Introduction

We welcome all new developers and contributors with very friendly community. Join #FlexGet @ freenode.

Technologies used

List of external libraries and utilities. As for new libraries, we require that all of them are installable on Windows trough pip. This will exclude things that require compilers like LXML.

Core

  • SQLAlchemy
  • BeautifulSoup
  • Feedparser
  • Python-Requests
  • PyNBZ
  • Jinja2
  • PyYaml
  • jsonschema
  • Some smaller misc libraries

HTTPServer

  • Flask
  • Jinja2
  • CherryPy

CherryPy is only used for WSGI server.

How do I get started?

Set up development environment, which is basically just two steps.

  1. GIT clone our repository.
    • Either the wheel package of setuptools and pip need to be in the directory with bootstrap.py
    • Or virtualenv package needs to be installed.
  3. Run bootstrap.py with Python 2.6.x - 2.7.x.

For easier collaboration we recommend forking us on github and sending pull request. Once we see any semi-serious input from a developer we will grant write permissions to our central repository. You can also request this earlier if you wish.

If you are new to Git there are several interactive tutorials you can try to get you started including tryGit and LearnGitBranching.

Environment

Once you have bootstrapped the environment you have fully functional FlexGet in a virtual environment in your clone directory. You can easily add or modify existing plugins in here and it will not mess your other FlexGet instances in any way. The commands in the documentation expect the virtual environment to be activated. If you don’t activate it you must run commands explicitly from under environment bin directory or scripts in windows. E.g. flexget would be bin/flexget (at project root) in unactivated virtual environment.

How to activate virtualenv under linux:

source bin/activate

FlexGet project uses paver to provide development related utilities and tasks. Run paver --help to see what commands are available. Some of these will be mentioned later.

Code quality

Unit tests

There are currently over 250 unit tests ensuring that existing functionality is not accidentally broken. Unit tests can be invoked with the installation of additionnal requirments:

pip install jenkins-requirements.txt

Easiest way to run tests is trough paver:

paver test

By default no online tests are executed, these can be enabled with --online argument. There are other ways to run the tests as well, more specifically we use nose framework.

Run single test file via nose:

nosetests test_file

Run single test suite (class):

nosetests test_file:class

Run single test case from suite:

nosetests test_file:class.case

Live example:

nosetests test_seriesparser:TestSeriesParser.test_basic

Note

Don’t use .py extension or include path with these. Configuration file setup.cfg defines needed parameters for Nose.

Project has Jenkins CI server which polls master branch and makes runs tests and makes new build if they pass.

Unit tests are not mandatory for a plugin to be included in the FlexGet distribution but it makes maintaining our code trough project life and refactoring so much easier.

Code Style

All code should be formatted according to Python PEP8 recommendations. With the exception of line length limit at 79 characters. FlexGet uses 120 characters instead.

To run PEP8 checker:

paver pep8

We do have some violations in our codebase, but new code should not add any.