Python geometry format conversion cheat-sheet

If your GIS system is based on the Python/Django/DRF/PostGIS stack, it is very common to perform geometry format conversions between its different layers. This short article includes a cheat-sheet of the most common conversions needed.

Formats

• GeoDjango Geometry Field: this is how Django exposes its geometry model fields; it is basically a thin wrapper over the C/C++ GEOS library's objects.
• Shapely Geometry: Shapely is a feature-rich Python library used for in-memory geometry manipulation. It also uses GEOS underneath, but operates in a higher abstraction layer.
• GeoJSON: this is a JSON-based format described by the RFC-7946, and it is very popular for payloads on web APIs.
• GeoJSON-like Python dict: same as GeoJSON, but parsed as a Python dict.
• WKT stands for “Well-Known Text”, it is a very simple text representation. The same specification also defines the WKB counterpart (“Well-Known Binary”), for direct binary transfer.

Conversions

Required imports:

import json
from shapely import from_wkb, from_wkt, from_geojson, to_geojson
from shapely.geometry import mapping, shape
from django.contrib.gis.geos import GEOSGeometry

Examples used in the table:

• django_field: instance of django.contrib.gis.geos.polygon.Polygon
• shapely_geo: instance of shapely.geometry.polygon.Polygon
• geojson_str: instance of str
• geojson_dict: instance of dict
• wkt_str: instance of str

Snippets:

From	To	Snippet
GeoDjango	Shapely	from_wkb(django_field.wkb.tobytes())
GeoDjango	GeoJSON string	django_field.geojson
GeoDjango	GeoJSON dict	json.loads(django_field.geojson)
GeoDjango	WKT	django_field.wkt
Shapely	GeoDjango	GEOSGeometry(memoryview(shapely_geo.wkb))
Shapely	GeoJSON string	to_geojson(shapely_geo)
Shapely	GeoJSON dict	mapping(shapely_geo)
Shapely	WKT	shapely_geo.wkt
GeoJSON string	GeoDjango	GEOSGeometry(geojson_str)
GeoJSON string	Shapely	from_geojson(geojson_str)
GeoJSON string	GeoJSON dict	json.loads(geojson_str)
GeoJSON string	WKT	from_geojson(geojson_str).wkt
GeoJSON dict	GeoDjango	GEOSGeometry(json.dumps(geojson_dict))
GeoJSON dict	Shapely	shape(geojson_dict)
GeoJSON dict	GeoJSON string	json.dumps(geojson_dict)
GeoJSON dict	WKT	shape(geojson_dict).wkt
WKT	GeoDjango	GEOSGeometry(wkt_str)
WKT	Shapely	from_wkt(wkt_str)
WKT	GeoJSON string	to_geojson(from_wkt(wkt_str))
WKT	GeoJSON dict	mapping(from_wkt(wkt_str))

Please notice: the snippets listed above are the most performant conversion paths. You may find alternative snippets somewhere else, but those listed here, especially the two directions between GeoDjango and Shapely, are the fastest versions possible.