Flow is a website that lets you plan courses with friends.
JavaScript Python CSS HTML Shell Makefile
Clone or download
jswu Merge pull request #323 from UWFlow/sandy-facebook-api-name
Update Facebook login to expect name
Latest commit 2c9c636 Jun 19, 2018
Failed to load latest commit information.
Docker Switch Dockerfile to point to UWFlow/rmc instead of JGulbronson/rmc Jun 10, 2016
analytics Add stats for ratio of users/courses that have/(have been) rated or r… Feb 27, 2014
config Disable compiling (minifying) JS files on deploy Apr 20, 2014
data Remove unused CritiqueCourse document Nov 23, 2016
emails Add clause about sharing with friends Mar 24, 2014
html_snapshots Prevent PhantomJS requests to Google Analytics Jun 3, 2015
kittens with, png -> jpg, copy changes Feb 17, 2014
migrations Revert "Revert "Merge pull request #28 from divad12/jlfwong-linter"" Jan 24, 2014
models Merge pull request #307 from JGulbronson/add-documentation Dec 4, 2017
server Update Facebook login to expect name Jun 18, 2018
server_setup Added rvm Aug 6, 2016
shared Use pagination for Facebook friends Nov 12, 2016
test Add test for removing user course Aug 5, 2014
third_party Added tests for the schedule parser, and moved it to it's own file Apr 19, 2014
tools Revert "Revert "Merge pull request #28 from divad12/jlfwong-linter"" Jan 24, 2014
.gitignore Course recommendation engine for UW Flow Jun 20, 2016
.gitmodules Revert "Revert "Change all references to old repos to point to the ne… Feb 25, 2014
.travis.yml Use Docker on Travis May 29, 2016
CONTRIBUTING.md Change CONTRIBUTING.md to be fork-and-pull instead of inter-branch Feb 28, 2014
Code of Conduct.md Proposing a code of conduct Dec 4, 2017
LICENSE.txt Add MIT License Feb 26, 2014
Makefile Merge branch 'master' into jlfwong-setup-instructions Dec 4, 2017
README.md Merge branch 'master' into jlfwong-setup-instructions Dec 4, 2017
Vagrantfile Add note about Vagrant version Apr 29, 2014
__init__.py copy some stuff to get courses into mongodb from old project Sep 10, 2012
commit-msg-hook Revert "Revert "Merge pull request #28 from divad12/jlfwong-linter"" Jan 24, 2014
deploy.sh Remove old Hipchat deploy notify script Dec 22, 2016
docker_on_connect.sh Skip needing `cd /rmc && source ~/.virtualenv/rmc/bin/activate` Jun 27, 2016
docker_shell.sh Skip needing `cd /rmc && source ~/.virtualenv/rmc/bin/activate` Jun 27, 2016
docker_tests.sh Use Docker on Travis May 29, 2016
js_test.sh Added Travis CI yml file and badge to README, currently only runs JS … Oct 9, 2014
lint_blacklist.txt Basic scholarships with React Jun 10, 2015
linux-setup.sh Fixed broken links in linux setup Dec 27, 2016
local_server.sh Add instructions to run UWFlow locally in Docker Jun 27, 2016
local_uwsgi_server.sh Add script to test uwsgi server in dev Dec 3, 2013
mac-setup.sh Update url for homebrew install May 25, 2015
package.json Basic scholarships with React Jun 10, 2015
prod_server.sh Attempt to fix prod_server.sh after uwsgi update Jan 24, 2014
requirements.txt updated version of uWSGI Nov 14, 2016
setup.sh Added mocha-phantomjs to make passing the js-tests a deploy blocker Apr 23, 2014


Build Status


Plan your courses

Getting up and running

The fastest way to get running is to use Docker. This will let you run Flow inside of a virtual machine to avoid dealing with package installation problems and to avoid polluting your development environment. It should also take less time.

To get started, install Docker.

Once you have docker installed, run:

$ make shell_in_docker

This will download and run a Docker image with all of Flow's dependencies already installed. Don't worry if you don't know what a Docker image is.

It will only need to download the image once, so booting up in the future should be much faster.

