def test_raise_nonpoly(dfs):
polydf, _ = dfs
pointdf = polydf.copy()
pointdf['geometry'] = pointdf.geometry.centroid
with pytest.raises(TypeError):
overlay(pointdf, polydf, how="union")
def test_overlay(dfs, how, use_sindex, expected_features):
"""
Basic overlay test with small dummy example dataframes (from docs).
Results obtained using QGIS 2.16 (Vector -> Geoprocessing Tools ->
Intersection / Union / ...), saved to GeoJSON and pasted here
"""
df1, df2 = dfs
result = overlay(df1, df2, how=how, use_sindex=use_sindex)
# construction of result
if how == 'identity':
expected = pd.concat([
GeoDataFrame.from_features(expected_features['intersection']),
GeoDataFrame.from_features(expected_features['difference'])
], ignore_index=True)
else:
expected = GeoDataFrame.from_features(expected_features[how])
# TODO needed adaptations to result
# if how == 'union':
# result = result.drop(['idx1', 'idx2'], axis=1).sort_values(['col1', 'col2']).reset_index(drop=True)
# elif how in ('intersection', 'identity'):
# result = result.drop(['idx1', 'idx2'], axis=1)
assert_geodataframe_equal(result, expected)
# for difference also reversed
if how == 'difference':
result = overlay(df2, df1, how=how, use_sindex=use_sindex)
expected = GeoDataFrame.from_features(
expected_features['difference_inverse'])
assert_geodataframe_equal(result, expected)
def test_preserve_crs(dfs, how):
df1, df2 = dfs
result = overlay(df1, df2, how=how)
assert result.crs is None
crs = {'init': 'epsg:4326'}
df1.crs = crs
df2.crs = crs
result = overlay(df1, df2, how=how)
assert result.crs == crs
def test_union_no_index(self):
# explicitly ignore indicies
dfB = overlay(self.polydf, self.polydf2, how="union", use_sindex=False)
self.assertEquals(dfB.shape, self.union_shape)
# remove indicies from df
self.polydf._sindex = None
self.polydf2._sindex = None
dfC = overlay(self.polydf, self.polydf2, how="union")
self.assertEquals(dfC.shape, self.union_shape)
def test_geometry_not_named_geometry(self):
# Issue #306
# Add points and flip names
polydf3 = self.polydf.copy()
polydf3 = polydf3.rename(columns={'geometry':'polygons'})
polydf3 = polydf3.set_geometry('polygons')
polydf3['geometry'] = self.pointdf.geometry.loc[0:4]
self.assertTrue(polydf3.geometry.name == 'polygons')
df = overlay(polydf3, self.polydf2, how="union")
self.assertTrue(type(df) is GeoDataFrame)
df2 = overlay(self.polydf, self.polydf2, how="union")
self.assertTrue(df.geom_almost_equals(df2).all())
def test_empty_intersection(dfs):
df1, df2 = dfs
polys3 = GeoSeries([Polygon([(-1, -1), (-3, -1), (-3, -3), (-1, -3)]),
Polygon([(-3, -3), (-5, -3), (-5, -5), (-3, -5)])])
df3 = GeoDataFrame({'geometry': polys3, 'col3': [1, 2]})
expected = GeoDataFrame([], columns=['col1', 'col3', 'geometry'])
result = overlay(df1, df3)
assert_geodataframe_equal(result, expected, check_like=True)
def test_overlay_nybb(how):
polydf = read_file(geopandas.datasets.get_path('nybb'))
# construct circles dataframe
N = 10
b = [int(x) for x in polydf.total_bounds]
polydf2 = GeoDataFrame(
[{'geometry': Point(x, y).buffer(10000), 'value1': x + y,
'value2': x - y}
for x, y in zip(range(b[0], b[2], int((b[2]-b[0])/N)),
range(b[1], b[3], int((b[3]-b[1])/N)))],
crs=polydf.crs)
result = overlay(polydf, polydf2, how=how)
cols = ['BoroCode', 'BoroName', 'Shape_Leng', 'Shape_Area',
'value1', 'value2']
if how == 'difference':
cols = cols[:-2]
# expected result
if how == 'identity':
# read union one, further down below we take the appropriate subset
expected = read_file(os.path.join(
DATA, 'nybb_qgis', 'qgis-union.shp'))
else:
expected = read_file(os.path.join(
DATA, 'nybb_qgis', 'qgis-{0}.shp'.format(how)))
# The result of QGIS for 'union' contains incorrect geometries:
# 24 is a full original circle overlapping with unioned geometries, and
# 27 is a completely duplicated row)
if how == 'union':
expected = expected.drop([24, 27])
expected.reset_index(inplace=True, drop=True)
# Eliminate observations without geometries (issue from QGIS)
expected = expected[expected.is_valid]
