Lakes and rivers sourced from openstreetmap data.

Ensure you've activated the conda environment.

Create the data directories:

mkdir -p data/sources/ && mkdir -p data/temp/ && mkdir -p data/output/

Use the provided osm download utility to download data from the Overpass API. Pass in the bounding box as the first argument, ensuring that the boxes' corners are in the format lat, long. The second argument is the Overpass API query (way[waterway];>;nwr[natural=water];>;), which selects the required relations and ways, plus their nodes.

./../../utils/download_osm.sh \
    "43.022586,-123.417224,45.278084,-118.980589" \
    "(way[waterway];>;nwr[natural=water];>;)"

Now you should have the data at data/sources/extract.osm.pbf.

To convert the openstreetmap data into polygons and lines, run:

python parse_waterways.py

Now you should have water and waterways files at data/temp/water.gpkg and data/temp/waterways.gpkg.

To place labels over water bodies such as lakes and reservoirs, we've built a tool called geopolymid. This tool will create smoothed lines that flow through the centers of each polygon and can be used to place a label.

First, filter the water dataset so only water bodies of the correct type that have names are included:

ogr2ogr \
    -f GPKG \
    data/temp/water-no-rivers.gpkg \
    data/temp/water.gpkg \
    -where "(type != 'river' OR type IS NULL) AND name IS NOT NULL"

See geopolymid for instructions on how to create label lines.

Once complete, name the labels as water-labels.gpkg and place in data/temp/water-labels.gpkg.

Now we can create a tiled version of the water polygons and lines:

./tile_waterways.sh

Now you should have data/output/water.pmtiles, which combines the water polygons, water labels, and waterway lines.