TileStache.Vector | index |
Provider that returns vector representation of features in a data source.
This is a provider that does not return an image, but rather queries
a data source for raw features and replies with a vector representation
such as GeoJSON. For example, it's possible to retrieve data for
locations of OpenStreetMap points of interest or street centerlines
contained within a tile's boundary.
Many Polymaps (http://polymaps.org) examples use GeoJSON vector data tiles,
which can be effectively created using this provider.
Vector functionality is provided by OGR (http://www.gdal.org/ogr/).
Thank you, Frank Warmerdam.
Currently two serializations and three encodings are supported for a total
of six possible kinds of output with these tile name extensions:
GeoJSON (.geojson):
See http://geojson.org/geojson-spec.html
Arc GeoServices JSON (.arcjson):
See http://www.esri.com/library/whitepapers/pdfs/geoservices-rest-spec.pdf
GeoBSON (.geobson) and Arc GeoServices BSON (.arcbson):
BSON-encoded GeoJSON and Arc JSON, see http://bsonspec.org/#/specification
GeoAMF (.geoamf) and Arc GeoServices AMF (.arcamf):
AMF0-encoded GeoJSON and Arc JSON, see:
http://opensource.adobe.com/wiki/download/attachments/1114283/amf0_spec_121207.pdf
Possible future supported formats might include KML and others. Get in touch
via Github to suggest other formats: http://github.com/migurski/TileStache.
Common parameters:
driver:
String used to identify an OGR driver. Currently, "ESRI Shapefile",
"PostgreSQL", "MySQL", Oracle, Spatialite and "GeoJSON" are supported as
data source drivers, with "postgis" and "shapefile" accepted as synonyms.
Not case-sensitive.
OGR's complete list of potential formats can be found here:
http://www.gdal.org/ogr/ogr_formats.html. Feel free to get in touch via
Github to suggest new formats: http://github.com/migurski/TileStache.
parameters:
Dictionary of parameters for each driver.
PostgreSQL:
"dbname" parameter is required, with name of database.
"host", "user", and "password" are optional connection parameters.
One of "table" or "query" is required, with a table name in the first
case and a complete SQL query in the second.
Shapefile and GeoJSON:
"file" parameter is required, with filesystem path to data file.
properties:
Optional list or dictionary of case-sensitive output property names.
If omitted, all fields from the data source will be included in response.
If a list, treated as a whitelist of field names to include in response.
If a dictionary, treated as a whitelist and re-mapping of field names.
clipped:
Default is true.
Boolean flag for optionally clipping the output geometries to the
bounds of the enclosing tile, or the string value "padded" for clipping
to the bounds of the tile plus 5%. This results in incomplete geometries,
dramatically smaller file sizes, and improves performance and
compatibility with Polymaps (http://polymaps.org).
projected:
Default is false.
Boolean flag for optionally returning geometries in projected rather than
geographic coordinates. Typically this means EPSG:900913 a.k.a. spherical
mercator projection. Stylistically a poor fit for GeoJSON, but useful
when returning Arc GeoServices responses.
precision:
Default is 6.
Optional number of decimal places to use for floating point values.
spacing:
Optional number of tile pixels for spacing geometries in responses. Used
to cut down on the number of returned features by ensuring that only those
features at least this many pixels apart are returned. Order of features
in the data source matters: early features beat out later features.
verbose:
Default is false.
Boolean flag for optionally expanding output with additional whitespace
for readability. Results in larger but more readable GeoJSON responses.
id_property:
Default is None.
Sets the id of the geojson feature to the specified field of the data source.
This can be used, for example, to identify a unique key field for the feature.
Example TileStache provider configuration:
"vector-postgis-points":
{
"provider": {"name": "vector", "driver": "PostgreSQL",
"parameters": {"dbname": "geodata", "user": "geodata",
"table": "planet_osm_point"}}
}
"vector-postgis-lines":
{
"provider": {"name": "vector", "driver": "postgis",
"parameters": {"dbname": "geodata", "user": "geodata",
"table": "planet_osm_line"}}
}
"vector-shapefile-points":
{
"provider": {"name": "vector", "driver": "ESRI Shapefile",
"parameters": {"file": "oakland-uptown-point.shp"},
"properties": ["NAME", "HIGHWAY"]}
}
"vector-shapefile-lines":
{
"provider": {"name": "vector", "driver": "shapefile",
"parameters": {"file": "oakland-uptown-line.shp"},
"properties": {"NAME": "name", "HIGHWAY": "highway"}}
}
"vector-postgis-query":
{
"provider": {"name": "vector", "driver": "PostgreSQL",
"parameters": {"dbname": "geodata", "user": "geodata",
"query": "SELECT osm_id, name, highway, way FROM planet_osm_line WHERE SUBSTR(name, 1, 1) = '1'"}}
}
"vector-sf-streets":
{
"provider": {"name": "vector", "driver": "GeoJSON",
"parameters": {"file": "stclines.json"},
"properties": ["STREETNAME"]}
}
Caveats:
Your data source must have a valid defined projection, or OGR will not know
how to correctly filter and reproject it. Although response tiles are typically
in web (spherical) mercator projection, the actual vector content of responses
is unprojected back to plain WGS84 latitude and longitude.
If you are using PostGIS and spherical mercator a.k.a. SRID 900913,
you can save yourself a world of trouble by using this definition:
http://github.com/straup/postgis-tools/raw/master/spatial_ref_900913-8.3.sql
Package Contents | ||||||
|
Classes | ||||||||||||||||||
|
Data | ||
pyamf_classes = {<class 'TileStache.Vector.Arc._amfFeatureSet'>: 'com.esri.ags.tasks.FeatureSet', <class 'TileStache.Vector.Arc._amfSpatialReference'>: 'com.esri.ags.SpatialReference', <class 'TileStache.Vector.Arc._amfFeature'>: 'com.esri.ags.Feature', <class 'TileStache.Vector.Arc._amfGeometryMapPoint'>: 'com.esri.ags.geometry.MapPoint', <class 'TileStache.Vector.Arc._amfGeometryPolyline'>: 'com.esri.ags.geometry.Polyline', <class 'TileStache.Vector.Arc._amfGeometryPolygon'>: 'com.esri.ags.geometry.Polygon'} |