Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rey-allan/aguefort
:game_die: A chatbot for asking about D&D tips powered by Dropout's Adventuring Academy knowledge and LLMs.
https://github.com/rey-allan/aguefort
claude-3-haiku dungeons-and-dragons faiss generative-ai gradio langchain large-language-models rag
Last synced: 13 days ago
JSON representation
:game_die: A chatbot for asking about D&D tips powered by Dropout's Adventuring Academy knowledge and LLMs.
- Host: GitHub
- URL: https://github.com/rey-allan/aguefort
- Owner: rey-allan
- License: mit
- Created: 2024-03-09T06:16:38.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2024-03-27T04:08:40.000Z (9 months ago)
- Last Synced: 2024-11-04T08:32:45.686Z (about 2 months ago)
- Topics: claude-3-haiku, dungeons-and-dragons, faiss, generative-ai, gradio, langchain, large-language-models, rag
- Language: Python
- Homepage:
- Size: 877 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# aguefort
:game_die: 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 protected]
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=
AWS_SECRET_ACCESS_KEY=
AWS_REGION=
```
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.
