This is a fancy way of saying it doesn't do anything useful yet, and probably won't run on your setup
Requires: Python 2.7.9, Docker, Redis (Only in Production)
Setup
# Get source code
git clone https://github.com/hamiltont/WebJuice.git
cd WebJuice
# Install dependencies
virtualenv env
source env/bin/activate
pip install -r requirements.txt
# If running in developer mode, use docker to setup a
# container running Redis
python start_redis.pyLaunch
Simply put, launching is as easy as calling foreman start. This will spin
up Flask, Flower, and Celery.
However, before you can launch you should set environment variables so we can connect to your Docker host. Using Docker Machine is easy, it will automatically setup Docker in a range of environments (OS X, Windows, Linux, Amazon, Digital Ocean, Rackspace, etc) and setup these environment variables
Docker environment variables:
- Required :
DOCKER_HOST - Required :
DOCKER_HOST_IP - Recommended:
DOCKER_TLS_VERIFY - Recommended:
DOCKER_CERT_PATH
After you setup a Docker provider using docker-machine, use
$ eval "$(docker-machine env)"
$ export DOCKER_HOST=`docker-machine ip`Management Interfaces
There are hyperlinks on the About page to important services, including Flask, Flower, and the RabbitMQ Management Interface
Halting
Using Ctrl-C with foreman works fine 95% of the time. Occasionally
you may need to kill orphaned python processes using killall -9 python
Ok, fair question for a project that benchmarks web frameworks...
I chose Python for consistency with the FrameworkBenchmarks project, full stop. I'll debate language preference as much as the next developer, but in this case I wanted language consistency to encourage maintainability and future contributions.
Why Flask? I hadn't used any Python web frameworks before this project,
so I made the best judgement I could.
Flask didn't require many new concepts for me (I suspect it won't for
most people---yay future contributions), and has a nice balance of
flexibility vs simplicity.
A big seller for me was also Flask's clean supports JADE templates
(via pyjade)--I like the cleanness of JADE and how its space-dependant
style is similar to Python.
For open source especially, I tend to value having working code available at a minimum level of performance--this codebase is my attempt to get a continuous system off the ground. I won't be surprised or sad at all if someone rewrites the entire app one day (in fact, it would be a huge success!). Happy hacking ;-)
