Skip to main content

Geospatial Extensions for Pyramid

Project description

Geospatial Extensions for the Pyramid web framework.

Install

Papyrus can be installed with easy_install:

$ easy_install papyrus

(Installing Papyrus in an isolated virtualenv is recommended.)

Often you’ll want to make Papyrus a dependency of your Pyramid application. For that add papyrus to the install_requires list defined in the Pyramid application setup.py:

install_requires = [
    'pyramid',
    'pyramid_sqla',
    'pyramid_handlers',
    'SQLAlchemy',
    'transaction',
    'repoze.tm2',
    'zope.sqlalchemy',
    'WebError',
    'papyrus'
    ]

Notes:

  • the pyramid_sqla is useful when using SQLAlchemy in the Pyramid application.

  • the pyramid_handlers package is required for creating handlers and actions (instead of view callbables) in your Pyramid application. Handlers basically emulate Pylons’ controllers, so people coming from Pylons may want to use pyramid_handlers in their Pyramid applications.

Run Papyrus Tests

To run the Papyrus tests the nose, mock, and psycopg2 packages must be installed in the Python environment. Also install coverage to be able to get a coverage report when running the tests.

To run the tests and get a coverage report use the following command at the root of the Papyrus tree:

$ nosetests --with-coverage

Currently, 100% of the Papyrus code is covered by tests, I’d like to preserve that.

GeoJSON Renderer

Papyrus provides a GeoJSON renderer, based on Sean Gillies’ geojson package.

To be able to use the GeoJSON renderer the GeoJSON renderer factory must be added to the application configuration.

For that you can either pass the factory to the Configurator constructor:

from pyramid.mako_templating import renderer_factory as mako_renderer_factory
from papyrus.renderers import geojson_renderer_factory
config = Configurator(
    renderers=(('.mako', mako_renderer_factory),
               ('geojson', geojson_renderer_factory))
    )

Or you can apply the add_renderer method to the Configurator instance:

from papyrus.renderers import geojson_renderer_factory
config.add_renderer('geojson', geojson_renderer_factory)

Make sure that add_renderer is called before any add_view call that names geojson as an argument.

To use the GeoJSON renderer in a view set renderer to geojson in the view config. Here is a simple example:

@view_config(renderer='geojson')
def hello_world(request):
    return {
        'type': 'Feature',
        'id': 1,
        'geometry': {'type': 'Point', 'coordinates': [53, -4]},
        'properties': {'title': 'Dict 1'},
        }

Views configured with the geojson renderer must return objects that implement the Python Geo Interface.

Here’s another example where the returned object is an SQLAlchemy (or GeoAlchemy) mapped object:

@view_config(renderer='geojson')
def features(request):
    return Session().query(Spot).all()

In the above example the Spot objects returned by the query call must implement the Python Geo Interface.

MapFish Web Services

Papyrus provides an implementation of the MapFish Protocol. This implementation relies on GeoAlchemy.

This section describes through an example how to build a MapFish web service web (i.e. a web service that conforms to the MapFish Protocol) in a Pyramid application.

So let’s assume we want to create a spots MapFish web service that relies on a spots database table.

Database Model

First of all we need an SQLAlchemy/GeoAlchemy mapping for that table. The pyramid_routesalchemy and pyramid_sqla templates places SQLAlchemy models in a models.py file, we’ll do likewise here.

Here’s what our models.py file looks like:

from sqlalchemy import Column
from sqlalchemy import Integer
from sqlalchemy import Unicode

from sqlalchemy.exc import IntegrityError
from sqlalchemy.ext.declarative import declarative_base

from sqlalchemy.orm import scoped_session
from sqlalchemy.orm import sessionmaker

from zope.sqlalchemy import ZopeTransactionExtension

from geoalchemy import GeometryColumn, Point, WKBSpatialElement

import geojson

from shapely.geometry import asShape
from shapely.wkb import loads

Session = scoped_session(
                sessionmaker(extension=ZopeTransactionExtension())
                )
Base = declarative_base()

class Spot(Base):
    __tablename__ = 'spots'
    id = Column(Integer, primary_key=True)
    name = Column(Unicode, nullable=False)
    geom = GeometryColumn('the_geom', Point(srid=4326))

    def __init__(self, feature):
        self.id = feature.id
        self.__update__(feature)

    def __update__(self, feature):
        geometry = feature.geometry
        if geometry is not None and \
           not isinstance(geometry, geojson.geometry.Default):
            shape = asShape(geometry)
            self.geom = WKBSpatialElement(buffer(shape.wkb), srid=4326)
            self._shape = shape
        self.name = feature.properties.get('name', None)

