mock_rec_experiment

This repository contains an adapted version of the STELLA Micro-Template, designed as an experimental system for recommendations within the Mock Recommendation System. The code specifically focuses on providing recommendations based on the RELISH documents, which are part of the system's dataset.

Overview

The mock_rec_exp repository serves as an experimental recommendation model within the Mock Recommendation System. Unlike the baseline system, this experimental model sorts recommendations for each PubMed article in the RELISH Database either alphabetically or in ascending order The goal is to integrate STELLA within the Mock Recommendation System and evaluate the effectiveness of the integration.

Purpose

The primary objective of this repository is to:

• Serve as an experimental system recommendation system for the Mock Recommendation System introducing variations in the recommendation sorting mechanism.
• Integrate STELLA into the recommendation pipeline, testing its performance in comparison to the baseline.
• Compare user statistics between the baseline and experimental systems, such as user interactions (e.g., clicks), to assess whether the new approach yields improvements in recommendation relevance or user engagement.

Developer Documentation

Prerequisites

Before starting, ensure that Docker is installed on your system. You can follow the installation instructions here.

Data

To download the necessary data files, execute the following script:

./data.sh

This will create a data folder and download two files "relish_text.jsonl" and "relish_recoms.jsonl".

Running the Application

1. Build the Docker Image:

To build the Docker image for the application, run the following command:

sudo docker build -t mock_rec_exp .

2. Run the Docker Container: After building the image, you can run the application in the terminal:

sudo docker run -it mock_rec_exp

Testing the Application

Once the application is running, you can test it by visiting the following URLs:

1. Check Indexing Status: Navigate to http://0.0.0.0:5000/index. The page should display the message: Indexing is done!

2. Test Recommendations: To test the recommendation functionality, visit: http://0.0.0.0:5000/recommendations/publications?item_id=22569528.

The response should look like this:

{
  "item_id": "22569528", 
  "itemlist": [
    17928366,
    18280112,
    18332145,
    18463290,
    18562239,
    18665890,
    18983981,
    19052640,
    19060905,
    19242111,
    19244124,
    19414607,
    19805545,
    19816936,
    20079430,
    20811985,
    22028468,
    22177953,
    23549785,
    23712012
  ], 
  "num_found": 20, 
  "page": 0, 
  "rpp": 20
}