funlib.geometry

A package for Coordinates, Rois and common operations in Coordinate spaces

Installation

pip install funlib.geometry

Usage

Coordinate

An immutable tuple of integers with element-wise arithmetic:

from funlib.geometry import Coordinate

a = Coordinate(1, 2, 3)
b = Coordinate(4, 5, 6)

a + b        # Coordinate(5, 7, 9)
a * 2        # Coordinate(2, 4, 6)
a * b        # Coordinate(4, 10, 18)
abs(-a)      # Coordinate(1, 2, 3)

Use None for unbounded dimensions:

Coordinate(1, None, 3) + Coordinate(4, 5, 6)  # Coordinate(5, None, 9)

Roi

A rectangular region defined by an offset and shape:

from funlib.geometry import Roi, Coordinate

roi = Roi((0, 0, 0), (10, 20, 30))

roi.offset   # Coordinate(0, 0, 0)
roi.shape    # Coordinate(10, 20, 30)
roi.begin    # Coordinate(0, 0, 0)
roi.end      # Coordinate(10, 20, 30)
roi.center   # Coordinate(5, 10, 15)
roi.size     # 6000

Arithmetic shifts the offset (+/-) or scales both offset and shape (*/////):

roi + Coordinate(1, 2, 3)  # Roi((1, 2, 3), (10, 20, 30))
roi * 2                     # Roi((0, 0, 0), (20, 40, 60))

Intersection, union, and containment:

a = Roi((0, 0), (10, 10))
b = Roi((5, 5), (10, 10))

a.intersect(b)   # Roi((5, 5), (5, 5))
a.union(b)       # Roi((0, 0), (15, 15))
a.contains(b)    # False
a.contains((3, 3))  # True

Snap to a grid and convert to numpy slices:

import numpy as np

roi = Roi((3, 7), (10, 10))
roi.snap_to_grid(Coordinate(4, 4), mode="grow")  # Roi((0, 4), (16, 16))

data = np.zeros((100, 100))
region = data[roi.to_slices()]  # equivalent to data[3:13, 7:17]