/flask_app_with_ml_model

Primary LanguagePython

Update 1.0.0

  • Taking the ML model from AWS s3 bucket
  • Storing the processed data and results on MongoDB hosted on AWS EC2

Introduction:

This project is demonstrating how we can integrate a Machine Learning(ML) model with a Python Flask application and generate an API. This code extract the model from AWS S3 and store the model outputs in MongoDB living on AWS EC2.

Problem

Incoming users on the website should see intelligent search results on the live website.

Solution

Building a Flask API which will allow user to insert the data and get results from ML model.

Requirements:

  • Python 3.7
  • Docker installed and set up (Docker has a very detailed documentation, https://docs.docker.com/engine/install/ubuntu/)
  • Virtual env with all project libraries installed
  • AWS account with S3 and MongoDB
  • Credential file with AWS creds
  • Model stored in s3

Setup:

  1. Create a virtual env
  • python -m venv venv
  1. Activate the virutal env
  • source venv/bin/activate
  1. Install project libraries
  • pip install -r requirements.txt
  1. set flask app with the command
  • SET FLASK_APP=test_2.py
  1. Then execute flask
  • flask run

Dockerizing Flask app:

Docker commands

  • Navigate to apps folder inside DevOps to run docker commands
  1. Remove already existing docker image (if any)
  • docker image rm --force stringdetectorml
  1. Build a new docker image
  • make build-string-detector
  1. Run docker image
  • docker run -p 5001:5001 stringdetectorml
  1. Investigate the image
  • docker run -it stringdetectorml sh