expected.reset_index(inplace=True, drop=True)
if how == 'identity':
expected = expected[expected.BoroCode.notnull()].copy()
# Order GeoDataFrames
expected = expected.sort_values(cols).reset_index(drop=True)
# TODO needed adaptations to result
result = result.sort_values(cols).reset_index(drop=True)
if how in ('union', 'identity'):
# concat < 0.23 sorts, so changes the order of the columns
# but at least we ensure 'geometry' is the last column
assert result.columns[-1] == 'geometry'
assert len(result.columns) == len(expected.columns)
result = result.reindex(columns=expected.columns)
assert_geodataframe_equal(result, expected, check_crs=False,
check_column_type=False,)
def test_geometry_not_named_geometry(dfs, how, other_geometry):
# Issue #306
# Add points and flip names
df1, df2 = dfs
df3 = df1.copy()
df3 = df3.rename(columns={'geometry': 'polygons'})
df3 = df3.set_geometry('polygons')
if other_geometry:
df3['geometry'] = df1.centroid.geometry
assert df3.geometry.name == 'polygons'
res1 = overlay(df1, df2, how=how)
res2 = overlay(df3, df2, how=how)
assert df3.geometry.name == 'polygons'
if how == 'difference':
# in case of 'difference', column names of left frame are preserved
assert res2.geometry.name == 'polygons'
if other_geometry:
assert 'geometry' in res2.columns
assert_geoseries_equal(res2['geometry'], df3['geometry'],
check_series_type=False)
res2 = res2.drop(['geometry'], axis=1)
res2 = res2.rename(columns={'polygons': 'geometry'})
res2 = res2.set_geometry('geometry')
# TODO if existing column is overwritten -> geometry not last column
if other_geometry and how == 'intersection':
res2 = res2.reindex(columns=res1.columns)
assert_geodataframe_equal(res1, res2)
df4 = df2.copy()
df4 = df4.rename(columns={'geometry': 'geom'})
df4 = df4.set_geometry('geom')
if other_geometry:
df4['geometry'] = df2.centroid.geometry
assert df4.geometry.name == 'geom'
res1 = overlay(df1, df2, how=how)
res2 = overlay(df1, df4, how=how)
assert_geodataframe_equal(res1, res2)
def test_overlay(dfs_index, how, use_sindex):
"""
Basic overlay test with small dummy example dataframes (from docs).
Results obtained using QGIS 2.16 (Vector -> Geoprocessing Tools ->
Intersection / Union / ...), saved to GeoJSON
"""
df1, df2 = dfs_index
result = overlay(df1, df2, how=how, use_sindex=use_sindex)
# construction of result
def _read(name):
expected = read_file(
os.path.join(DATA, 'polys', 'df1_df2-{0}.geojson'.format(name)))
expected.crs = None
return expected
if how == 'identity':
expected_intersection = _read('intersection')
expected_difference = _read('difference')
expected = pd.concat([
expected_intersection,
expected_difference
], ignore_index=True, sort=False)
expected['col1'] = expected['col1'].astype(float)
else:
expected = _read(how)
# TODO needed adaptations to result
if how == 'union':
result = result.sort_values(['col1', 'col2']).reset_index(drop=True)
elif how == 'difference':
result = result.reset_index(drop=True)
assert_geodataframe_equal(result, expected, check_column_type=False)
# for difference also reversed
if how == 'difference':
result = overlay(df2, df1, how=how, use_sindex=use_sindex)
result = result.reset_index(drop=True)
expected = _read('difference-inverse')
assert_geodataframe_equal(result, expected, check_column_type=False)
def test_union_non_numeric_index(self):
import string
letters = list(string.ascii_letters)
polydf_alpha = self.polydf.copy()
polydf2_alpha = self.polydf2.copy()
polydf_alpha.index = letters[:len(polydf_alpha)]
polydf2_alpha.index = letters[:len(polydf2_alpha)]
df = overlay(polydf_alpha, polydf2_alpha, how="union")
assert type(df) is GeoDataFrame
assert df.shape == self.union_shape
assert 'value1' in df.columns and 'Shape_Area' in df.columns
def test_correct_index(dfs):
# GH883 - case where the index was not properly reset
df1, df2 = dfs
polys3 = GeoSeries([Polygon([(1, 1), (3, 1), (3, 3), (1, 3)]),
Polygon([(-1, 1), (1, 1), (1, 3), (-1, 3)]),
