def test_urls(self):
import rowgenerators as rg
gdf = rg.geoframe('censusgeo://CA/140')
print(gdf.set_index('geoid').head())
return
u = rg.parse_app_url('census://CA/140/B17001')
t = u.get_resource().get_target()
print(t, t.year, t.release)
self.assertEqual('census://CA/140/B17001', str(t))
self.assertEqual(2016, t.year)
self.assertEqual(5, t.release)
u = rg.parse_app_url('census://2015/3/CA/140/B17001')
t = u.get_resource().get_target()
print(t, t.year, t.release)
self.assertEqual('census://2015/3/CA/140/B17001', str(t))
self.assertEqual(2015, t.year)
self.assertEqual(3, t.release)
gdf = t.geoframe()
self.assertEqual(43.083, gdf.area.sum().round(3))
gdf = rg.geoframe('census://CA/140/B17001')
self.assertEqual(43.083, gdf.area.sum().round(3))
gdf = rg.geoframe('censusgeo://CA/140')
self.assertEqual(43.083, gdf.area.sum().round(3))
def test_shell_not_linear_ring(self):
import rowgenerators as rg
t = rg.geoframe('censusgeo://2019/5/CA/tract')
print(len(t))
def test_national_geo(self):
from rowgenerators import geoframe
gf = geoframe('censusgeo://US/cbsa')
print(len(gf))
def _f_get_split_blocks(st, cache, url):
k = f'blocks/geo/{st}'
if not cache.exists(k):
df = rg.geoframe(url).to_crs(4326)
df['geoid'] = df.geoid20.apply(lambda v: str(Block.parse(v).as_acs()))
df = df.rename(columns={
'aland20': 'aland',
'awater20': 'awater',
'intptlat20': 'lat',
'intptlon20': 'lon',
})
df = df[['geoid', 'aland', 'awater', 'lat', 'lon', 'geometry']]
df['lat'] = df.lat.astype(float)
df['lon'] = df.lon.astype(float)
cache.put(k, df)
return k
def tracts(self):
if self._tracts is None:
logger.info("Building tracts")
url_t = self.pkg.reference('us_tracts_template').url
frames = [
rg.geoframe(url_t.format(st=st))
for st in tqdm(stusab.values())
]
tracts = pd.concat(frames).to_crs(4326)
# Mark the tracts in the continential US
tracts['continential'] = tracts.statefp.isin(
self.states.statefp.unique()).astype(int)
tracts['tract_id'] = tracts.reset_index().index
# Need to convert to each UTM zone to get accurate area
# computation.
# tracts = tracts.sort_values('geoid').reset_index()
# frames = [ g.to_crs(int(idx)).area for idx, g in tracts.groupby('utm_epsg')]
# t = pd.concat(frames).to_frame('utm_area')
# tracts = tracts.join(t)
tracts['geohash'] = tracts[[
'intptlat', 'intptlon'
]].astype(float).apply(lambda r: gh.encode(r.intptlat, r.intptlon),
axis=1)
tracts['gh4'] = tracts.geohash.str.slice(0, 4)
self._tracts = tracts[[
'geoid', 'tract_id', 'geohash', 'statefp', 'intptlat',
'intptlon', 'geometry', 'aland', 'awater', 'gh4',
'continential'
]]
return self._tracts
def census_geo(state, sl='state', year=2018, release=5):
import rowgenerators as rg
return rg.geoframe(f'censusgeo://{year}/{release}/{state}/{sl}')