Once the image is downloaded and running, you should find yourself in a /bin/bash shell. This is running inside a Docker container (effectively a virtual machine). Inside this container, the rmc repository can be found in /rmc, which is where you should be right after running make shell_in_docker.

To start running Flow locally, run the following inside this new shell.

$ make local

If you point your browser at http://localhost:5000/, you should now see the Flow homepage running on your computer!

Congratulations! You now have Flow running locally.

Getting seed data

To do anything interesting in Flow, you need data in your database. This is where information about courses, professors, and scheduling information is stored.

To get started, open a new terminal (different from the one running make local), and start a new shell inside Docker by running

$ make shell_in_docker

Inside the Docker container, run the following to get some basic course data into the DB.

$ make init_data

Directory structure

If you're eager to dive into the code, you might want to read this first. This isn't exhaustive, but it should be enough to get you started if you want to contribute.

  • config/: Configuration for frameworks, databases, or anything that might vary between the development environment and production.
  • data/: This is where we collect data and load it into the database
    • crawler.py downloads data by scraping pages and hitting APIs
    • processor.py processes the data grabbed by crawler.py and loads it into the DB
    • aggregator.py is run on a regular schedule (daily for the most part) to keep our data up to date
  • models/: "Schema" definitions for our models backed by MongoEngine
  • server/: Request handlers, static assets, and templates
    • templates/: Jinja2 templates
      • Files in here ending with _page.html (e.g. course_page.html) are rendered directly by the Flask server with render_template calls, with the exception of the base_*_page.html files which other _page.html templates inherit from.
      • Most of the other files (e.g. course.html) contain Underscore templates used to render stuff on the client-side
    • static: Static assets eventually ending up as files served directly by nginx when on production
      • js: All our JavaScript code, organized into RequireJS modules
        • ext/: All third party JavaScript code
        • main.js: The entry point for JavaScript executing on page load
      • sass: We don't write CSS directly for Flow, we use the SCSS flavor of Sass, which compiles down to CSS
    • server.py: The majority of the request handlers for the application, written in Flask

Using the REPL

If you need a REPL to fool around with the database or test out some code, check out tools/devshell.py.

It automatically loads some imports and connects to the database for you. This setup code can be found in tools/devshell_eval.py.

Here's what an example session might look like:

$ tools/devshell.py
Python 2.7.1 (r271:86832, Jul 31 2011, 19:30:53)
Type "copyright", "credits" or "license" for more information.

IPython 0.13.1 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]: m.User.objects(first_name__in=['Jamie', 'David', 'Sandy', 'Mack'],
last_name__in=['Wong', 'Hu', 'Duan', 'Wu'])
Out[1]: [<User: David Hu>, <User: Mack Duan>, <User: Sandy Wu>, <User: Jamie Wong>]

Why RMC?

It might seem funny that this repository and a bunch of the code references rmc.

RMC stands for "Rate My Courses", which was the prototype name for this project before it was given the (slightly) better name of Flow.

Because of the profileration of this 3 letter prefix throughout the code, and the unfortunate coupling of the repository name and our python namespace, we decided to leave it be.

Running tests

To run all the tests in the entire system:

make alltest

To run all the tests except the really slow ones (namely Selenium tests):

make test

To run all the tests under a specific directory tree or in a specific file:

PYTHONPATH=.. nosetests server/api
PYTHONPATH=.. nosetests server/api/v1_test.py


When you're ready to contribute, take a look at the contributing guidelines and our style guide.

Setting up without Docker

If you'd prefer to avoid the docker route, you can install the dependencies directly on your own machine.

To set up your dev environment, run make install.

We work inside a virtualenv, so remember to source ~/.virtualenv/rmc/bin/activate whenever you're working within the repo.

You should now be ready to boot the local server, with make local.

Once it starts running, point your browser to http://localhost:5000/

MongoDB error on Linux

If you are getting a connection refused error when trying to run make local and are on Linux, this is most likely due to MongoDB taking too long to start the first time it's run. To fix this, run mongod --config config/mongodb_local.conf and let it warm up for about 30 seconds to 1 minute. Then kill the process, and run make local again. It should work now.