Polygon([(3, 3), (5, 3), (5, 5), (3, 5)])])
df3 = GeoDataFrame({'geometry': polys3, 'col3': [1, 2, 3]})
i1 = Polygon([(1, 1), (1, 3), (3, 3), (3, 1), (1, 1)])
i2 = Polygon([(3, 3), (3, 5), (5, 5), (5, 3), (3, 3)])
expected = GeoDataFrame([[1, 1, i1], [3, 2, i2]],
columns=['col3', 'col2', 'geometry'])
result = overlay(df3, df2)
assert_geodataframe_equal(result, expected)
def raster_gt(box, gt):
transform = rio.transform.from_bounds(*box.geometry.values[0].bounds, 768,
768)
species_encoding = {1001: 1, 1005: 2}
inter = gpd.overlay(gt, box, how='intersection')
shapes = ((row.geometry, species_encoding[row.Species])
for _, row in inter.iterrows())
rastered_shape = rio.features.rasterize(shapes=shapes,
out_shape=(768, 768),
transform=transform)
rgb_mask = np.zeros((768, 768, 3))
rgb_mask[:, :, 0] = rastered_shape == 1
rgb_mask[:, :, 1] = rastered_shape == 2
return rgb_mask
def build_mask(data, grid):
dI = gpd.overlay(grid.df,
data.to_crs(epsg=grid.epsg),
how="intersection",
keep_geom_type=False)
dO = gpd.overlay(grid.df,
data.to_crs(epsg=grid.epsg),
how="difference",
keep_geom_type=False)
dataI = np.array(dI["geometry"][0].array_interface()["data"]).reshape(
-1, 2)
dataO = np.array(dO["geometry"][0].array_interface()["data"]).reshape(
-1, 2)
dataM = pd.DataFrame(np.vstack((dataI, dataO)), columns=["x", "y"])
dataM["mask"] = np.nan
dataM.iloc[:dataI.shape[0], 2] = True
dataM.iloc[dataI.shape[0]:, 2] = False
dataM = dataM.sort_values(["x", "y"])
mask_array = dataM["mask"].values.astype(int).reshape(grid.nx, grid.ny).T
return mask_array
def spatialSwathFilter(region,
h5files,
subgroup='//All_Data/ATMS-SDR-GEO_All/'):
swathGroups = groupSwathFiles(h5files)
geoms, sdrnames, geonames = [], [], []
for sg in swathGroups:
geoFile, sdrFile = sg
geoBase = 'HDF5:"{0}":{1}{2}'
lats = gdal.Open(geoBase.format(geoFile, subgroup,
'Latitude')).ReadAsArray()
lons = gdal.Open(geoBase.format(geoFile, subgroup,
'Longitude')).ReadAsArray()
view = gdal.Open(
geoBase.format(geoFile, subgroup,
'SatelliteZenithAngle')).ReadAsArray()
yp, xp = np.where(view < 50)
minindex, maxindex = xp.min(), xp.max()
lons = lons[:, minindex:maxindex]
lats = lats[:, minindex:maxindex]
wVerts = [(lons[i, 0], lats[i, 0]) for i in range(lats.shape[0])
if (lons[i, 0] > -200) and (lats[i, 0] > -200)][::-1]
nVerts = [(lons[0, i], lats[0, i]) for i in range(lats.shape[1])
if (lons[0, i] > -200) and (lats[0, i] > -200)]
eVerts = [(lons[i, -1], lats[i, -1]) for i in range(lats.shape[0])
if (lons[i, -1] > -200) and (lats[i, -1] > -200)]
sVerts = [(lons[-1, i], lats[-1, i]) for i in range(lats.shape[1])
if (lons[-1, i] > -200) and (lats[-1, i] > -200)]
sVerts.append(wVerts[0])
verts = wVerts + nVerts + eVerts + sVerts
geoms.append(geometry.Polygon(verts))
sdrnames.append(sdrFile)
geonames.append(geoFile)
swathGeo = gpd.GeoDataFrame(pd.DataFrame({
'sdr': sdrnames,
'geo': geonames,
'geometry': geoms
}),
geometry=geoms)
swathGeo.crs = {'init': 'epsg:4326'}
intersection = gpd.overlay(region, swathGeo, how='intersection')
return list(intersection.sdr), list(intersection.geo)
def clipMask(crs, shpFile, minX, maxX, minY, maxY, **kwargs):
'''Clip a shapefile to extent
For better performance first spatial join with intersect is computed, and then overlapping polygons are clipped with geopandas overlay
New polygon area is calculated. Note that it assumes that crs is projected for area calculation
:param minX:
:param maxX
:param minY:
:param maxY:
:param shpFile:
:return:
'''
logger.info('Starting to clip polygon minX=%s, maxX=%s, minY=%s, maxY=%s',
minX, maxX, minY, maxY)
logger.info('Opening mask file: %s', shpFile)
# create extent geodataframe
start = time.time()
# open mask shapefile
masksShp = gp.read_file(shpFile)
logger.info('Creating clip polygon from extent...')
extent = gp.GeoSeries([
Polygon([(minX, minY), (minX, maxY), (maxX, maxY), (maxX, minY),
(minX, minY)])
])
dfExtent = gp.GeoDataFrame(geometry=extent)
dfExtent.crs = crs
logger.info('Intersecting Shapefile with extent...')
