Jarvis-MCMC API

Jarvis-MCMC is an API microservice that assists in running MCMC estimation procedures whether it be simulations or fitting models to new data.

Getting Started

Simple instructions to get set up in development environment are provided here

Deployment

The API is Docker-capable. Full instructions for getting started are here

Documentation

Documentation root for the project is provided here

To build the docs into html:

$ cd fibonacci_api/docs
$ make html

Documentation appears in $DEV_HOME/jarvis_mcmc_api/static/index.html

Testing

Run the tests with nosetests

$ cd $DEV_HOME/jarvis_mcmc_api
$ nosetests --with-coverage --cover-html --cover-package=jarvis_mcmc_api --cover-erase
.....

Name                         Stmts   Miss  Cover   Missing
----------------------------------------------------------
fibonacci_api.py                 0      0   100%   
fibonacci_api/common.py         16      2    88%   37-38
fibonacci_api/fibonacci.py      13      0   100%   
fibonacci_api/main.py           31      4    87%   23, 53-54, 75
----------------------------------------------------------
TOTAL                           60      6    90%   
----------------------------------------------------------------------
Ran 5 tests in 0.027s

Quick Start for the impatient

$ git clone https://github.com/drackham/jarvis-mcmc.git
$ cd jarvis_mcmc/jarvis_mcmc_api
$ virtualenv --no-site-packages python
$ source python/bin/activate               # ensure you are using the correct virtualenv python
$ pip install -r requirements.txt
$ python main.py

Use curl to interact with the API.

$ curl http://localhost:5000/jarvis_mcmc/list?count=10
{
  "answer": [
    0,
    1,
    1,
    2,
    3,
    5,
    8,
    13,
    21,
    34
  ]
}

Technology Stack / Dependencies

• Python 2.7 - All components are build in Python
• Nginx - HTTP Proxy to sitting in front of the Python Service
• Flask - Web application framework suitable for creating microservices in Python.
• Gunicorn - Pre-fork application server to efficiently serve the Flask application in production.
• Docker - Container technology to create simple and consistent deployments.
• Nose - Unit testing framework.
• Sphinx - Documentation generator