Field geometry and mixture design

Canonical representation

A Field is an arbitrary set of two-dimensional planting coordinates.

field = cm.Field.from_coordinates([
    (0, 0), (1, 0), (2, 0),
    (0.5, 1), (1.5, 1),
])

This representation avoids building the epidemiological model around square matrices.

cm.Field.rectangular(rows=20, columns=40, spacing=1.0)
cm.Field.from_mask(mask, spacing=1.0)
cm.Field.from_polygon(boundary, spacing=1.0)

For measured fields, supply the actual planting coordinates directly.

A design is an assignment vector

\[ z=(z_1,\ldots,z_N), \]

where z_i is the variety planted at site i.

design = cm.MixtureDesign(
    field,
    ("A", "A", "B", "C", "B"),
)

Or generate exact counts randomly:

design = cm.MixtureDesign.random(
    field,
    counts={"A": 40, "B": 35, "C": 25},
    seed=123,
)