HTTP REST service for text summarization. Build with Flask, MySQL, and Hugging Face Transformers.

Why?

• The goal was to have a working application that could be build on top of.
• To practice creating a small system that utilizes Flask, MySQL, and Machine learning
• Learning/Practice

There are two endpoints of interest:

This will summarize text and store it. Returns a json object with the following:

{ 
  summary_id=summary_id, 
  summary=text_summary
}

content-type: application/x-www-form-urlencoded

POST parameters: text, where text is the text to be summarized.

Retrieve the text data using summary_id. Returns a Json object with the following:

{
  long_text=long_text, 
  text_summary=text_summary
}

where long_text is the original text and text_summary is the summarized text. Parameters content-type: application/json POST parameters: summary_id, the PRIMARY KEY ID of the summarized text.

The app runs on Flask, with MySQL as a database to store text summaries. Further, the app has been containerized using Docker and docker-compose.

To run:

sudo docker-compose up

and then navigate to http://0.0.0.0:5000/ in a browser.

This brings up a small webpage that makes use of the summary API. The webpage is intended to just showcase how the endpoints work, and to test summarizing text.

If you do not have docker-compose installed, or are unfamiliar with it, one could also do the following:

• using any python environment, install requirements.txt
• initialize a database summary using MySQL, and run ./db/init.sql
• update local MySQL credentials in ./app/backend/database/connector.py:
• run python wsgi.py
• navigate to http://0.0.0.0:5000/