An interface to the EPSG Geodetic Parameter Dataset at

pip install python-epsg==0.1.5


Python EPSG

Build Status


This package provides an API for accessing the data in the EPSG registry. The epsg.schema module provides an object model that closely maps to the GML available as an export from the online registry.

Note that this package does not provide any functionality for performing reprojections or coordinate transformations: its sole purpose is to act as an API to access the data available at the EPSG registry.

The object model builds on SQLAlchemy to provide persistence and querying of the object model from within a SQL database.


The epsg.Registry class represents a local database copy of the online EPSG registry. The default is an in-memory sqlite database if no other database engine is passed in using the engine constructor argument. e.g.

>>> from epsg import Registry
>>> registry = Registry()   # use in-memory database

This can take a while as data is retrieved from the online EPSG registry at

epsg.Registry implements the Python MutableMapping interface. Keys represent EPSG identifiers and the values are the objects themselves:

>>> epsg4326 = registry['urn:ogc:def:crs:EPSG::4326']
>>> print epsg4326
<GeodeticCRS('urn:ogc:def:crs:EPSG::4326','WGS 84')>

These objects can be introspected to provide access to the EPSG information:

u'WGS 84'
>>> epsg4326.geodeticDatum.realizationEpoch, 1, 1)

The object model is defined in epsg.schema but closely mirrors the EPSG GML format. The GML can be obtained from the online EPSG registry as follows:

>>> from epsg.service import Service
>>> service = Service()
>>> service.connect() # open an HTTP connection to the online registry
>>> gml = service.export() # get the GML as a string

The following classes compose the object model:

>>> set((type(v) for v in registry.itervalues()))
set([<class 'epsg.schema.PrimeMeridian'>,
     <class 'epsg.schema.AreaOfUse'>,
     <class 'epsg.schema.Ellipsoid'>,
     <class 'epsg.schema.GeodeticDatum'>,
     <class 'epsg.schema.VerticalDatum'>,
     <class 'epsg.schema.EngineeringDatum'>,
     <class 'epsg.schema.GeodeticCRS'>,
     <class 'epsg.schema.EllipsoidalCS'>,
     <class 'epsg.schema.CartesianCS'>,
     <class 'epsg.schema.VerticalCS'>,
     <class 'epsg.schema.SphericalCS'>,
     <class 'epsg.schema.EngineeringCRS'>,
     <class 'epsg.schema.CoordinateSystemAxis'>,
     <class 'epsg.schema.AxisName'>,
     <class 'epsg.schema.ProjectedCRS'>,
     <class 'epsg.schema.VerticalCRS'>,
     <class 'epsg.schema.CompoundCRS'>])

Changes to the instances are persisted in the registry (and its underlying database):

>>> name = 'World Geodetic System 1984'
>>> = name
>>> del epsg4326
>>> assert registry['urn:ogc:def:crs:EPSG::4326'] == name

Querying the registry

Complex registry queries can be performed by using the SQLAlchemy API, based on objects in the schema module. This is done using the Repository.session property which is a sqlalchemy.orm.Session instance.

  • Obtain all Ellipsoid objects containing the case insensitive substring airy:
>>> from epsg import schema
>>> registry.session.query(schema.Ellipsoid).filter('%airy%')).all()
  • Get a particular projected coordinate reference system:
>>> registry.session.query(schema.ProjectedCRS).filter_by(identifier = 'urn:ogc:def:crs:EPSG::6594').first()
  • Find out how many coordinate reference systems are contained within the longitudes of -76 and -75:
>>> registry.session.query(schema.ProjectedCRS).join(schema.ProjectedCRS.domainOfValidity).filter(schema.AreaOfUse.eastBoundLongitude.between(-76,-75), schema.AreaOfUse.westBoundLongitude.between(-76,-75)).count()

See querying in SQLAlchemy for further details.

Loading registry data

Registries can be initialised with specific data by using specific Loader instances. Registry.getLoader provides a shortcut for creating a loader from the latest data in the online registry: the following statements are equivalent:

# using the default loader upon initialisation
>>> registry2 = Registry()

# using `getloader` with the constructor
>>> loader = registry.getLoader()
>>> registry2 = Registry(loader=loader)

The init method can be used to completely re-create and re-populate a registry database:

>>> registry2.init() # use the default loader
>>> registry2.init(loader)  # specify a loader
>>> registry2.init(loader=False) # re-create but don't populate

Loaders can be created from XML files...

>>> from epsg.load import XML, XMLLoader
>>> xml = XML.FromFile('./tests/test.xml')
>>> loader = XMLLoader(xml)
>>> loader.load() # create the objects from the XML

...or from XML strings...

>>> xml = XML.FromString(gml)
>>> loader = XMLLoader(xml)
>>> loader.load()

...which is equivalent to:

>>> loader = registry.getLoader(gml)

Updating registries

Registry objects implement the MutableMapping interface which means they can be updated from other dictionary like objects that contain appropriate epsg.schema instances. Registry objects themselves provide the correct interface...

>>> registry2 = Registry(loader=false) # create an empty registry
>>> registry2.update(registry) # copy the registry
>>> assert len(registry2) == len(registry) # they are the same do Loader objects:

>>> registry2.update(loader)

Copying registries

Copying registries is simply a case of initialising a registry with another registry or loader:

>>> registry2 = Registry(loader=registry)
>>> registry2 = Registry(loader=loader)

Persisting registries

For efficiency reasons an application will most likely not want to obtain its data from the online EPSG registry every time it needs to access the data. The solution is to use a SQLAlchemy database engine attached to a local, persistent database. The local database acts as a cache which can be updated as required:

>>> from sqlalchemy import create_engine
>>> engine = create_engine('sqlite:///./epsg-registry.sqlite')
>>> registry = Registry(engine)
>>> registry.init() # refresh as required



From PyPI

pip install python-epsg

From source

Download the source from either or, then run the following from the root distribution directory:

python install

It is recommended that you also run:

python test

This exercises the comprehensive package test suite. Note that the tests require an internet connection to access the EPSG registry web service.


  • This is a new and immature package: please treat it as beta quality software and report any bugs in the github issue tracker.

  • The implementation of the GML object model in python is incomplete. See the file for a list of GML elements that are currently not present within the python object model.