# intersect with extent
maskIntersect = gp.sjoin(masksShp, dfExtent, how='inner', op='intersects')
#drop columns except geometry and Area
maskIntersect.drop(maskIntersect.columns.difference(['geometry', 'Area']),
1,
inplace=True)
#rename Area column to area_old
maskIntersect.rename(columns={'Area': 'area_old'}, inplace=True)
logger.info('Clip overlapping polygons...')
maskClipped = gp.overlay(maskIntersect, dfExtent, how='intersection')
logger.info('Total time used for clipping %s seconds',
'{0:.3g}'.format(time.time() - start))
return maskClipped
def run(self, region: MultiPolygon, period: DateRange,
granularity: TimeAggregation,
pollutant: Pollutant) -> tuple[DataFrame, GeoDataFrame]:
self._validate(region, period, granularity, pollutant)
self._state = Status.RUNNING
self._progress = 0
# Generate data frame with random emission values per GNFR sector
data = self._create_gnfr_table(pollutant)
for sector in GNFR:
data.loc[sector] = [random() * 100, random() * 18, random() * 22]
# Add totals row at the bottom
data.loc["Totals"] = data.sum(axis=0)
self._progress = 50
# Generate bogus grid with random emission values
geo_data, _ = self._create_grid(region, .1, .1, snap=False)
geo_data = overlay(geo_data,
GeoDataFrame({'geometry': [region]},
crs="EPSG:4326"),
how='intersection')
geo_data.insert(0, "Area [km²]",
geo_data.to_crs(epsg=8857).area /
10**6) # Equal earth projection
geo_data.insert(1, f"Total {pollutant.name} emissions [kg]",
[random() * 100 for _ in range(geo_data.shape[0])])
geo_data.insert(2, "Umin [%]", 42)
geo_data.insert(3, "Umax [%]", 42)
geo_data.insert(4, "Number of values [1]", len(period))
geo_data.insert(5, "Missing values [1]", 0)
if granularity is TimeAggregation.YEARLY:
headings = self._create_column_headings_per_year(period, pollutant)
elif granularity is TimeAggregation.MONTHLY:
headings = self._create_column_headings_per_month(
period, pollutant)
else:
headings = [
f"{day} {pollutant.name} emissions [kg]" for day in period
]
for count, heading in enumerate(headings):
geo_data.insert(6 + count, heading, [
random() * 100 / len(headings)
for _ in range(geo_data.shape[0])
])
self._progress = 100
self._state = Status.READY
return self._create_result_tuple(data, geo_data)
def overlay():
polys1 = geopandas.GeoSeries([Polygon([(0,0), (2,0), (2,2), (0,2)]),
Polygon([(2,2), (4,2), (4,4), (2,4)])])
polys2 = geopandas.GeoSeries([Polygon([(1,1), (3,1), (3,3), (1,3)]),
Polygon([(3,3), (5,3), (5,5), (3,5)])])
df1 = geopandas.GeoDataFrame({'geometry': polys1, 'df1':[1,2]})
df2 = geopandas.GeoDataFrame({'geometry': polys2, 'df2':[1,2]})
#原始叠加显示
ax = df1.plot(color='red')
df2.plot(ax=ax, color='green', alpha=0.5)
plt.title('data')
#联合
res_union = geopandas.overlay(df1, df2, how='union')
ax = res_union.plot(alpha=0.5, cmap='tab10')
df1.plot(ax=ax, facecolor='none', edgecolor='k')
df2.plot(ax=ax, facecolor='none', edgecolor='k')
plt.title('union')
#相交
res_intersection = geopandas.overlay(df1, df2, how='intersection')
ax = res_intersection.plot(alpha=0.5, cmap='tab10')
df1.plot(ax=ax, facecolor='none', edgecolor='k')
df2.plot(ax=ax, facecolor='none', edgecolor='k')
plt.title('intersection')
#交集取反
res_symdiff = geopandas.overlay(df1, df2, how='symmetric_difference')
ax = res_symdiff.plot(alpha=0.5, cmap='tab10')
df1.plot(ax=ax, facecolor='none', edgecolor='k')
df2.plot(ax=ax, facecolor='none', edgecolor='k')
plt.title('symmetric_difference')
plt.show()
def test_overlay_overlap(how):
"""
Overlay test with overlapping geometries in both dataframes.
Test files are created with::
import geopandas
from geopandas import GeoSeries, GeoDataFrame
from shapely.geometry import Point, Polygon, LineString
s1 = GeoSeries([Point(0, 0), Point(1.5, 0)]).buffer(1, resolution=2)
s2 = GeoSeries([Point(1, 1), Point(2, 2)]).buffer(1, resolution=2)
df1 = GeoDataFrame({'geometry': s1, 'col1':[1,2]})
df2 = GeoDataFrame({'geometry': s2, 'col2':[1, 2]})
ax = df1.plot(alpha=0.5)
df2.plot(alpha=0.5, ax=ax, color='C1')
df1.to_file('geopandas/geopandas/tests/data/df1_overlap.geojson',
driver='GeoJSON')
df2.to_file('geopandas/geopandas/tests/data/df2_overlap.geojson',
driver='GeoJSON')
and then overlay results are obtained from using QGIS 2.16
(Vector -> Geoprocessing Tools -> Intersection / Union / ...),
saved to GeoJSON.
