######################################

Steps to test REST web service

######################################

• Make sure Docker is installed

• Open folder in terminal

• Run:

  $ docker load -i woz.tar
  $ docker run -d -p 5000:5000 woz:latest
  

• Open browser and make test request by putting this in the address bar:

    http://0.0.0.0:5000/woz/?m2=80&single=20&married_nokids=20&notmarried_nokids=20&married_kids=20&notmarried_kids=20&single_parent=20&total=120
  

• If this does not work, check what host the server is on. Run this in terminal:

    $ docker ps
  

• Replace 0.0.0.0 in the original link with the host the server is on.

• Change the values of the parameters to get different output!

• When done testing, find the container id with:

    $ docker ps
  

• Then use this id to run the following command and shut down the container:

    $ docker kill [container id]
  

##################################

Other files in this folder

##################################

Everything else in this folder is what I used to analyze and process data, train a model and test the webservice locally.

Contents:

WOZ 
│   README.txt                  - this file 
│   requirements.txt            - required python packages
│   data_analysis.py            - script to analyze data, can be used both before and after processing data
│   data_processing.py          - script that processes data, as described in report
│   evaluate_random_forest.py   - script to evaluate the random forest algorithm on train/test split of the data
│   train_random_forest.py      - script to train model on all data
│   utils.py                    - some utilitary functions that are used by other files
│   test_request.py             - script I used for easy local testing of web service
│   woz.tar                     - docker image (not tracked by git because of size)
│   .gitignore
│
└───data
│   │   original .xlsx files with provided data
│   │   .npy files with processed data
│   
└───out
|   │   data_analysis.txt                                            - results of data_analysis.py, before processing
|   │   data_analysis_after_processing.txt                           - results of data_analysis.py, after processing
|   │   trained_model.sav                                            - model trained in train_random_forest.py
|   │   .png files with histograms showing features' distributions   - data visualization obtained with data_processing.py
|
└───RESTWebService
    │   app.ini                - configuration of wsgi
    │   app.py                 - the restful flask api
    │   Dockerfile
    │   requirements.txt       - required python packages
    │   trained_model.sav      - model used by server
    │   wsgi.py                - server's entry point
    │   .dockerignore