dopelines 0.1.0

Douglas-Peucker line simplification

DoPe

Douglas-Peucker line simplification (data reduction).

Reduces the number of points in a dataset, while preserving its most striking features.

The resulting dataset is a subset of the original dataset.

Similar algorithms are used e.g. when zooming on a geographical map. However, this type of algoithm is also useful for general data reduction, as an alternative to conventional filtering.

Currently we only offer a recursive implementation (depth-first), which may not be the most efficient. An iterative implementation may follow (breadth-first).

Installation

Normal installation:

pip install dope

With plot support (adds matplotlib):

pip install dope[plot]

With development tools:

pip install dope[dev]

Example

from dope import DoPeR
import numpy

data_original = [
    [0, 0], [1, -1], [2, 2], [3, 0], [4, 0], [5, -1], [6, 1], [7, 0]
]

dp = DoPeR(data=data_original)

# either use epsilon threshold (i.e. max. error w.r.t. normalized data)
data_simplified_eps = dp.simplify(epsilon=0.2)

# or use maximum recursion depth
data_simplified_depth = dp.simplify(max_depth=4)

# compare original data and simplified data in a plot
dp.plot()

Also see examples in tests.

References:

Douglas DH, Peucker TK. Algorithms for the reduction of the number of points required to represent a digitized line or its caricature. Cartographica: the international journal for geographic information and geovisualization. 1973 Dec 1;10(2):112-22.