In this project I created a webapp using flask that you can go to at project. In this website you can input some text into a form that will then generate a sentence from that text. It does this by creating a markov chain with the text. If you don't know what a markov chain is you can go to my website where introduce markov chains. But in its simplest form it creates a dictionary with probabilities. These probabilities can be used to predict what words come after each other. With this we can create a sentence. Down below will be a visual to even further understand.
The image above is a simple markov chain. As you can see Point E is more likely to go to Point A, but Point A is just slightly more likely to go back to itself. A real life example of this is the weather. Using a markov chain you can predict what the weather will be given the weather before. If we create a visual representation for the weather a simple visual would look like this. In pi
To install the requirements we can do pip install -r requirements.txt. Because Im using flask for this project to start the project we will need to run export FLASK_APP=app.py and lastly flask run
This app is currently deployed on heroku here. In a production environment, the application’s front end will live on a different server from the markov model. However, this model is lightweight so to keep things simpler I left it on the same server. Ideally the model would be on a seperate gcp for aws server and my heroku server that hosts the frontend would just call my model server and return some output. I was previously able to do this by using google cloud platform for my model server and heroku for frontend. I needed the free gcp credits for a larger project so for now the model doesn't have a seperate server. Therefore api.py is not being used.