Set up a computer to fetch data and build regions locally. District builds run on the build VM (see Dataset operations).
You need GDAL (gdalbuildvrt, gdal_translate) with working LERC decoding,
and Tippecanoe. The pipeline checks both before it fetches. Python packages do
not add codecs to system GDAL; see
LERC dependency check.
Leave plenty of disk space: imagery alone can be several GB per region.
From the project folder:
python3 -m venv .venv
. .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -r requirements.txt
python -c "import geopandas, rasterio, shapely, laspy, pvlib; print('ready')"
The last line should print ready. Run . .venv/bin/activate in each new
terminal; the prompt then starts with (.venv).
The pipeline downloads building outlines, elevation and imagery from the LINZ Data Service, which needs an API key.
.env from the example and add the key:cp .env.example .env
LINZ_API_KEY=your-key
.env sits beside config.py and Git ignores it. Never commit the key or
paste it into issues or chat. The web pipeline’s optional locality grid uses
the LINZ boundary and needs no additional API key or dataset.
VS Code may say terminal environment injection is disabled. Ignore it: the
fetch scripts read .env themselves.