picogeojson 0.4.4

Minimal GeoJSON parser and emitter

Straightforward and compliant GeoJSON parsing and serialization with zero dependencies. Easily ingest or output GeoJSON adhering to RFC 7946.

Usage

GeoJSON files or strings are read using fromfile() or fromstring() (alias loads()).

pt = picogeojson.fromstring('{"type": "Point", "coordinates": [1.0, 3.0]}')
# -> Point(coordinates=[1.0, 3.0])

Sometimes a particular type of GeoJSON object is expected (e.g. from an API GET request), but for safety the type needs to be checked. Alternatively, the result_fromstring() function can be used, which returns an object with safe accessors for specific GeoJSON types.

result = picogeojson.result_fromstring(api_response.decode("utf-8"))

# Expecting one or more points or multipoints
for geom in result.points:
    # do something with points
    # ...

for geom in result.multilinestrings:
    # do something with multilinestrings
    # ...

This works for Features too, and we can filter by the .properties member.

result = picogeojson.result_fromstring(api_response.decode("utf-8"))

for feature in result.features("Polygon", {"type": "Lake", "state": "Oregon"}):
    # do something with lakes in Oregon
    # ...

GeoJSON objects may be constructed in Python and composed (merge()) or split (burst()).

points = [picogeojson.Point((1, 2)),
          picogeojson.Point((3, 4)),
          picogeojson.Point((5, 6))]

merged_points = picogeojson.merge(points)
# -> MultiPoint(coordinates=[(1, 2), (3, 4), (5, 6)])

split_points = picogeojson.burst(merged_points)
# -> [Point((1, 2)), Point((3, 4)), Point((5, 6))]

GeoJSON objects are serialized with tostring() (alias dumps()).

picogeojson.tostring(
    picogeojson.Point([1.0, 3.0])
)
# -> {"coordinates": [1.0, 3.0], "type": "Point"}'

Keyword arguments can be passed to tostring() that - enforce Polygon/MultiPolygon rotation direction, with counterclockwise for external rings and clockwise for internal rings (enforce_poly_winding) - split objects that cross the international dateline into multipart objects, for easier processing (antimeridian_cutting) - control whether a bbox member is computed and added to the JSON output (write_bbox)

picogeojson will leverage ujson as a backend if it is installed. Otherwise, it uses Python’s built-in json module.

Performance

The read benchmark involves reading a list of earthquake features. The write benchmark involves serializing the continent of Australia.

Module	Read	Write
json	1.49	2.00
geojson	6.74	same
picogeojson	1.84	same*
picogeojson+ujson	1.63	0.31*

*antimeridian cutting and polygon winding check set to False

This is a standalone Python package extracted from the Karta geojson submodule.