Skip to main content

GIS Tools

Project description


This package aims to bring you close to a ready-usable area search for existing “addresses” within a given distance from a starting point.

This was developed for germany, where any place has a zip-code. We wanted to have a search functionality to find objects that are “close” (i.e. within a defined distance) to a given location.

You could also use any other key which points to coordinates with little effort. The only requirement is that the key is unique.

For this, zip codes are the best available solution for us. We have a free list of zip codes and corresponding gis information for germany (see data directory). This list is from the opengeodb project and licensed as public domain:


First, let’s load the required zcml as this is a unit test.
>>> from Products.Five.zcml import load_config, load_string
>>> import d9t.gis
>>> import Products.Five
>>> import five.localsitemanager
>>> load_config('configure.zcml', package=Products.Five)
>>> load_config('configure.zcml', package=five.localsitemanager)
>>> load_config('configure.zcml', package=d9t.gis)


Measurable Objects

A measurable object has to provide ICoordinate, which also may be provided by an adapter.

Let’s define a generic address class (minimalistic and incomplete, i know).

>>> from zope.interface import implements, Interface
>>> from zope.component import queryUtility, adapts
>>> from d9t.gis.interfaces import ICoordinate, ICoordinateProvider
>>> class Address(object):
...     implements(ICoordinate)
...     def __init__(self, country, zip):
...         zip_coordinate_provider = queryUtility(ICoordinateProvider)
...         coordinate = zip_coordinate_provider.coordinate(country, zip)
...         self.latitude = coordinate.latitude
...         self.longitude = coordinate.longitude

Now let’s generate a few addresses.

>>> a1 = Address("DE", "89073")
>>> a2 = Address("DE", "88299")

You can also create simple Coordinates. A demo Coordinate class is included:

>>> from d9t.gis.coordinate import Coordinate
>>> c1 = Coordinate(9.99968113200213, 48.4052825255534)   # DE 89073
>>> c2 = Coordinate(10.0191253966503, 47.8117109627977)   # DE 88299

It should be noted that ZipCoordinateProvider loads the gis data from csv on zope startup. This takes far less then a second and we decided that storing the data in zodb wouldn’t have any advantage.

Calculating Distance

>>> from d9t.gis.interfaces import IDistanceCalculation
>>> distance_util = queryUtility(IDistanceCalculation)

You can measure distance between anything that provides ICoordinate. So: Between Coordinates

>>> distance_util.distance(c1, c2)

Between Addresses

>>> distance_util.distance(a1, a2)

Between Coordinates and Addresses

>>> distance_util.distance(a1, c2)
>>> distance_util.distance(a1, c1)

Common usage

Probably you already have some address objects which are unaware of gis information. Maybe you want - and this is indeed a good idea - to prevent your address - objects to even know about gis information.

You can have all gis features by providing an ICoordinate Adapter for your objects.

Imagine we have a already existing class which is unaware of any gis info:

>>> class IMyAddress(Interface):
...     """ """
>>> class MyAddress(object):
...     implements(IMyAddress)
...     address = ""
...     zip_code = ""
...     city = ""
...     country = ""
...     def __init__(self, address, zip_code, city, country):
...         self.address, self.zip_code,, = address, zip_code, city, country

You would then have objects of that type:

>>> my_a1 = MyAddress("Ilextwiete 12", "22455", "Hamburg", "DE")
>>> my_a2 = MyAddress("Gangweg 2", "80797", "Muenchen", "DE")

For measuring distance, your objects have to provide ICoordinate. So let’s create an adapter which simply uses a ready-to-use utility to get the coordinates.

>>> class MyAddressCoordinate(object):
...     implements(ICoordinate)
...     adapts(IMyAddress)
...     def __init__(self, my_address):
...         self.my_address = my_address
...         zip_coordinate_provider = queryUtility(ICoordinateProvider)
...         coordinate = zip_coordinate_provider.coordinate(, self.my_address.zip_code)
...         self.latitude = coordinate.latitude
...         self.longitude = coordinate.longitude

You would usually provide the adapter in your zcml, but as this is a testcase, we’ll do it here:

>>> from import ztapi
>>> ztapi.provideAdapter(IMyAddress, ICoordinate, MyAddressCoordinate)

Then you can measure as usual:

>>> distance_util.distance(my_a1, my_a2)

Get nearby places

If you want to know which of several addresses are the closest to a given one, just give the utility a list of known ICoordinate and one to look for. You will get back a list of tuples where [0] is the distance and [1] is the original object. Btw: As you will see any adaptable object will be ok and returned as-is.