"""
df1 = read_file(os.path.join(DATA, 'overlap', 'df1_overlap.geojson'))
df2 = read_file(os.path.join(DATA, 'overlap', 'df2_overlap.geojson'))
result = overlay(df1, df2, how=how)
if how == 'identity':
raise pytest.skip()
expected = read_file(os.path.join(
DATA, 'overlap', 'df1_df2_overlap-{0}.geojson'.format(how)))
if how == 'union':
# the QGIS result has the last row duplicated, so removing this
expected = expected.iloc[:-1]
# TODO needed adaptations to result
result = result.reset_index(drop=True)
if how == 'union':
result = result.sort_values(['col1', 'col2']).reset_index(drop=True)
assert_geodataframe_equal(result, expected, check_column_type=False,
check_less_precise=True)
def extract_tile_items(raster_features, labels, min_x, min_y, tile_width,
tile_height):
"""Extract label items that belong to the tile defined by the minimum
horizontal pixel `min_x` (left tile limit), the minimum vertical pixel
`min_y` (upper tile limit) and the sizes ̀tile_width` and `tile_height`
measured as a pixel amount.
The tile is cropped from the original image raster as follows:
- horizontally, between `min_x` and `min_x+tile_width`
- vertically, between `min_y` and `min_y+tile_height`
This method takes care of original data projection (UTM 37S, Tanzania
area), however this parameter may be changed if similar data on another
projection is used.
Parameters
----------
raster_features : dict
Raw image raster geographical features (`north`, `south`, `east` and
`west` coordinates, `weight` and `height` measured in pixels)
labels : geopandas.GeoDataFrame
Raw image labels, as a set of geometries
min_x : int
Left tile limit, as a horizontal pixel index
min_y : int
Upper tile limit, as a vertical pixel index
tile_width : int
Tile width, measured in pixel
tile_height : int
Tile height, measured in pixel
Returns
-------
geopandas.GeoDataFrame
Set of ground-truth labels contained into the tile, characterized by
their type (complete, unfinished or foundation) and their geometry
"""
area = get_tile_footprint(raster_features, min_x, min_y, tile_width,
tile_height)
bdf = gpd.GeoDataFrame(crs=fiona.crs.from_epsg(raster_features["srid"]),
geometry=[area])
reproj_labels = labels.to_crs(epsg=raster_features["srid"])
tile_items = gpd.sjoin(reproj_labels, bdf)
if tile_items.shape[0] == 0:
return tile_items[["condition", "geometry"]]
tile_items = gpd.overlay(tile_items, bdf)
tile_items = tile_items.explode() # Manage MultiPolygons
return tile_items[["condition", "geometry"]]
def test_overlay_overlap(how):
"""
Overlay test with overlapping geometries in both dataframes.
Test files are created with::
import geopandas
from geopandas import GeoSeries, GeoDataFrame
from shapely.geometry import Point, Polygon, LineString
s1 = GeoSeries([Point(0, 0), Point(1.5, 0)]).buffer(1, resolution=2)
s2 = GeoSeries([Point(1, 1), Point(2, 2)]).buffer(1, resolution=2)
df1 = GeoDataFrame({'geometry': s1, 'col1':[1,2]})
df2 = GeoDataFrame({'geometry': s2, 'col2':[1, 2]})
ax = df1.plot(alpha=0.5)
df2.plot(alpha=0.5, ax=ax, color='C1')
df1.to_file('geopandas/geopandas/tests/data/df1_overlap.geojson',
driver='GeoJSON')
df2.to_file('geopandas/geopandas/tests/data/df2_overlap.geojson',
driver='GeoJSON')
and then overlay results are obtained from using QGIS 2.16
(Vector -> Geoprocessing Tools -> Intersection / Union / ...),
saved to GeoJSON.
