# If you have a fork, you'll want to clone it instead
git clone git://github.com/lemur/lemur.git
# Create a python virtualenv
mkvirtualenv lemur
# Make the magic happen
make
Running ``make`` will do several things, including:
* Setting up any submodules (including Bootstrap)
* Installing Python requirements
* Installing NPM requirements
..note::
You will want to store your virtualenv out of the ``lemur`` directory you cloned above,
otherwise ``make`` will fail.
Create a default Lemur configuration just as if this were a production instance:
::
lemur init
You'll likely want to make some changes to the default configuration (we recommend developing against Postgres, for example). Once done, migrate your database using the following command:
Lemur follows the guidelines laid out in `pep8 <http://www.python.org/dev/peps/pep-0008/>`_ with a little bit
of flexibility on things like line length. We always give way for the `Zen of Python <http://www.python.org/dev/peps/pep-0020/>`_. We also use strict mode for JavaScript, enforced by jshint.
You can run all linters with ``make lint``, or respectively ``lint-python`` or ``lint-js``.
To help developers maintain the above standards, Lemur includes a configuration file for Yelp's `pre-commit <http://pre-commit.com/>`_. This is an optional dependency and is not required in order to contribute to Lemur.
The test suite consists of multiple parts, testing both the Python and JavaScript components in Lemur. If you've setup your environment correctly, you can run the entire suite with the following command:
::
make test
If you only need to run the Python tests, you can do so with ``make test-python``, as well as ``test-js`` for the JavaScript tests.
You'll notice that the test suite is structured based on where the code lives, and strongly encourages using the mock library to drive more accurate individual tests.
..note:: We use py.test for the Python test suite, and a combination of phantomjs and jasmine for the JavaScript tests.
Static Media
------------
Lemur uses a library that compiles it's static media assets (LESS and JS files) automatically. If you're developing using
runserver you'll see changes happen not only in the original files, but also the minified or processed versions of the file.
If you've made changes and need to compile them by hand for any reason, you can do so by running:
::
lemur compilestatic
The minified and processed files should be committed alongside the unprocessed changes.
It's also important to note that Lemur's frontend and API are not tied together. The API does not serve any of the static assets, we rely on nginx or some other file server to server all of the static assets.
During development that means we need an additional server to serve those static files for the GUI.
The gulp task compiles all the JS/CSS/HTML files and opens the Lemur welcome page in your default browsers. Additionally any changes to made to the JS/CSS/HTML with be reloaded in your browsers.
Because Lemur is just Flask, you can use all of the standard Flask functionality. The only difference is you'll be accessing commands that would normally go through manage.py using the ``lemur`` CLI helper instead.
For example, you probably don't want to use ``lemur start`` for development, as it doesn't support anything like
automatic reloading on code changes. For that you'd want to use the standard builtin ``runserver`` command:
::
lemur runserver
DDL (Schema Changes)
--------------------
Schema changes should always introduce the new schema in a commit, and then introduce code relying on that schema in a followup commit. This also means that new columns must be NULLable.
Removing columns and tables requires a slightly more painful flow, and should resemble the follow multi-commit flow:
All patches should be sent as a pull request on GitHub, include tests, and documentation where needed. If you're fixing a bug or making a large change the patch **must** include test coverage.
Uncertain about how to write tests? Take a look at some existing tests that are similar to the code you're changing, and go from there.
You can see a list of open pull requests (pending changes) by visiting https://github.com/netflix/lemur/pulls