    @property
    def __geo_interface__(self):
        id = self.id
        if hasattr(self, '_shape') and self._shape is not None:
            geometry = self_shape
        else:
            geometry = loads(str(self.geom.geom_wkb))
        properties = dict(name=self.name)
        return geojson.Feature(id=id, geometry=geometry, properties=properties)

def initialize_sql(engine):
    Session.configure(bind=engine)
    Base.metadata.bind = engine

Note that the Spot class implements the Python Geo Interface (though the __geo_interface__ property), and defines __init__ and __update__ methods.

Implementing the Python Geo Interface is required for being able to serialize Spot objects into GeoJSON (for example using Papyrus’ GeoJSON renderer).

The __init__ and __update__ methods are required for inserting and updating objects, respectively. Both the __init__ and __update__ methods receive a GeoJSON feature (geojson.Feature) as an argument.

The GeoInterface Mixin

Papyrus provides a mixin to help create SQLAlchemy/GeoAlchemy mapped classes that implement the Python Geo Interface, and define __init__ and __update__ as expected by the MapFish protocol. The mixin is named GeoInterface, and is provided by the papyrus.geo_interface module.

Using GeoInterface our Spot class looks like this:

from papyrus.geo_interface import GeoInterface

class Spot(GeoInterface, Base):
    __tablename__ = 'spots'
    id = Column(Integer, primary_key=True)
    name = Column(Unicode, nullable=False)
    geom = GeometryColumn('the_geom', Point(srid=4326))

GeoInterface represents a convenience method. Often, implementing one’s own __geo_interface__, __init__, and __update__ definitions is a better choice than relying on GeoInterface.

When using GeoInterface understanding its code can be useful. It can also be a source of inspiration for those who don’t use it.

Handler

Now that database model is defined we can now create the core of our MapFish web service.

The web service itself can be defined in a handler class, or through view callables, typically functions. This section shows how to define a MapFish web service in a handler class.

Here is what our handler looks like (typically defined in the application’s handlers.py file):

from pyramid_handlers import action

from myproject.models import Session, Spot
from papyrus.protocol import Protocol

# create the protocol object. 'geom' is the name
# of the geometry attribute in the Spot model class
proto = Protocol(Session, Spot, 'geom')

class SpotHandler(object):
    def __init__(self, request):
        self.request = request

    @action(renderer='geojson')
    def read_many(self):
        return proto.read(self.request)

    @action(renderer='geojson')
    def read_one(self):
        id = self.request.matchdict.get('id', None)
        return proto.read(self.request, id=id)

    @action(renderer='string')
    def count(self):
        return proto.count(self.request)

    @action(renderer='geojson')
    def create(self):
        return proto.create(self.request)

    @action(renderer='geojson')
    def update(self):
        id = self.request.matchdict['id']
        return proto.update(self.request, id)

    @action()
    def delete(self):
        id = self.request.matchdict['id']
        return proto.delete(self.request, id)

The six actions of the SpotHandler class entirely define our MapFish web service.

We now need to provide routes to these actions. This is done by calling add_papyrus_handler() on the Configurator. Here’s what the __init__.py file looks like:

from pyramid.config import Configurator
import pyramid_beaker
import pyramid_sqla
from pyramid_sqla.static import add_static_route
import papyrus
from papyrus.renderers import geojson_renderer_factory

def main(global_config, **settings):
    """ This function returns a Pyramid WSGI application.
    """
    config = Configurator(settings=settings)

    # Initialize database
    pyramid_sqla.add_engine(settings, prefix='sqlalchemy.')

    # Configure Beaker sessions
    session_factory = pyramid_beaker.session_factory_from_settings(settings)
    config.set_session_factory(session_factory)

    # Configure renderers
    config.add_renderer('.html', 'pyramid.mako_templating.renderer_factory')
    config.add_renderer('geojson', geojson_renderer_factory)

    config.add_subscriber('myproject.subscribers.add_renderer_globals',
                          'pyramid.events.BeforeRender')

