def test_facility_sampler_missing_activity_fail():
facility_df = pd.DataFrame({
'id': [1, 2, 3, 4],
'activity': ['home', 'work', 'home', 'education']
})
points = [Point((1, 1)), Point((1, 1)), Point((3, 3)), Point((3, 3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a': [1, 2, 3], 'b': [4, 5, 6]})
polys = [
Polygon(((0, 0), (0, 2), (2, 2), (2, 0))),
Polygon(((2, 2), (2, 4), (4, 4), (4, 2))),
Polygon(((4, 4), (4, 6), (6, 6), (6, 4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(
facility_gdf,
zones_gdf, ['home', 'work', 'education'],
random_default=False,
fail=
test_applt_discrete_joint_distribution_sampler_to_fred_not_carefully)
with pytest.raises(UserWarning):
sampler.sample(0, 'education')
def test_facility_sampler_missing_activity_return_None_weighted():
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['home','work','home','education'], 'floorspace': [0,200,700,100]})
points = [Point((1,1)), Point((1,1)), Point((3,3)), Point((3,3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2), (2,2), (2,0))),
Polygon(((2,2), (2,4), (4,4), (4,2))),
Polygon(((4,4), (4,6), (6,6), (6,4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'], random_default=False, fail=False, weight_on='floorspace')
assert sampler.sample(0, 'education') is None
def test_facility_sampler_weighted():
# zero-weight home facility is ignored:
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['home','home','home','education'], 'floorspace': [0,200,700,100]})
points = [Point((1,1)), Point((1.5,1.5)), Point((3,3)), Point((3,3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2), (2,2), (2,0))),
Polygon(((2,2), (2,4), (4,4), (4,2))),
Polygon(((4,4), (4,6), (6,6), (6,4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'], weight_on='floorspace')
assert sampler.sample(0, 'home') == Point((1.5,1.5))
def test_facility_sampler_missing_activity_random_sample():
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['home','work','home','education']})
points = [Point((1,1)), Point((1,1)), Point((3,3)), Point((3,3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2), (2,2), (2,0))),
Polygon(((2,2), (2,4), (4,4), (4,2))),
Polygon(((4,4), (4,6), (6,6), (6,4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'])
assert isinstance(sampler.sample(0, 'education'), Point)
def test_facility_sampler_weighted_maxwalk():
# amongst three workplace alternatives, only one is within walking distance from a PT stop:
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['work','work','work','education'], 'floorspace': [100000,200000,700,100],
'transit':[4000,3000,800,10]})
points = [Point((1,1)), Point((1.5,1.5)), Point((1.8,1.8)), Point((3,3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2), (2,2), (2,0))),
Polygon(((2,2), (2,4), (4,4), (4,2))),
Polygon(((4,4), (4,6), (6,6), (6,4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'], weight_on='floorspace', max_walk=1000)
assert sampler.sample(0, 'work', mode='bus') == Point((1.8,1.8))
def test_facility_dict_build():
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['home','work','home','education']})
points = [Point((1,1)), Point((1,1)), Point((3,3)), Point((3,3))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2), (2,2), (2,0))),
Polygon(((2,2), (2,4), (4,4), (4,2))),
Polygon(((4,4), (4,6), (6,6), (6,4)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'])
assert len(sampler.samplers) == 2
assert sampler.samplers[0]['education'] == None
assert isinstance(sampler.samplers[0]['work'], GeneratorType)
def test_facility_sampler_weighted_distance():
# distance-based sampling
# a 10-min walking trip at 5kph would have a radius of approx 1250 meters.
facility_df = pd.DataFrame({'id':[1,2,3,4], 'activity': ['work','work','work','education'], 'floorspace': [1,1,10,10]})
points = [Point((0.01,0.01)), Point((1000,750)), Point((100,100)), Point((30000,30000))]
facility_gdf = gp.GeoDataFrame(facility_df, geometry=points)
zones_df = pd.DataFrame({'a':[1,2,3], 'b': [4,5,6]})
polys = [
Polygon(((0,0), (0,2000), (2000,2000), (2000,0))),
Polygon(((20000,20000), (20000,40000), (40000,40000), (40000,20000))),
Polygon(((40000,40000), (40000,60000), (60000,60000), (60000,40000)))
]
zones_gdf = gp.GeoDataFrame(zones_df, geometry=polys)
sampler = facility.FacilitySampler(facility_gdf, zones_gdf, ['home', 'work', 'education'], weight_on='floorspace')
sampled_facilities = []
for i in range(20):
sampled_facilities.append(sampler.sample(0, 'work', mode='walk', previous_duration=pd.Timedelta(minutes=15), previous_loc=Point(0,0)))
assert pd.Series(sampled_facilities).value_counts(normalize=True).idxmax() == Point((1000,750))
