/enquir-backend

APIs using Information Retrieval and Natural Language Processing to generate notecards and questions

Primary LanguageJavaScript

enquir's API

The backend for a mobile app to help with studying

Users will be able to create notecards with its content automatically populated by typing in a topic or speaking the topic name to the app. The app will then generate questions given the information and send them to the user via push notifications.

Details

The api consists of two microservices:

Information Retrieval

The Information Retrieval API uses axios, Google's Custom Search Engine, and cheerio to fetch the top webpages on a given topic, scrape the pages, and return their content.

Hosted with AWS Lambda.

Question Generation

The Question Generation API uses spaCy's API to map the syntax tree, tag the part-of-speech, and perform named entity recognition. With these, the API parses a given blurb of information for its nominal subject, root verb, auxillary verb (if necessary), and object, and then creates WH-type questions.

Hosted with Heroku

4/30/2020: Question generator works best with simple Subject-Verb-Object sentence structures.