"""
df1 = read_file(os.path.join(DATA, 'overlap', 'df1_overlap.geojson'))
df2 = read_file(os.path.join(DATA, 'overlap', 'df2_overlap.geojson'))
result = overlay(df1, df2, how=how)
if how == 'identity':
raise pytest.skip()
expected = read_file(os.path.join(
DATA, 'overlap', 'df1_df2_overlap-{0}.geojson'.format(how)))
if how == 'union':
# the QGIS result has the last row duplicated, so removing this
expected = expected.iloc[:-1]
# TODO needed adaptations to result
result = result.reset_index(drop=True)
if how == 'union':
result = result.sort_values(['col1', 'col2']).reset_index(drop=True)
assert_geodataframe_equal(result, expected, check_column_type=False,
check_less_precise=True)
def test_non_overlapping(how):
p1 = Polygon([(0, 0), (2, 0), (2, 2), (0, 2)])
p2 = Polygon([(3, 3), (5, 3), (5, 5), (3, 5)])
df1 = GeoDataFrame({"col1": [1], "geometry": [p1]})
df2 = GeoDataFrame({"col2": [2], "geometry": [p2]})
result = overlay(df1, df2, how=how)
if how == "intersection":
expected = GeoDataFrame(
{
"col1": np.array([], dtype="int64"),
"col2": np.array([], dtype="int64"),
"geometry": [],
},
index=pd.Index([], dtype="object"),
)
elif how == "union":
expected = GeoDataFrame(
{
"col1": [1, np.nan],
"col2": [np.nan, 2],
"geometry": [p1, p2],
}
)
elif how == "identity":
expected = GeoDataFrame(
{
"col1": [1.0],
"col2": [np.nan],
"geometry": [p1],
}
)
elif how == "symmetric_difference":
expected = GeoDataFrame(
{
"col1": [1, np.nan],
"col2": [np.nan, 2],
"geometry": [p1, p2],
}
)
elif how == "difference":
expected = GeoDataFrame(
{
"col1": [1],
"geometry": [p1],
}
)
assert_geodataframe_equal(result, expected)
def get_user_data_by_vicinity():
all_user_data_result = get_all_user_data()
if not const.HTTP_ERROR:
# Set the area proximity geodataframe
df_wgs84 = pd.DataFrame({
'id': [1],
'latitude': [1.23],
'longitude': [-0.213]
})
df_wgs84_geom = [
Point(xy) for xy in zip(df_wgs84.longitude, df_wgs84.latitude)
]
wgs84_crs = {'init': 'epsg:4326'}
geo_df_wgs84 = gpd.GeoDataFrame(df_wgs84,
crs=wgs84_crs,
geometry=df_wgs84_geom)
df_web_mercator = geo_df_wgs84.to_crs(epsg=3857)
df_web_mercator['geometry'] = df_web_mercator.geometry.buffer(
const.PROXIMITY)
# Set the user data geodataframe
all_user_df = pd.DataFrame.from_dict(all_user_data_result,
orient='columns')
all_user_df['longitude'] = pd.to_numeric(all_user_df['longitude'],
downcast='float')
all_user_df['latitude'] = pd.to_numeric(all_user_df['latitude'],
downcast='float')
all_user_geom = [
Point(xy)
for xy in zip(all_user_df.longitude, all_user_df.latitude)
]
all_user_geo_df = gpd.GeoDataFrame(all_user_df,
crs=wgs84_crs,
geometry=all_user_geom)
all_user_geo_df_web_mercator = all_user_geo_df.to_crs(epsg=3857)
# Overlay the user data geodataframe with the area proximity geodataframe to retrieve target users data
target_users_gdf = gpd.overlay(all_user_geo_df_web_mercator,
df_web_mercator,
how='intersection')
# Set the target users list
target_users_id_list = []
for index, row in target_users_gdf.iterrows():
target_users_id_list.append(row['id_1'])
target_users_item_list = []
for user_item in all_user_data_result:
if user_item['id'] in target_users_id_list:
target_users_item_list.append(user_item)
return target_users_item_list
else:
return all_user_data_result
def _generate_ground_truth(w, h, crop_size, annotation_polygon: Polygon,
pixel_annotation_value):
"""
:param w:
:param h:
:param crop_size:
:param annotation_polygon:
:param pixel_annotation_value:
:return:
"""
patch_mask_polygon = Polygon([(w, h), (w + crop_size, h),
(w + crop_size, h + crop_size),
(w, h + crop_size)])
patch_mask_polygon = gpd.GeoSeries(patch_mask_polygon)
annotation_polygon = gpd.GeoSeries(annotation_polygon)
# Get the intersection of the `patch mask and an annotation
#
# 'patch_mask' would fed into GeoDataFrame as dataset.
gdf_mask = gpd.GeoDataFrame({
'geometry': patch_mask_polygon,
'patch_mask': pixel_annotation_value
})
gdf_curr_annotation = gpd.GeoDataFrame({'geometry': annotation_polygon})
gdf_mask_curr_anno_diff = gpd.overlay(gdf_mask,
gdf_curr_annotation,
how='intersection')
if not gdf_mask_curr_anno_diff.empty:
# 'geom' work as boundary box
mask_curr_anno_intersection_rasterized = \
make_geocube(vector_data=gdf_mask_curr_anno_diff,
resolution=(1., 1.),
geom=json.dumps(mapping(box(w, h, w+crop_size, h+crop_size))),
fill=opt.pixel_anno_ignore)
# TODO: refactor a transformation of geocube data to numpy array
intersection_data = mask_curr_anno_intersection_rasterized.to_dict()
intersection_data = intersection_data['data_vars']['patch_mask'][
'data']
patch_ground_truth = np.array(intersection_data)
return patch_ground_truth
return np.full((crop_size, crop_size),
pixel_annotation_value).astype(np.float)
def get_country_geometries(country_names=None, extent=None, resolution=10):
"""Returns a gpd GeoSeries of natural earth multipolygons of the
specified countries, resp. the countries that lie within the specified
extent. If no arguments are given, simply returns the whole natural earth
dataset.
Take heed: we assume WGS84 as the CRS unless the Natural Earth download
utility from cartopy starts including the projection information. (They
are saving a whopping 147 bytes by omitting it.) Same goes for UTF.