Let’s measure what are the nearest 3 to Muenchen (my_a2) from this list:

>>> nearest = distance_util.nearest(my_a2, (my_a1, my_a2, c1, c2, a1, a2))
>>> ["%s (%s)" % (n[0], n[1].__class__) for n in nearest]
["0.0 (<class 'MyAddress'>)", "175.137634687 (<class 'Address'>)", "175.137634687 (<class 'd9t.gis.coordinate.Coordinate'>)", "175.24792832 (<class 'Address'>)", "175.24792832 (<class 'd9t.gis.coordinate.Coordinate'>)", "624.795549599 (<class 'MyAddress'>)"]

You can also limit the search to e.g. 3 results (sorted of course):

>>> nearest = distance_util.nearest(my_a2, (my_a1, my_a2, c1, c2, a1, a2), 3)
>>> ["%s (%s)" % (n[0], n[1].__class__) for n in nearest]
["0.0 (<class 'MyAddress'>)", "175.137634687 (<class 'Address'>)", "175.137634687 (<class 'd9t.gis.coordinate.Coordinate'>)"]

Get nearby ZIPs

In case you need all zips within a given distance around a given coordinate, you might find the INearbyZips utility useful.

>>> from zope.component import getUtility
>>> from d9t.gis.interfaces import INearbyZips, IDistanceCalculation
>>> nbz = getUtility(INearbyZips)
>>> distance_util = getUtility(IDistanceCalculation, name="km")
>>> nbz.nearbyZips(c1, distance_util.toRadiant(10))
set([('DE', '89077'), ('DE', '89231'), ('DE', '89075'), ('DE', '89073')])

This was for 10km. Attention! This only works away from radiant bounderies. Stay away from +-180 degrees! This is due to speed optimizations. Sorry ;)


Nearby places with portal_catalog

When using portal_catalog, you only get brains back which have no usable interface to adapt to. Then, please don’t getObject() anything. It’s a waste.

Instead, create a decorator for the brains like that:

>>> class MyAddressBrainCoordinateDecorator(object):
...     implements(ICoordinate)
...     def __init__(self, brain):
...         self.brain = brain
...         zip_coordinate_provider = queryUtility(ICoordinationProvider)
...         coordinate = zip_coordinate_provider.coordinate(brain.getCountry, brain.getZip)
...         self.latitude = coordinate.latitude
...         self.longitude = coordinate.longitude

Then decorate your brains before using them for nearest search and get them back after the search:

>>> brains = []
>>> decorated_brains = [MyAddressBrainCoordinateDecorator(brain) for brain in brains]
>>> nearest = distance_util.nearest(my_a2, decorated_brains, 5)
>>> brains = [decorated_brain.brain for decorated_brain in decorated_brains]

Too bad, that way the laziness of the portal_catalog search is gone. But with a result set of less than 100 that shouldn’t really matter. If your set is big enough for performance impacts, any ideas are welcome.

Have fun ;)


d9t.gis - 0.4 [20120525]

  • Fixed a bug where a rounding error (caused by float) crashed with ValueError: math domain error when the lookup-coordinate were identical to one in the list, i.e. when the distance should have been 0.

d9t.gis - 0.3 [20081217]

  • Added nearby Zips utility (FAST, no. DAMN FAST!) [Daniel Kraft, Oliver Roch]
  • Added named distance utility for miles and km [Daniel Kraft, Oliver Roch]
  • Made zip database pluggable. You may now code your sql implementation. [Daniel Kraft, Oliver Roch]

d9t.gis - 0.2

  • Fully functional and complete doctest available. [Daniel Kraft]

d9t.gis - 0.1 Unreleased

  • Initial package structure. [zopeskel]

Project details

Download files

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

Filename, size & hash SHA256 hash help File type Python version Upload date
d9t.gis-0.4.tar.gz (418.4 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page