    # Set up routes and views
    config.include(papyrus)
    config.add_papyrus_handler('spots', '/spots',
                               'myproject.handlers:SpotHandler')
    config.add_handler('home', '/', 'myproject.handlers:MainHandler',
                       action='index')
    config.add_handler('main', '/{action}', 'myproject.handlers:MainHandler',
        path_info=r'/(?!favicon\.ico|robots\.txt|w3c)')
    add_static_route(config, 'myproject', 'static', cache_max_age=3600)

    return config.make_wsgi_app()

There are multiple things to note in the above code:

  • the call to add_render for the addition of geojson renderer,

  • the call to include for the inclusion of papyrus (this is what makes add_papyrus_handler available on the Configurator),

  • the call to add_papyrus_handler for the creation of routes to our handler actions.

The add_papyrus_handler method is a convenience. Here’s what does under the hood:

config.add_handler('spots_read_many', '/spots',
                   'myproject.handlers:SpotHandler',
                   action='read_many', request_method='GET')
config.add_handler('spots_read_one', '/spots/{id}',
                   'myproject.handlers:SpotHandler',
                   action='read_one', request_method='GET')
config.add_handler('spots_count', '/spots/count',
                   'myproject.handlers:SpotHandler',
                   action='count', request_method='GET')
config.add_handler('spots_create', '/spots',
                   'myproject.handlers:SpotHandler',
                   action='create', request_method='POST')
config.add_handler('spots_update', '/spots/{id}',
                   'myproject.handlers:SpotHandler',
                   action='update', request_method='PUT')
config.add_handler('spots_delete', '/spots/{id}',
                   'myproject.handlers:SpotHandler',
                   action='delete', request_method='DELETE')

View functions

Using view functions instead of a handler class and actions here’s how our web service implementation looks like:

from myproject.models import Session, Spot
from papyrus.protocol import Protocol

# 'geom' is the name of the mapped class' geometry property
proto = Protocol(Session, Spot, 'geom')

@view_config(route_name='spots_read_many', renderer='geojson')
def read_many(request):
    return proto.read(request)

@view_config(route_name='spots_read_one', renderer='geojson')
def read_one(request):
    id = request.matchdict.get('id', None)
    return proto.read(request, id=id)

@view_config(route_name='spots_count', renderer='string')
def count(request):
    return proto.count(request)

@view_config(route_name='spots_create', renderer='geojson')
def create(request):
    return proto.create(request)

@view_config(route_name='spots_update', renderer='geojson')
def update(request):
    id = request.matchdict['id']
    return proto.update(request, id)

@view_config(route_name='spots_delete')
def delete(request):
    id = request.matchdict['id']
    return proto.delete(request, id)

Again we need to add routes, one route for each view function. This is done by calling add_route on the Configurator:

config.add_route('spots_read_many', '/spots', request_method='GET')
config.add_route('spots_read_one', '/spots/{id}', request_method='GET')
config.add_route('spots_count', '/spots/count', request_method='GET')
config.add_route('spots_create', '/spots', request_method='POST')
config.add_route('spots_update', '/spots/{id}', request_method='PUT')
config.add_route('spots_delete', '/spots/{id}', request_method='DELETE')

Note: if you use view callables as described in this section the pyramid_handlers package isn’t required as an application’s dependency.

TODO

Changes

0.2

  • Add the papyrus.geo_inteface.GeoInterface mixin

  • Add the papyrus.add_papyrus_handler configurator directive

0.1

  • Initial version

Project details


Download files

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

Source Distribution

papyrus-0.2.tar.gz (17.0 kB view details)

Uploaded Source

File details

Details for the file papyrus-0.2.tar.gz.

File metadata

  • Download URL: papyrus-0.2.tar.gz
  • Upload date:
  • Size: 17.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No

File hashes

Hashes for papyrus-0.2.tar.gz
Algorithm Hash digest
SHA256 d743fec794a3580b14a3091493a659d41504221005d136b0c97c34c1e9a80d76
MD5 1c56e6188fe329539d65f7b347170fbf
BLAKE2b-256 f208d12ddeff61d9c1eb7a1d99d6a75bb7b7ace9ba0c2c9f4d3909f102ab4c14

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page