In this article i will show the core part of this simple dockerized flask app used for showing a regression plot image.
At first we should create a Dockerfile that uses a requirements file to install all needed dependencies:
matplotlib==3.3.0
seaborn==0.10.1
pandas==1.0.5
flask==2.1.0
Now let’s see the Dockerfile:
FROM python:3.8
COPY ./requirements.txt /requirements.txt
WORKDIR /
RUN pip3 install -r requirements.txt
COPY . /
EXPOSE 5000
ENTRYPOINT ["python3"]
CMD [ "app/app.py" ]
We are ready to setup flask and create the app.py file and the plotdata.py file used to plot the image
# app.py
from flask import Flask, send_file
from plotdata import regression_plot
app = Flask(__name__)
@app.route('/', methods=['GET'])
def regr_plot():
image = regression_plot()
return send_file(image,
attachment_filename='regplot.png',
mimetype='image/png')
if __name__ == '__main__':
app.run(host='0.0.0.0', debug=False)
#plotdata.py
import pandas as pd
import matplotlib
import seaborn as sns
import io
if __name__ =='__main__':
from PIL import Image
matplotlib.use('agg')
def regression_plot():
df = pd.read_csv('tempYearly.csv')
sns_plot = sns.regplot(x='Rainfall', y='Temperature', data=df)
image = io.BytesIO()
sns_plot.figure.savefig(image, format='png')
image.seek(0)
return image
if __name__ == '__main__':
image = regression_plot()
im = Image.open(image)
im.save('regress.png')
Finally we need to build the image using this command:
docker build --tag flask-plotting-app .
then running the container:
docker run -it --name flaskpltapp -p 5000:5000 flask-plotting-app:latest
This is the final result
This is the repository link of this project: Dockerized Flask regression app
