This guide walks you through configuring a custom Ollama model and using it to enhance Python code with docstrings and summaries for use in a Retrieval-Augmented Generation (RAG) pipeline.

Before proceeding, ensure the following are installed:

• Ollama
• Python 3.10+
• A virtual environment (via venv or conda)
• Python dependencies from the project's requirements.txt

# Example setup with venv
python -m venv venv
.\venv\Scripts\activate
pip install -r requirements.txt

Download an Ollama-compatible model. After testing several options, we recommend:

ollama pull falcon3:7b

✅ Note: falcon3:7b strikes a good balance of quality and speed for structured code summarization. It has yeilded the best results after several rounds of testing.

1. Navigate to the directory containing your Modelfile.

2. If you choose a different model, update the FROM line in the Modelfile accordingly:

   FROM <your-model-here>

3. Create the custom model:

   ollama create pattern-rag-gen -f Modelfile

This model will specialize in summarizing and annotating code for the design pattern RAG project.