Clothing Recommender

Authors: Pablo Kvitca, Tatiana Ediger

This project models a Clothing domain with different types of clothes, their attributes, and relationships on a Graph based model. We use the Neo4j Database for storage and analysis of this graph. The project allows us to add, list, see, describe, find, and most interestingly recommend clothes to purchase/wear to the user.
Additionally, we use Testcontainers to simplify out Data-Access Layer Integration testing process.

Technologies Used

• Neo4j Database - Data storage & modelling
• Neo4j-OGM - Interfacing with DB from Java
• Logback - Logging
• JUnit 5 - Jupiter - Testing Framework
• Testcontainers - Neo4j - Integration Testing
• Docker

Project Status

The project currently just has a demo Command Line based interface, intended to show the capabilities of the Graph model and database. Currently some of the recommendation options are implemented (recommend to a user, recommend to a user using related items, and recommend to a user using similar items). To see this in action you can use the Main program as a command line interface, as follows:

Requirements

• Java 11.0.6
• Neo4j Database - IMPORTANT: should have user called neo4j with password demo - and should be actively running with the Bolt protocol (see Neo4j documentaito) on port #7687. Note that these are the default network configs.
• For running tests: Docker should be running.

Available Demo Commands

1. filldb - seed the database with demo data. SHOULD BE RAN FIRST

• list [label] -- lists all available of label
  • clothings -- lists all available of clothings
  • bottoms -- lists all available of bottoms
  • tops -- lists all available of tops
  • footwears -- lists all available of footwears
  • users -- lists all available of users
  • descriptors -- lists all available of descriptors
  • colors -- lists all available of colors
  • brands -- lists all available of brands
  • fanciness -- lists all available of fanciness
  • subtypes -- lists all available of subtypes
  • groupings -- lists all available of groupings
  • collections -- lists all available of collections
  • sets -- lists all available of sets
• closet [username] -- lists every clothing owned by the given user
  • [username] is the unique identifier for the user
• describe-clothing [catalogID] -- lists every descriptor and grouping the clothing is in
  • [catalogID] is the unique identifier for the clothing
• described-by [descriptor type] [descriptor name] -- lists every clothing related to the given descriptor
  • [descriptor type] can be: "color", "brand", "subtype", or "fanciness"
  • [descriptor name] should be a type of descriptor on the database. Example: blue, Vans, jeans, fancy
• grouped-by [grouping type] ["][grouping name]["] -- lists every clothing related to the given grouping
  • [grouping type] can be: "collection" or "set"
  • [grouping name] shoud be the name of the grouping (can be between quotes if the name has spaces)
• add-clothing [clothing type] ["][clothing name]["] [catalogID] [[descriptor type] [descriptor name] [descriptor type] [descriptor name] ...] -- adds the given clothing to the DB, with the given name, catalog id, type, and descriptors. Groupings need to be added on a separate command.
  • [clothing type] can be "bottom", "top", or "footwear"
  • [clothing name] can be any string for the name of the clothing (use quotes if it contains spaces)
  • [catalogID] should be a unique string identifier for the clothing
  • An unlimited number of [descriptor type] [descriptor name] pairs is then accepted. These can be existing or new to the DB
• add-to-closet [username] [catalogID] -- adds the given clothing as owned by the given user
  • [username] is the unique identifier for the user
  • [catalogID] is the unique identifier for the clothing
• add-user [username] ["][readable name]["] -- adds the given user to the DB
  • [username] is the unique identifier for the user
  • [readable name] is the printable name of the user (use quotes if it contains spaces)
• add-to-grouping [catalogID] [grouping type] ["][grouping name]["] -- adds the given clothing as being on the given grouping
  • [catalogID] is the unique identifier for the clothing
  • [grouping type] can be: "collection" or "set"
  • [grouping name] shoud be the name of the grouping (can be between quotes if the name has spaces). This can be existing or new to the DB
• recommend-purchase-related [username] [catalogID] -- recommends items related (by groupings) to the given clothing, that the user does NOT own already
  • [username] is the unique identifier for the user
  • [catalogID] is the unique identifier for the clothing
• recommend-purchase-similar [username] [catalogID] -- recommends items similar (by descriptors) to the given clothing, that the user does NOT own already
  • [username] is the unique identifier for the user
  • [catalogID] is the unique identifier for the clothing
• recommend-purchase [username] -- recommends items based on similar users to the given user, that the user does NOT own already
  • [username] is the unique identifier for the user

Future commands

• recommend-wear-related [username] [catalogID] -- recommends items related (by groupings) to the given clothing, that the user DOES own already
  • [username] is the unique identifier for the user
  • [catalogID] is the unique identifier for the clothing
• recommend-wear-similar [username] [catalogID] -- recommends items similar (by descriptors) to the given clothing, that the user DOES own already
  • [username] is the unique identifier for the user
  • [catalogID] is the unique identifier for the clothing
• recommend-wear [username] -- recommends items based on similar users to the given user, that the user DOES own already
  • [username] is the unique identifier for the user