RESTful-API with Flask

Example how to use Flask to build an API for Machine Learning data requests

Optimal architecture:

API<->Wrapper<->Model<-training Model

File

- ml_api.py: The API implemented as WSGI

Setup (on Ubuntu)

The pathlib library is included in all versions of python >= 3.4. Therefore I recommend using the most up2date version of Python 3. In addition we want to use the packages Flask for our API and Scikit Learn for loading our trained model.

Install Python packages:

$ sudo pip install -U sklearn, flask

Clone the repo or just download the file:

$ sudo apt-get install git
$ git clone https://github.com/.../....git api_directory

Run

WARNING!

• Debug mode should never be used in a production environment!
• Flask Dev Server should never be used in a production environment!

Run Dev Server with:

$ FLASK_APP=ml_api.py flask run
 * Running on http://localhost:5000/

Simple Example

Now, one can use tools like wget, cURL, Python (Requests) or your web browser to communicate with the API. If they aren't already installed, use:

$ sudo apt-get install curl

Outlook

REST-Framework

https://github.com/encode/django-rest-framework