A DataSource is a backend data datasource. These data datasources provide 
persistent storage for features.

OGR
===
The OGR datasource allows you to take any OGR datasource -- such as a shapefile,
PostGIS database, GML file, or other formats supported by OGR -- and use it
as a backend for a FeatureServer layer.

The OGR layer requires a 'dsn' property and a 'layer' property. In many cases, 
the layer is part of the DSN. An example OGR configuration might look like:

::

    [myshape]
    type=OGR
    dsn=/home/example/myshape.shp
    layer=myshape

Dependancies:
 * OGR
 * ogr Python bindings

DBM
===
The DBM datasource uses anydbm combined with pickle to store features in a file
on disk. It should work on any platform. A sample DBM configuration might
look like:

::

    [scribble]
    type=DBM
    file=/home/example/scribble.db

Depending on the system on which it is running, the internals of this database
may be different. 

Dependancies:
 * None

DBM Extra Features:
 * Allows for non-fixed attribute schemas: every feature can have a different
   set of attributes.
 * Allows for the setting of a 'unique' property on the layer. If unique is 
   set, features will be 'merged' based on this key on creation, instead of 
   a new feature being created. 

BerkleyDB
=========
The BerkleyDB datasource is a subclass of the DBM datasource. It uses a 
BerkleyDB module for storage. The bsddb module is not available under OS X, 
however, this datasource should work on both Windows and Linux. It is 
recommended to use  this datasource instead of the DBM datasource when 
possible. 

Performance testing suggests that this datasource is approximately 25% faster
than the default anydbm module on Linux.

A sample BerkleyDB configuration might look like: 

::

    [scribble]
    type=BerkleyDB
    file=/home/example/scribble.db

Dependancies:
 * working bsddb module

PostGIS
=======
The PostGIS datasource implements a direct connection to PostGIS. This 
datasource requires dsn parameter and a layer parameter. It also accepts a 
variety of optional parameters, some of which you may need to override the
default value of, depending on your schema.  Please see the example below
for details.

A sample PostGIS configuration might look like:

::

    [mylayer]
    type=PostGIS
    dsn=host=localhost dbname=mydata user=myuser password=1234
    layer=mylayer
    fid=gid #defaults to ogc_fid
    geometry=wkb_geometry # defaults to the_geom
    srid=26910 #defaults to 4326
    attribute_cols=name,some_interesting_column #optional
    order=cost #optional

Dependancies:
 * psycopg or psycopg2

SQLite
======
A simple sqlite datasource.

Dependancies:
 * None under Python2.5
 * pysqlite under Python2.4

WFS
===
The WFS datasource implements read-only access to WFS servers. Required 
parameters are:
 * url
 * typename

A sample WFS layer configuration might look like:

::

    [wfs]
    type=WFS
    url=http://zuluviz.sdsu.edu:8080/geoserver/wfs
    typename=topp:states

Dependancies:
 * OGR
 * ogr Python bindings


Flickr
======
The Flickr datasource loads data from geotagged Flickr photos. Filtering
the results can be done by using any of the parameters listed at:

http://flickr.com/services/api/flickr.photos.search.html

For example, a flickr source limited to all photos which match all the 
tags 'mit', 'food', '2007' from the user with ID '56541240@N00'::

::

    [flickr]
    type=Flickr
    user_id=56541240@N00
    tags=mit,food,2007
    tag_mode=all

By default, this only shows public photos. To grant the datasource access
to your non-public photos, you must fetch a 'read' permissions token from
the API. You can use the Flickr.py as a command line application to do this:

$ ./FeatureServer/DataSource/Flickr.py
Open the following URL in a browser:
http://flickr.com/services/auth/?api_key=[...]&perms=read&frob=[...]
Press enter when complete.

The result of this will be something like:

Your token is:

    721576-5d6044

Add the following line to a Flickr data source to complete configuration:

auth_token=721576-5d6044  

With this example, a Flickr datasource would look like:

::

    [flickr]
    type=Flickr
    user_id=56541240@N00
    auth_token=721576-5d6044

Twitter
=======
Powered by Twittervision, the twitter datasource lets you use twittervision's
API to display the current location of a user.

Very simplistic, doesn't do much.

::

    [twitter-crschmidt]
    type=Twitter
    username=crschmidt

OSM
===
Allows for requests to hit the OpenStreetMap API to request data. No
configuration neccesary. Only supports query-by-bounding-box and query-by-id:
the latter will only search for 'ways', not nodes or relations. Relations
are not handled at all.

It is possible to add an 'osmxapi=yes' option. This will use the OSM Extended
API instead of the main API. This API may be several hours out of date, but 
it does allow attribute queries: see 
http://wiki.openstreetmap.org/index.php/Osmxapi . It does no support 
ID queries, and if one is sent, it will instead be redirected to the main
API.

:: 
  
   [osm]
   type=OSM
   osmxapi=no