Skip to main content
Help the Python Software Foundation raise $60,000 USD by December 31st!  Building the PSF Q4 Fundraiser

Cython wrapper for the Boost Voronoi library (version 1.59.0)

Project description


A wrapper for Boost's Voronoi diagram library



Cython dependency is optional. Cpp sources generated with Cython are available in releases.

Note on using the operates in 2 modes that are based on the presence of the dev file in the root of the project.

* When dev is **present**, Cython will be used to compile the .pyx sources. This is the development mode (as you get it in the git repository).

* When dev is **absent**, C/C++ compiler will be used to compile the .cpp sources (that were prepared in in the development mode). This is the distribution mode (as you get it on PyPI).

This way the package can be used without or with an incompatible version of Cython.

The idea comes from Matt Shannon's bandmat library.

From PyPI

Cython not required.

``pip install pyvoronoi``

From source

Cython required.

Clone the repository:

``git clone``


``python install``

After every modification of .pyx files compile with Cython:

``python build_ext --inplace``


Create a new instance, passing the scaling factor into the constructor:
import pyvoronoi
pv = pyvoronoi.Pyvoronoi(10)

Since the voronoi library uses integer representation for points, the scaling factor chosen must be high enough
to avoid roundoff error when converting from point coordinates to integers.

Add points and segments:

.. code:: python

pv.AddPoint([0, 0])

Call ``Construct()`` and get the edges and vertices:

.. code:: python

edges = pv.GetEdges()
vertices = pv.GetVertices()
cells = pv.GetCells()

Note that vertices, edges, and cells, can be accessed individually. The methods above are just convenience wrappers around
the following functions:

* GetVertex

* GetEdge

* Get Cell

.. code:: python

def GetVertices(self):
count = self.CountVertices()
output = []
for index in range(count):
return output

.. code:: python

def GetEdges(self):
count = self.CountEdges()
output = []
for index in range(count):
return output

.. code:: python

def GetCells(self):
count = self.CountCells()
output = []
for index in range(count):
return output

If you are running python 2.x, you might want to write your own wrappers using xrange. This will be more efficient.

Edges have the following properties:

* ``start, end`` contain the indices of the start and end vertices or ``-1`` if the edge is infinite at that end.
* ``is_primary`` is true if the edge is not coincident with any of the source inputs.
* ``is_linear`` is true if the edge is linear (not curved).
* ``cell`` is the identifier of the cell this segment is part of.
* ``twin`` is the identifier of the twin segment as defined in the boost voronoi API.

Cells have the following properties:

* ``cell_identifier`` is the index of the cell.
* ``site`` is the index of the site which generated this cell (same as site1, site2 on the edges).
* ``contains_point`` is true if the site was generated by a point.
* ``contains_segment`` is true if the site was generated by a segment.
* ``is_open`` is true if any of the cell's edges is infinite.
* ``is_degenerate`` is true if the cell doesn't have an incident edge. Can happen if a few input segments share a common endpoint.
* ``vertices`` contains indices into the vertex array.
* ``edges`` contains indices into the edge array.

.. code:: python

pv = pyvoronoi.Pyvoronoi(100)

edges = pv.GetEdges()
vertices = pv.GetVertices()
cells = pv.GetCells()
print("Cell Count: {0}".format(len(cells)))
for c in cells:
print("Cell contains point: {0}. Contains segment: {1}. Is open: {2}, Site Index: {3}".format(c.contains_point, c.contains_segment, c.is_open,
for sIndex in c.edges:
print("Start Index: {0}, End Index = {1}".format(edges[sIndex].start, edges[sIndex].end))

Some output edges returned by the boost voronoi API are suposed to be curved. In the C++ API, it is up to you to code it. Luckily, you can do it in python using the following the function DiscretizeCurvedEdge.
The sample below shows you how to do that:

.. code-block:: python

for cIndex in range(len(cells)):
cell = cells[cIndex]
if cell.is_open == False:
for i in range(len(cell.edges)):
e = edges[cell.edges[i]]
startVertex = vertices[e.start]
endVertex = vertices[e.end]

max_distance = distance([startVertex.X, startVertex.Y], [endVertex.X, endVertex.Y]) / 10
if startVertex != -1 and endVertex != -1:
if(e.is_linear == True):
array = [[startVertex.X, startVertex.Y],[endVertex.X, endVertex.Y]]
points = pv.DiscretizeCurvedEdge(cell.edges[i], max_distance)
for p in points:
print "{0},{1}".format(p[0], p[1])

The curve interpolation code can return 2 exceptions.
*FocusOnDirectixException: this happens when the input point is on the segment side. In that cases, it makes no sense to interpolate a parabola between those two geometries since a parabola equation is supposed to find an equidistant point between the two geometries.
*UnsolvableParabolaEquation: there are cases where the point returned by boost does not fit with the parabola equation (for a same position on the x-axis, we get 2 different points, both equidistant). Understanding this issue is still under investigation. It is possible to mitigate this issue by setting an optional 3rd parameter of the function DiscretizeCurvedEdge). A higher value means more tolerance to this exception. The recommended value would be 1 / Scaling Factor.


- Pyvoronoi is available under `MIT
license <>`__.
- The core Voronoi library is available under `Boost Software
License <>`__. Freeware for both
open source and commercial applications.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pyvoronoi, version 1.0.0
Filename, size File type Python version Upload date Hashes
Filename, size pyvoronoi-1.0.0-cp27-none-win32.whl (86.7 kB) File type Wheel Python version 2.7 Upload date Hashes View
Filename, size (154.4 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page