Skip to main content

Run your first AI service locally

This walks you through the AgriFoodData Starter Kit — from install to a running service that ingests data, processes it, and writes results back.

Prerequisites​

  • OS — Windows, macOS, or Linux
  • Python 3.10+
  • Docker 20.10+
  • Git (recommended)

Verify:

python --version    # Python 3.10.x or higher
docker --version    # Docker version 20.10.x or higher

1. Install the AgriFoodData SDK​

pip install git+https://github.com/NaLamKI/SDK
pip show nalamkisdk

The SDK repo and the import name will move to agrifooddata as part of the ongoing rebrand — see the Roadmap.

2. Clone the AgriFoodData Starter Kit​

git clone https://github.com/NaLamKI/Starterkit.git
cd Starterkit

3. Set up a virtualenv and install​

python -m venv venv
source venv/bin/activate          # macOS/Linux
# venv\Scripts\activate             # Windows
pip install -r src/requirements.txt

4. Run the example service​

The default starter detects green colours in images — a placeholder for a real vegetation/crop model.

python test/test.py

Inputs are read from test/action/input/, outputs land in test/action/output/.

To visualise the outputs:

python src/visualize_outputs.py

5. What just happened​

  1. The service loads images from test/action/input/.
  2. It processes each image (here: green-pixel detection).
  3. It writes results to test/action/output/.
  4. The visualisation script renders them.

Project structure​

StarterKit/
├── src/                       # Your service code
│   ├── service.py             # Main service class
│   ├── requirements.txt
│   └── visualize_outputs.py
├── test/                      # Local test harness
│   ├── service.py
│   ├── test.py
│   └── action/
│       ├── input/
│       └── output/
└── Dockerfile

Next steps​