llm_recommendation

App for recommending anime with LLMs

Quick Start Guide

Prerequisites

• Docker installed on your machine.
• An OpenAI API key (if using OpenAI services).

Build the Docker Image and Run the Project

To build the Docker image and run the project, run the following commands in the project directory:

docker build -t llm_rec .
docker run -d --name my_flask_app -e OPENAI_API_KEY=$OPENAI_API_KEY -p 5000:5000 llm_rec

API Endpoints

• GET /: Returns a welcome message.
• GET /anime: Queries the anime dataset based on various filters. E.g. /anime?mt_episodes=200&lt_episodes=300&mt_rating=7&mt_members=1000&genre=Drama queries for a Drama anime with 200 to 300 episodes, a rating of more than 7, and more than 1000 members.
• GET /user: Queries the user dataset based on various filters. E.g. /user?user_id=567&mt_rating=5&lt_rating=7 queries for all entries of user 567 with a rating of more than 5 and less than 7.
• GET /recommend/int:user_id: Provides recommendations for a specified user. The user can specify the number of recommendations to return using the k parameter. If not specified, the default value is 5.
• GET /anime: Provides a more natural experience for the user to get recommendations.

Querying Anime Data

You can query anime data using the /anime endpoint with various parameters. Here are some examples:

List of anime filters:

• anime_id - Get anime information by ID.
• name - Get anime information by name.
• genre - Get anime information by genre.
• type - Get anime information by type.
• episodes - Get anime information by episode count.
• mt_episodes - Get anime information by more than a certain episode count.
• lt_episodes - Get anime information by less than a certain episode count.
• rating - Get anime information by rating.
• mt_rating - Get anime information by more than a certain rating.
• lt_rating - Get anime information by less than a certain rating.
• members - Get anime information by number of members.
• mt_members - Get anime information by more than a certain number of members.
• lt_members - Get anime information by less than a certain number of members.

You can combine multiple parameters to refine your search. For example:

GET /anime?genre=Adventure&type=Movie&mt_rating=7.5

This will return anime that are of the Adventure genre, are Movies, and have a rating greater than 7.5.

Querying User Data

You can query user data using the /user endpoint with various parameters. Here are some examples:

List of user filters:

• user_id - Get user information by ID.
• anime_id - Get user information by anime ID.
• rating - Get user information by rating.
• mt_rating - Get user information by more than a certain rating.
• lt_rating - Get user information by less than a certain rating.

You can combine multiple parameters to refine your search. For example:

GET /user?anime_id=456&lt_rating=7.5

This will return users that watched anime with ID 456 and have rated it less than 7.5.

Querying Recommendations

You can query recommendations using the /recommend/<int:user_id> endpoint with various parameters. Here are some examples:

List of recommendation filters:

• k - Get recommendations for a specified user. The user can specify the number of recommendations to return using the k parameter. If not specified, the default value is 5.

You can combine multiple parameters to refine your search. For example:

GET /recommend/123?k=7

This will return the top 7 anime recommendations for user with ID 123.

Querying Natural Language

You can query the natural language interface using the /anime endpoint with various parameters. Here are some examples:

GET /talk/123/I%20want%20a%20happy%20anime

This will return anime recommendations for user with ID 123 that are uplifting and happy.