It requires EDM, along with a Python bootstrap environment equipped with click and setuptools.

To create a development environment, run:

python -m ci build

from the top-level of the repository, within the Python bootstrap environment.

To run tests for the traits_futures EDM environment, do:

python -m ci test

To run tests under coverage:

python -m ci coverage

To run a style check:

python -m ci flake8

To build the documentation:

python -m ci doc

The example scripts can be run with:

python -m ci example <example-name>

Run python -m ci example to see the list of available examples.

All of the above commands take two options. The --python-version option lets you specify the Python version to use for the development environment. The --toolkit option allows you to specify a GUI backend. Run python -m ci <command> --help for more information on any of these commands.