Analyze text with LLM

This workshop is based on a collaboration with Babylotse, a program offered by Kinderschutzbund – Bezirksverband Frankfurt am Main e. V.. We extend our sincere thanks to Nicola Küpelikilinc ([email protected]) for her permission to publish and use the data.

License

Code

Data

The survey data are licensed under CC-BY 4.0. Attribution can be done in the form:

🇬🇧: "Family satisfaction in Frankfurt - Babylotse survey 2024" by Babylotse Frankfurt is licensed under CC BY 4.0 and accessible on GitHub.

🇩🇪 "Zufriedenheit von Familien in Frankfurt - Babylotse Umfrage 2024" von Babylotse Frankfurt ist lizensiert unter CC BY 4.0 und verfügbar auf GitHub.

We thank Nicola Küpelikilinc ([email protected]) for her permission to publish and use the data under open licenses.

Setup

1. Install uv
2. Clone this repo git clone https://github.com/CorrelAid/workshop-babylotse.git
3. Open your terminal and navigate to the repository (which I will call root directory)
4. Run uv venv to create a new virtual environment .venv with all the dependencies listed in pyproject.toml. Alternatively, you can also run uv sync, if you are already in a fresh environment (like conda)
5. Now you can use either your IDE or jupyter lab to access the notebooks

Using LLM

To be able to use a LLM, you need to either have one installed locally (see Ollama), or you need an API key.

• Use OpenAI Platform, register for an account and create your API key
• Use Groq Cloud, register for an account and create your API key

I have listed the packages for each API as an optional dependency, thus you have to install the related Python library.

For openai:

uv sync --extra openai

For Groq:

uv sync --extra groq

Now you have to put your API key into the .env file:

1. Copy .env.default and rename it to .env
2. Fill in your API key for either OpenAI or Groq

You are ready to start with the notebooks!