League of Legends Analysis with LLMs

Overwiew

This project automates the extraction and analysis of YouTube videos related to professional League of Legends matches. The script connects to the YouTube API, retrieves the transcript of a given video, processes it, and sends it to GPT-4 with a carefully designed prompt. The LLM then generates an analytical summary that highlights key moments, plays, and strategic decisions. This tool is intended to help those who are not fully familiar with the game understand and enjoy professional matches.

Features

• Extracts transcripts from YouTube videos using the YouTube API.
• Processes and cleans the transcript data to improve analysis quality.
• Sends the processed text to GPT-4 with a structured prompt.
• Generates an analytical summary that highlights:
  • Champs selection.
  • Key plays and strategies.
  • Crucial moments in the match.
  • Tactical decisions and their impact.
• Provides a user-friendly interface via Gradio for easy interaction.

Installation

To run the notebook, install the required dependencies:

pip install youtube_transcript_api openai gradio

Usage

1. Extract Video ID: The script extracts the YouTube video ID from a given URL.
2. Download Transcript: It fetches the transcript for the video ID.
3. Process Transcript: The text is cleaned and structured for better readability.
4. Generate Summary: The cleaned transcript is sent to GPT-4 for analysis, producing an insightful breakdown of the match.
5. Deploy with Gradio: The project can be deployed with a Gradio interface for easier usage.

Example function calls:

video_id = extract_video_id("https://www.youtube.com/watch?v=exampleID")
transcript = download_transcript(video_id)
summary = generate_match_summary(transcript)

To run the Gradio interface:

import gradio as gr
iface = gr.Interface(fn=generate_match_summary, inputs="text", outputs="text")
iface.launch()

Requirements

• An OpenAI API key is required to use GPT-4. Set it as an environment variable or provide it in your script.

export OPENAI_API_KEY='your-api-key-here'

Error Handling

• If a transcript is unavailable, the script handles errors gracefully and provides debugging information.
• Ensures the prompt sent to GPT-4 is optimized for match analysis.

Purpose

This project is designed to bridge the gap for newcomers to League of Legends esports by providing high-quality, AI-generated summaries that explain matches in an accessible way.

License

This project is for educational and competitive use only. Please ensure compliance with YouTube's API policies and OpenAI's terms of use.