Skip to content

Architecture and Design

Jump to bottom
Liad edited this page Apr 20, 2018 · 1 revision

The project consists of 3 different databases:

  • Document DB
  • Users DB
  • Recommendations DB

Document DB

The document DB is an key-value DB, like redis or MongoDB. given a document ID or URL (key), it retrieves the related document data

The results of the NLP pre-processing as well as the extracted topics are saved in the value part as well

Users DB

This is where the user contact details and authentication details are stored. It can be either Tabular, key-value, or any other idea.

It should be flexible enough to be able to group the users (for example by organizations)

Recommendations DB

The information regarding:

  • which document is recommended to which user
  • which user liked/didn't liked which document recommendation

The DBs are wrapped with a single layer - Data access layer - and therefore they can be distributed and even replaced, while the API stays the same, and the settings information is saved in the layer / docker container.

Clone this wiki locally