README

This exercise demonstrates a small http web server providing a RESTful API endpoint to calculate the sum of a list of integers in stream mode

• endpoints
  • /total/
    • method: post
    • payload: [1, 2, 3, 4]
      • a list of integers in json format. Other format or number type may cause an error message.
    • response: { "total": 10 }
      • a dictionary which contains a key named "total" whose value is the sum, in json format.

Prerequisites

• In order to benefit from Tornado's full capability, please run it on Linux. Otherwise, please just use docker instead.
• The current implementation is tested against Python 3.6, which is normally defined by the embedded Dockerfile.

Installation

• On a linux work station directly
  • Install python dependencies

    pip install -r requirements.txt

  • Install using setup.py

    python setup.py install

• Using container tool
  • Build docker image

    docker build -t sum_cal .

How to use?

• Launch the web server after an installation with Python

  calsum --http-address=0.0.0.0 --app-debug=false --enable-auto-fork

• Launch it using docker-compose

  docker-compose -f docker-compose.yml build
  docker-compose -f docker-compose.yml up

• Test it with a small python snippet

  • submit a big calculation

    import requests
    import json
    r = requests.post("http://localhost:8000/total/", data=json.dumps(list(range(10000001))))
    print(r.json())

    expected output

    {'total': 50000005000000}

  • submit some calculation requests in parallel

    import requests
    import json
    from multiprocessing import Pool
    
    
    def test_request(x):
        r = requests.post("http://localhost:8000/total/", data=json.dumps(list(range(100001+x))))
        return json.loads(r.content.decode()).get("total")
    
    def gen_result(x):
        return sum(range(100001 + x))
    
    with Pool(10) as p:
        expected = [gen_result(x) for x in range(100)]
        for e, r in zip(expected, p.map(test_request, [_ for _ in range(100)])):
            if e != r:
                raise Exception("{}!={}".format(e, r))
        print("ok at this scale")

    expected output

    ok at this scale

How to develop?

• Install dependencies

  pip install -r requirements.txt

• Install in development mode and run the server

  python setup.py develop
  calsum --http-address=0.0.0.0 --app-debug=true

• Run tests

  python setup.py test