MongoShapes

This project is a complete replacement of the GeoJSON field functionality of MongoEngine which exploits the geo_interface protocol to handle BSON serialization and deserialization from MongoDB. This is accomplished by integrating Shapely's corresponding GeoJSON-like geometry types:

• Point
• LineString
• Polygon
• MultiPoint
• MultiLineString
• MultiPolygon
• GeometryCollection

For substituting the MongoEngine field types:

• PointField
• LineStringField
• PolygonField
• MultiPointField
• MultiLineStringField
• MultiPolygonField

And additionally providing:

• GeometryCollectionField

Wherein the field value stored in memory is the corresponding geometry instance rather than a GeoJSON mapping. This provides direct access to all of Shapely's geometry operations and manipulations as well as access to an array_interface protocol for NumPy.

Field validation is overridden with Shapely's implementation of OpenGIS standards. Geo-querying via MongoDB's 2dsphere is also supported (with exception of GeometryCollection) in the usual manner.

Usage

Usage is intended to be consistent with the MongoEngine documentation:

import random
import mongoengine as me
import mongoshapes as ms

from shapely.geometry import box, mapping
from shapely.geometry.base import BaseGeometry


class Doc(me.Document):
    point = ms.PointField()


def bbox(geometry: BaseGeometry) -> ms.PolygonDict:
    buffered = geometry.buffer(5)
    bounds = buffered.bounds
    polygon = box(*bounds)
    mapped = mapping(polygon)
    return mapped


me.connect('mydb')

new = ms.Point(map(lambda x: random.uniform(-x, x), (180, 90)))

doc = Doc()
doc.point = new
doc.validate()
doc.save()

qry = doc.objects(point__geo_within=bbox(new))
assert qry.first() is not None

Testing

Using provided examples from MongoDB's GeoJSON documentation as fixtures, a full py.test suite exists:

• unit for testing GeoJSON fixtures with Shapely and MongoEngine.
• func for testing validation and querying with new field types.