Ollama RAG

NOTE: Utilized PC with 3 NVIDIA RTX A6000 in a Windows Environment

Prerequisites

Create and activate a Python 3.10 virtual environment

python -m venv [name_of_venv]

.\[name_of_venv]\Scripts\activate

pip install -r requirements.txt

NOTE: If your virtual environment is created using Windows, you must ensure that the Ollama models are pulled within a PowerShell environment.

From PowerShell terminal, download a model from HERE
```
ollama run "name_of_your_model"
```

From WSL terminal, determine your WSL IP address (look under the eth# interface)
IP address listed will be used to host and interact with your chromadb/chroma Docker container
```
ip a
```

Open WSL terminal and run the following command:
```
curl -fsSL https://ollama.com/install.sh | sh
```

From WSL terminal, download a model from HERE
```
ollama run "name_of_your_model"
```

From WSL terminal, determine your WSL IP address (look under the eth# interface)
```
ip a
```

Open VSCode and modify chroma_client.py:
- Replace "YOUR_WSL_IP_GOES_HERE" with your WSL IP.
Modify rag_query.py:
- Replace "YOUR_WSL_IP_GOES_HERE" with your WSL IP.
- Replace "YOUR_OLLAMA_MODEL_GOES_HERE" with your downloaded Ollama model.

Load and Split Documents
- Load PDF docs from the data directory and split each into chunks.
```
python loader.py
```
Initialize Docker Container
- Pull and initiate chromadb/chroma container from Docker
```
sudo docker run -p 8000:8000 chromadb/chroma
```
Create Vector Database
- Initialize the data directory containing your documents to create a vector database
```
python chroma_client.py
```
- Note: Each time you modify the documents in the data directory, re-run chroma_client.py.
Launch Interactive RAG System
```
python rag_query.py
```