Skip to content

Latest commit

 

History

History
60 lines (46 loc) · 2.42 KB

README.md

File metadata and controls

60 lines (46 loc) · 2.42 KB

aguefort

🎲 A chatbot for asking about D&D tips powered by Dropout's Adventuring Academy knowledge and LLMs.

Prerequisites

  • AWS Account
  • Docker
  • Poetry

Downloading Adventuring Academy Data

  1. Clone dropout-dl repo: git clone https://github.com/mosswg/dropout-dl.git
  2. Build Docker image: cd dropout-dl && docker build -t dropout-dl:latest .
  3. Remove the dropout-dl repo, it's no longer needed: cd .. && rm -rf dropout-dl
  4. Clone aguefort repo: git clone https://github.com/rey-allan/aguefort.git
  5. Install all dependencies and create virtual environment: poetry install && poetry shell
  6. Create login file at the root of the repo with your Dropout email and password, one per line
    email@example.com
    123456789
    
  7. Run the script: python3 scripts/download_captions.py
  8. Go grab a coffee while the script runs, it will take some time
  9. Captions will be saved to data/captions in the format season-episode_number.vtt

Generating the Knowledge Base

Note: Make sure to download the Adventuring Academy data before (see above)!

  1. Run the script: python3 scripts/generate_kb.py
  2. Go to sleep while the script runs, it will take a long time
  3. Knowledge base will be saved to data/adventuring_academy_kb

Running the App

Note: Make sure to generate the knowledge base before (see above)!

  1. Request access to Claude family of models via Bedrock in your AWS Account
    • Make sure to specify your use of these models will be for personal projects only
    • Access should be granted in a few minutes
  2. Create an IAM user in your AWS Account with the AmazonBedrockFullAccess policy attached to it, and note down your access and secret keys
  3. Create an .env file inside app/ folder with the following keys:
    AWS_ACCESS_KEY_ID=<YOUR_ACCESS_KEY>
    AWS_SECRET_ACCESS_KEY=<YOUR_SECRET_KEY>
    AWS_REGION=<REGION_WITH_BEDROCK_ACCESS>
    
  4. Run the app: python3 app/app.py --history $HISTORY
    • Where $HISTORY is the max number of previous messages from the chat history to use as context. It defaults to 5 messages.
  5. Open the app in your favorite browser, the default URL should look like this: http://127.0.0.1:7860
  6. Enjoy!

Screenshot

This is a screenshot of the kind of questions you can ask. And how you can also ask follow-up questions and continue chatting about the topics.