# RAGged Scribe
The Ragged Scribe is a RAG AI bot that will use indexed documents to help
you query and contents.
## Installation
- **Install [Ollama](https://ollama.com)**
- `ollama pull nomic-embed-text`
- `ollama pull llama3` (or the model you prefer)
> modify `OllamaLLM(model="llama3")` accordingly!
- **Start Ollama**
- Put some PDF and/or TXT documents into [static/files](static/files)
> We assume you are working in the `ragged-scribe` directory from here on.
- **Setup python environment**
- Create a virtual environment
- `python -m venv venv`
- `. ./venv/activate`
- Install dependencies
- `pip install -r requirements.txt`
- Index the newly added files:
> It may be required to create a `chroma` directory first!
- `python ./rag_indexer.py --db chroma static/files`
> The indexer will scan for PDF, text, and markdown documents, parse them
> and will add it in batches to the chroma db. Then you are ready!
- **Start the bot backend**
- `python ./rag_interface.py`
> If you need a different host or port, use `--host <host>` and `--port <number>`
> or FLASK_HOST/FLASK_PORT environment variables
- **Query your documents**
- https://localhost:5000/
----
# LICENSE (MIT)
See [LICENSE](LICENSE) for details.
Roughly based on https://github.com/pixegami/rag-tutorial-v2
