Python library for working with spatial databases (and services) and Who's On First documents.
sudo pip install -r requirements.txt .In an effort to make using this library in as broad a fashion as possible specific dependencies for the PostGIS "interfaces" have been excluded from the package's requirements.txt file because they depend on a whole bunch of stuff that involves extra work to install. So you will need to do that yourself in order to get stuff to work:
sudo pip install psycopg2 shapelyIt's not ideal and if there's a way to add these as "try to install these but don't freak out if it fails" dependencies to the requirements.txt file we'd love to hear about it.
First of all, we're using the term "interface" loosely here. I don't think Python has interfaces so what we're really talking about are the methods defined in mapzen.whosonfirst.spatial.base class. This class is then subclassed by "client" (discussed below) who override them, since they all raise an exception by default. Ideally each client would implement every method but in reality they won't for a variety of reasons.
Perform a point in polygon query, with 0 or more optional filters.
Return all the things that intersect with feature, with 0 or more optional filters.
Return all the things that intersect with feature, with 0 or more optional filters. Take care of paginating the results. Pagination details are left to the discretion of individual clients.
Convert a client's internal representation in to a proper GeoJSON Feature thingy. Ideally it would be the actual Who's On First record but that's still a bit squishy as of this wriging.
Index feature in the client's data store. Practically speaking this is still pretty specific to the PostGIS client right now but that may change.
We index and query spatial data in a variety of ways both to keep us honest and because different circumstances have different constraints and burdens. Ideally we'd like to arrive at a place where every index would store (or at least return) the same data, at least in so far as consumers are concerned. We're not quite there yet.
This implements all of "base" interface assuming a PostGIS database. All of this code is interwoven with code in the go-whosonfirst-pgis package which we use for indexing (because it's faster than doing in Python). As of this writing that package is where the database schema is defined.
Currently this implements the point_in_polygon and row_to_feature interfaces using the Who's On First API.
Currently this implements the point_in_polygon and row_to_feature interfaces using the Who's On First PIP server.
import mapzen.whosonfirst.spatial.whosonfirst
import mapzen.whosonfirst.spatial.postgres
pip_client = mapzen.whosonfirst.spatial.whosonfirst.pip()
pg_client = mapzen.whosonfirst.spatial.postgres.postgis()
for row in pip_client.point_in_polygon(40.661367, -111.500959):
f = pip_client.row_to_feature(row)
print f["properties"]
print "--"
for row in pg_client.point_in_polygon(40.661367, -111.500959):
f = pg_client.row_to_feature(row)
print f["properties"]This would print:
{'wof:name': u'Park City', 'wof:placetype': u'locality', 'wof:id': 101727553}
{'wof:name': u'84060', 'wof:placetype': u'postalcode', 'wof:id': 554749823}
{'wof:name': u'Summit', 'wof:placetype': u'county', 'wof:id': 102083555}
{'wof:name': u'Utah', 'wof:placetype': u'region', 'wof:id': 85688567}
{'wof:name': u'United States', 'wof:placetype': u'country', 'wof:id': 85633793}
--
{u'wof:repo': u'whosonfirst-data-postalcode-us', 'geom:longitude': -111.501929, 'geom:latitude': 40.652347, u'wof:name': u'84060', 'wof:placetype': 'postalcode', u'wof:country': u'US', 'wof:parent_id': 101727553L, u'wof:hierarchy': [{u'region_id': 85688567, u'continent_id': 102191575, u'country_id': 85633793, u'locality_id': 101727553, u'county_id': 102083555, u'postalcode_id': 554749823}], 'wof:id': 554749823L}
{u'wof:repo': u'whosonfirst-data-constituency-us', 'geom:longitude': -111.328529, 'geom:latitude': 40.944919, u'wof:name': u'Utah Congressional District 1', 'wof:placetype': 'constituency', u'wof:country': u'us', 'wof:parent_id': 85688567L, u'wof:hierarchy': [{u'continent_id': 102191575, u'country_id': 85633793, u'region_id': 85688567}], 'wof:id': 1108737501L}
{u'wof:repo': u'whosonfirst-data-constituency-us', 'geom:longitude': -111.878816, 'geom:latitude': 39.098999, u'wof:name': u'Utah', 'wof:placetype': 'constituency', u'wof:country': u'US', 'wof:parent_id': 85633793L, u'wof:hierarchy': [{u'continent_id': 102191575, u'country_id': 85633793, u'region_id': 85688567}], 'wof:id': 1108746635L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -111.230352, 'geom:latitude': 40.861066, u'wof:name': u'Summit', 'wof:placetype': 'county', u'wof:country': u'US', 'wof:parent_id': 85688567L, u'wof:hierarchy': [{u'continent_id': 102191575, u'country_id': 85633793, u'region_id': 85688567, u'county_id': 102083555}], 'wof:id': 102083555L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -96.999668, 'geom:latitude': 39.715956, u'wof:name': u'United States', 'wof:placetype': 'country', u'wof:country': u'US', 'wof:parent_id': 102191575L, u'wof:hierarchy': [{u'continent_id': 102191575, u'country_id': 85633793, u'empire_id': 136253057}], 'wof:id': 85633793L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -101.328273, 'geom:latitude': 38.309137, u'wof:name': u'North America', 'wof:placetype': 'continent', u'wof:country': u'', 'wof:parent_id': -1L, u'wof:hierarchy': [], 'wof:id': 102191575L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -106.298325, 'geom:latitude': 42.095168, u'wof:name': u'America/Denver', 'wof:placetype': 'timezone', u'wof:country': u'', 'wof:parent_id': 85633793L, u'wof:hierarchy': [{u'continent_id': 102191575, u'timezone_id': 102047459, u'country_id': 85633793}], 'wof:id': 102047459L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -111.878816, 'geom:latitude': 39.098999, u'wof:name': u'Utah', 'wof:placetype': 'region', u'wof:country': u'US', 'wof:parent_id': 85633793L, u'wof:hierarchy': [{u'continent_id': 102191575, u'country_id': 85633793, u'region_id': 85688567}], 'wof:id': 85688567L}
{u'wof:repo': u'whosonfirst-data', 'geom:longitude': -111.496653, 'geom:latitude': 40.64482, u'wof:name': u'Park City', 'wof:placetype': 'locality', u'wof:country': u'US', 'wof:parent_id': 102083555L, u'wof:hierarchy': [{u'continent_id': 102191575, u'locality_id': 101727553, u'country_id': 85633793, u'region_id': 85688567, u'county_id': 102083555}], 'wof:id': 101727553L}_The example response for the pip_client is actually a bit misleading. As of this writing it has been taught to fetch the source WOF record over the network so there is a lot more data. That's what it would look like if it didn't make a network request._
This package was substantially different once upon a time. If you're look for the old code there is a snapshot release of it dated 20170512.