Parameters:
country_names (list, optional): list with ISO3 names of countries, e.g
['ZWE', 'GBR', 'VNM', 'UZB']
extent (tuple, optional): (min_lon, max_lon, min_lat, max_lat) assumed
to be in the same CRS as the natural earth data.
resolution (float, optional): 10, 50 or 110. Resolution in m. Default:
10m
Returns:
GeoDataFrame
"""
resolution = nat_earth_resolution(resolution)
shp_file = shapereader.natural_earth(resolution=resolution,
category='cultural',
name='admin_0_countries')
nat_earth = gpd.read_file(shp_file, encoding='UTF-8')
if not nat_earth.crs:
nat_earth.crs = NE_CRS
if country_names:
if isinstance(country_names, str):
country_names = [country_names]
out = nat_earth[nat_earth.ISO_A3.isin(country_names)]
elif extent:
bbox = Polygon([
(extent[0], extent[2]),
(extent[0], extent[3]),
(extent[1], extent[3]),
(extent[1], extent[2])
])
bbox = gpd.GeoSeries(bbox, crs=nat_earth.crs)
bbox = gpd.GeoDataFrame({'geometry': bbox}, crs=nat_earth.crs)
out = gpd.overlay(nat_earth, bbox, how="intersection")
else:
out = nat_earth
return out
def test_correct_index(dfs):
# GH883 - case where the index was not properly reset
df1, df2 = dfs
polys3 = GeoSeries([
Polygon([(1, 1), (3, 1), (3, 3), (1, 3)]),
Polygon([(-1, 1), (1, 1), (1, 3), (-1, 3)]),
Polygon([(3, 3), (5, 3), (5, 5), (3, 5)]),
])
df3 = GeoDataFrame({"geometry": polys3, "col3": [1, 2, 3]})
i1 = Polygon([(1, 1), (1, 3), (3, 3), (3, 1), (1, 1)])
i2 = Polygon([(3, 3), (3, 5), (5, 5), (5, 3), (3, 3)])
expected = GeoDataFrame([[1, 1, i1], [3, 2, i2]],
columns=["col3", "col2", "geometry"])
result = overlay(df3, df2, keep_geom_type=True)
assert_geodataframe_equal(result, expected)
def intersect_aois(shps):
logger.info('Finding intersection among provided polygons...')
logger.info('Reading: {}'.format(shps[0]))
shp = gpd.read_file(shps[0])
if len(shp) > 1:
shp = dissolve_gdf(shp)
for s in shps[1:]:
logger.info('Reading: {}'.format(s))
shp2 = gpd.read_file(s)
if len(shp2) > 1:
shp2 = dissolve_gdf(shp2)
logger.info('Finding intersection...')
shp = gpd.overlay(shp, shp2)
return shp
def get_area_and_start_point():
#! load crawl area
area = gpd.read_file('../input/area_test.geojson')
df_roads = gpd.read_file("../input/深圳市_osm路网_道路.geojson")
df_nodes = gpd.read_file("../input/深圳市_osm路网_节点.geojson")
roads = gpd.overlay(df_roads, area, how="intersection")
nodes = gpd.overlay(df_nodes, area, how="intersection")
if False:
ax = map_visualize(roads, color='red', scale=0.1)
nodes.plot(ax=ax, )
ax.axis('off')
ax = map_visualize(roads.set_geometry('start'), color='red', scale=0.1)
ax.axis('off')
roads.loc[:, 'start'] = roads.geometry.apply(
lambda i: Point(i.xy[0][0], i.xy[1][0]))
roads.loc[:, 'end'] = roads.geometry.apply(
lambda i: Point(i.xy[0][-1], i.xy[1][-1]))
roads.loc[:, 'start_bd_mc'] = roads.start.apply(
lambda i: wgs_to_bd_mc(*i.coords[0]))
return roads.start_bd_mc.values.tolist(), area.loc[0].geometry
def dist_cont(point_df,dist_list,outside,buff_res):
if point_df.crs != outside.crs:
print('Point df and Outside df are not the same CRS')
return None
# Making outside area out dissolved object
out_cop = outside[['geometry']].copy()
out_cop['Constant'] = 1
out_cop = out_cop.dissolve('Constant')
# Make sure points are inside area
inside = point_df.within(out_cop['geometry'][1])
point_cop = point_df[inside].copy()
point_cop = point_df.copy()
point_cop['Constant'] = 1 #Constant for dissolve
point_cop = point_cop[['Constant','geometry']].copy()
res_buffers = []
for i,d in enumerate(dist_list):
print(f'Doing buffer {d}')
if i == 0:
res = dissolve_buff(point_cop, d, buff_res)
res_buffers.append(res.copy())
else:
res_new = dissolve_buff(point_cop, d, buff_res)
res_buffonly = gpd.overlay(res_new, res, how='difference')
res = res_new.copy()
res_buffers.append( res_buffonly.copy() )
# Now take the difference with the larger area
print('Working on leftover difference now')
leftover = gpd.overlay(out_cop, res, how='difference')
