AchilleasKn/flask_api_python

Deploying a Machine Learning Model as a REST API with Flask

Model Deployment as API | The Iris Dataset

Deploying a Machine Learning Model as a REST API with Flask

Data Set Information

This is perhaps the best known database to be found in the pattern recognition literature. The data set contains 3 classes of 50 instances each, where each class refers to a type of iris plant. One class is linearly separable from the other 2; the latter are NOT linearly separable from each other.

Predicted attribute: class of iris plant.

Attribute Information

1. sepal length in cm
2. sepal width in cm
3. petal length in cm
4. petal width in cm
5. class: -- Iris Setosa -- Iris Versicolour -- Iris Virginica

Steps

1. Build and train the machine learning model in a Jupyter Notebook (file: model/Iris_model.ipynb),
2. save the model in a (pickle) file (file: api/iris_model.pkl)
3. create an API application that uses the pre-trained model to generate predictions (file: api/api.py),
4. encapsulate the application in a Docker container (file: api/Dockerfile),
5. deploy the application to a cloud server.

Technical Requirements

• Python 3.4+,
• Docker,
• The required Python libraries used can be installed from the included requirements.txt file:

Running the application locally

Directly

# Clone the project
git clone https://github.com/AchilleasKn/flask_api_python.git

# Change Directory
cd flask_api_python/api

# Install pip for Python3
apt install python3-pip

# Install the requirements
pip3 install -r requirements.txt

# Run the script in Python
python3 api.py

On Docker

Available images:

• achilleaskn/flask_api_python:latest

This image is based on the python:3.6-jessie official image

• achilleaskn/flask_api_python:alpine.latest

This image is based on Alpine Linux image which is a lightweight version of Linux

From scratch

# Clone the project
git clone https://github.com/AchilleasKn/flask_api_python.git

# Change Directory
cd flask_api_python/api

# Build the docker image
docker build -t flask_api .

# For the alpine version run the following
#docker build -f Dockerfile.alpine -t flask_api .

# Run the flask_api image and expose the 5000 port 
docker run -d -p 5000:5000 flask_api

# To see the running containers
docker ps 

# To see the logs of our running container
docker logs <Container ID>

With Docker Pull

# Pull the docker image
docker pull achilleaskn/flask_api_python:latest

# For the alpine version run the following
#docker pull achilleaskn/flask_api_python:alpine.latest

# Run the flask_api image and expose the 5000 port 
docker run -d -p 5000:5000 achilleaskn/flask_api_python:latest

# For the alpine version run the following
#docker run -d -p 5000:5000 achilleaskn/flask_api_python:alpine.latest

# To see the running containers
docker ps 

# To see the logs of our running container
docker logs <Container ID>

Testing the application

Once it is running, the API can be queried using HTTP POST requests. I recommend using postman for testing.

URL: http://0.0.0.0:5000/predict

• Sample query for "Setosa" type:

{
	"feature_array":[4.9, 2.9, 1.2, 0.3]
}

The response should look like this:

{
    "prediction": [
        0
    ]
}

• Sample query for "Versicolour" type:

{
	"feature_array":[6.4, 3.2, 4.5, 1.5]
} 

The response should look like this:

{
    "prediction": [
        1
    ]
}

• Sample query for "Virginica" type:

{
	"feature_array":[6.2, 3.1, 5.3, 2.4]
} 

The response should look like this:

{
    "prediction": [
        2
    ]
}