Index Builder 
This is a nice little interactive survey using completely randomized data in which users can:
- create unique usernames
- select factors by giving them a weight, pro/anti direction & supporting reasons
- lock-in selections in order to see what kind of index your combination of factors have built
- compare performance of your index against some default indexes as well as other users
- an admin user can:
- view interesting breakdowns of all factor selections made by users as well as make these breakdowns available for users to see
- archive off user selections for further future analysis and to reset the app for another round of users
Demo
Live demo (choose any username you'd like): index-builder
Getting Started
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system. Once you've checked out this project you must make sure you have the following installed:
Include instructions on setting up virtualenv
To get up and running perform the following: pip install -r requirements.txt
In order to install scipy I had to run the following: sudo apt-get install gfortran libopenblas-dev liblapack-dev python-dev pip install scipy==0.15.1 easy_install numpy==1.9.2 easy_install pandas==0.16.2
You'll need node, npm & yarn installed yarn install npm run build
Prerequisites
- pip
- easy_install
- virtualenv
- node
- npm
- yarn
Installing
Setting up your virtual environment
1) mkdir -p ~/pyenvs
2) virtualenv ~/pyenvs/index_builder
3) source ~/pyenvs/index_build/bin/activate
4) pip install -r requirements.txt
5) sudo apt-get install gfortran libopenblas-dev liblapack-dev python-dev (assuming you're running linux)
6) pip install scipy==0.15.1
7) easy_install numpy==1.9.2
8) easy_install pandas==0.16.2
Building the javascript
1) yarn install
2) npm run build
Running the Flask back-end
python index_builder/server.py --HOST 0.0.0.0 --PORT 8080 --AUTH true
Running the Flask back-end with gunicorn
gunicorn server:app --config gunicorn_config.py --bind 0.0.0.0:8080 -w 8 -t 120 --limit-request-line 8190 --log-file=- --access-logfile ~/stdout --error-logfile ~/stderr --log-level=info
Building the javascript files in "watch" mode (this is good for development so it will hot swap changes)
npm run watch
Running the tests
Python tests
python setup.py test
All JS tests
npm run test
Specific JS tests
TEST=static/__tests__/factor_viewer/*-test.jsx npm run test-file
JS coverage report
npm run test-with-coverage
JS Duplication report
npm run report-duplicate-code
JS formatting
npm run format
JS linting
npm run lint -s
Deployment
You can run this application on a larger scale using gunicorn
export AUTH=true (if you want authentication turned on)
gunicorn server:app --config gunicorn_config.py --bind 0.0.0.0:8080 -w 8 -t 120 --limit-request-line 8190 --log-file=- --access-logfile ~/stdout --error-logfile ~/stderr --log-level=info
TBA: notes on docker deployment
Built With
- Flask - The back-end web framework used
- jinja - Core HTML templating
- yarn - JS Dependency Management
- Webpack - JS Bundler
- React - Core JS library our components are built on
Authors
- Andrew Schonfeld - Initial work - plutoplate
See also the list of contributors who participated in this project.
License
This project is licensed under the MIT License - see the LICENSE.md file for details
Acknowledgments
- Wilfred Hughes was truly an inspiration and a driving force behind many of the solutions to initial infrastructure headaches