Django speaking WFS 2.0 (exposing GeoDjango model fields)
Project description
django-gisserver
Django speaking WFS 2.0 to expose geo data.
Features
- WFS 2.0 simple implementation.
- GML 3.2 output.
- Standard and spatial filtering (FES 2.0)
- GeoJSON and CSV export formats.
- Extensible view/operations.
- Uses GeoDjango queries for filtering.
- Streaming responses for large datasets.
Usage
Create a model that exposes a GeoDjango field:
from django.contrib.gis.db.models import PointField
from django.db import models
class Restaurant(models.Model):
name = models.CharField(max_length=200)
location = PointField(null=True)
def __str__(self):
return self.name
Write a view that exposes this model as a WFS feature:
from gisserver.features import FeatureType, ServiceDescription
from gisserver.geometries import CRS, WGS84
from gisserver.views import WFSView
from .models import Restaurant
RD_NEW = CRS.from_srid(28992)
class PlacesWFSView(WFSView):
"""An simple view that uses the WFSView against our test model."""
xml_namespace = "http://example.org/gisserver"
# The service metadata
service_description = ServiceDescription(
title="Places",
abstract="Unittesting",
keywords=["django-gisserver"],
provider_name="Django",
provider_site="https://www.example.com/",
contact_person="django-gisserver",
)
# Each Django model is listed here as a feature.
feature_types = [
FeatureType(
Restaurant.objects.all(),
fields="__all__",
other_crs=[RD_NEW]
),
]
Use that view in the URLConf:
from django.urls import path
from . import views
urlpatterns = [
path("/wfs/places/", views.PlacesWFSView.as_view()),
]
You can now use http://localhost:8000/wfs/places/ in your GIS application. It will perform requests such as:
- http://localhost:8000/wfs/places/?SERVICE=WFS&REQUEST=GetCapabilities&ACCEPTVERSIONS=2.0.0,1.1.0,1.0.0
- http://localhost:8000/wfs/places/?SERVICE=WFS&REQUEST=DescribeFeatureType&VERSION=2.0.0&TYPENAMES=restaurant
- http://localhost:8000/wfs/places/?SERVICE=WFS&REQUEST=GetFeature&VERSION=2.0.0&TYPENAMES=restaurant&STARTINDEX=0&COUNT=1000&SRSNAME=urn:ogc:def:crs:EPSG::28992
By adding &OUTPUTFORMAT=geojson
or &OUTPUTFORMAT=csv
to the GetFeature
request, the GeoJSON and CSV outputs are returned.
The CSV output has an unlimited page size, as it's quite performant.
Configuring fields
By default, only the geometry field is exposed as WFS attribute.
Use the FeatureType(..., fields=[...])
parameter to define which fields should be exposed.
It's possible to expose foreign key relations too as complex fields:
from gisserver.features import FeatureType, field
class CustomWFSView(WFSView):
...
feature_types = [
FeatureType(
Restaurant.objects.all(),
fields=[
"id",
"name",
"location",
field("owner", fields=["id", "name", "phonenumber"])
"created"
],
),
]
Customization hooks
There are a few places that allow to customize the WFS logic:
- Overriding methods from
WFSView
, such asget_feature_types()
ordispatch()
for authorization. - Subclassing the
FeatureType
class, to override it'sget_queryset()
method. - Subclassing the
FeatureField
class for theFeatureType(fields=...)
parameter. - Registering new FES functions to
gisserver.parsers.fes_function_registry
. - Registering new stored procedures to
gisserver.queries.stored_query_registry
.
Debugging questions
- The error "Operation on mixed SRID geometries" often indicates the table SRID differs
from the
GeometryField(srid=..)
configuration in Django.
Standards compliance
Nearly all operations for the WFS simple conformance class are implemented. You should be able to view the WFS server QGis. The unit tests validate the output against WFS 2.0 XSD schema.
Some remaining parts for the "WFS simple" conformance level are not implemented yet:
- KVP filters:
propertyName
,aliases
,namespaces
. - Remote resolving:
resolve
,resolveDepth
,resolveTimeout
. - Some
GetCapabilities
features:acceptFormats
andsections
. - Temporal filtering (high on todo)
Low-prio items:
Anything outside WFS simple could be implemented, but is very low on the todo-list:
- The methods for the WFS transactional, locking and inheritance conformance classes.
- HTTP POST requests.
- SOAP requests.
- Other protocols (WMS, WMTS, WCS)
- Other output formats (shapefile, KML, GML 3.1) - but easy to add.
Development
When you follow the source of the WFSView
, WFSMethod
and Parameter
classes,
you'll find that it's written with extensibility in mind. Extra parameters and operations
can easily be added there. You could even do that within your own projects and implementations.
A lot of the internal classes and object names are direct copies from the WFS spec. By following these type definitions, a lot of the logic and code structure follows naturally.
The Makefile
gives you all you need to start working on the project.
Typing make
gives an overview of all possible shortcut commands.
The WFS specification and examples be found at:
- https://www.opengeospatial.org/standards/ (all OGC standards)
- https://docs.opengeospatial.org/ (HTML versions)
Some deeplinks:
- https://www.opengeospatial.org/standards/common (OGC Web Service Common)
- https://www.opengeospatial.org/standards/wfs#downloads (OGC WFS)
- https://portal.opengeospatial.org/files/09-025r2 (WFS 2.0 spec, PDF)
- https://portal.opengeospatial.org/files/09-026r1 (OpenGIS Filter Encoding 2.0, PDF)
- https://portal.opengeospatial.org/files/07-036 (GML 3.2.1)
Other links:
- http://schemas.opengis.net/wfs/2.0/ (XSD and examples)
- https://cite.opengeospatial.org/teamengine/ (test suite)
- https://mapserver.org/development/rfc/ms-rfc-105.html (more examples)
Why this code is shared
The "datapunt" team of the Municipality of Amsterdam develops software for the municipality. Much of this software is then published as Open Source so that other municipalities, organizations and citizens can use the software as a basis and inspiration to develop similar software themselves. The Municipality of Amsterdam considers it important that software developed with public money is also publicly available.
This package is initially developed by the City of Amsterdam, but the tools and concepts created in this project can be used in any city.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for django_gisserver-0.8.1-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8e27d1398e73ce68091cbf2de960ad5fb25191c2b7c4ca060780049e8fa3a27f |
|
MD5 | 055062db3696f20eebd1cfaa5d9e48d4 |
|
BLAKE2b-256 | a37a395c8c0acabd1ad907f14adb32ce478f404b7a75fce3c349f64393112965 |