res_buffers.append(leftover)
for i,d in enumerate(dist_list):
res_buffers[i]['Distance'] = str(d)
res_buffers[-1]['Distance'] = 'Outside'
# New geopandas DF
comb_df = pd.concat(res_buffers)
comb_df.reset_index(inplace=True, drop=True)
return comb_df
def stackUnion(add, stack, thing, sliver_size=0.001):
# Prepare to fail
failures = []
backup = stack.copy()
# Ensure the geometries are all valid
add = ensureValid(add)
stack = ensureValid(stack)
# Union the new layer to the overlay
try:
try:
stack = gpd.overlay(add, stack, "union")
except:
stack = roundCoords(stack, 5)
try:
stack = gpd.overlay(add, stack, "union")
except:
add = roundCoords(add, 5)
stack = gpd.overlay(ensureValid(add), ensureValid(stack),
"union")
# Round the coordinates
stack = roundCoords(stack, 5)
print(f" Added {thing}{' '*(21-len(thing))}{now()}")
except Exception as e:
failures.append(thing)
print(e)
print(f"--- FAILED TO ADD {thing} ---------\n")
return backup, failures
# Return the new union
stack = ensureValid(stack)
return stack, failures
def overlay_grid_and_plants(gdf, grid, img_path):
gdf_rd = gdf.to_crs(epsg=28992)
#create new columns
grid['id'] = range(len(grid))
grid['area'] = np.nan
#intersect data and grid
intersect = gpd.overlay(grid, gdf_rd, how='intersection')
#calculate area per grid cell and write area to column
for ID in intersect.id:
temp = intersect[intersect['id'] == ID]
area = temp.geometry.area.sum()
grid.loc[ID, 'area'] = float(area)
#write grid to file as output
grid.to_file(os.path.dirname(img_path) + '/plant_area_grid.shp')
return grid
def filter_map(gdf, bbox=[0, 0, 180, 90]):
p1 = Point(bbox[0], bbox[3])
p2 = Point(bbox[2], bbox[3])
p3 = Point(bbox[2], bbox[1])
p4 = Point(bbox[0], bbox[1])
np1 = (p1.coords.xy[0][0], p1.coords.xy[1][0])
np2 = (p2.coords.xy[0][0], p2.coords.xy[1][0])
np3 = (p3.coords.xy[0][0], p3.coords.xy[1][0])
np4 = (p4.coords.xy[0][0], p4.coords.xy[1][0])
filter = gpd.GeoDataFrame(gpd.GeoSeries(Polygon([np1, np2, np3, np4])),
columns=['geometry'],
crs=gdf.crs)
return gpd.overlay(gdf, filter, how='intersection')
def overlap_calc(_id, poly, grid_file, weight, service_type):
value_dict = Counter()
if type(poly.iloc[0][service_type]) != type(None):
value = float(poly[service_type]) * weight
intersect = gpd.overlay(grid_file, poly, how='intersection')
intersect['overlapped'] = intersect.area
intersect['percent'] = intersect['overlapped'] / intersect['area']
intersect = intersect[intersect['percent'] >= 0.5]
intersect_region = intersect['id']
for intersect_id in intersect_region:
try:
value_dict[intersect_id] += value
except:
value_dict[intersect_id] = value
return (_id, value_dict)
def test_correct_index(dfs):
# GH883 - case where the index was not properly reset
df1, df2 = dfs
polys3 = GeoSeries([
Polygon([(1, 1), (3, 1), (3, 3), (1, 3)]),
Polygon([(-1, 1), (1, 1), (1, 3), (-1, 3)]),
Polygon([(3, 3), (5, 3), (5, 5), (3, 5)])
])
df3 = GeoDataFrame({'geometry': polys3, 'col3': [1, 2, 3]})
i1 = Polygon([(1, 1), (1, 3), (3, 3), (3, 1), (1, 1)])
i2 = Polygon([(3, 3), (3, 5), (5, 5), (5, 3), (3, 3)])
expected = GeoDataFrame([[1, 1, i1], [3, 2, i2]],
columns=['col3', 'col2', 'geometry'])
result = overlay(df3, df2)
assert_geodataframe_equal(result, expected)
def test_keep_geom_type_geometry_collection2():
polys1 = [
box(0, 0, 1, 1),
box(1, 1, 3, 3).union(box(1, 3, 5, 5)),
]
polys2 = [
box(0, 0, 1, 1),
box(3, 1, 4, 2).union(box(4, 1, 5, 4)),
]
df1 = GeoDataFrame({"left": [0, 1], "geometry": polys1})
df2 = GeoDataFrame({"right": [0, 1], "geometry": polys2})
result1 = overlay(df1, df2, keep_geom_type=True)
expected1 = GeoDataFrame(
{
"left": [0, 1],
"right": [0, 1],
"geometry": [box(0, 0, 1, 1), box(4, 3, 5, 4)],
}
)
assert_geodataframe_equal(result1, expected1)
result1 = overlay(df1, df2, keep_geom_type=False)
expected1 = GeoDataFrame(
{
"left": [0, 1, 1],
"right": [0, 0, 1],
"geometry": [
box(0, 0, 1, 1),
Point(1, 1),
GeometryCollection([box(4, 3, 5, 4), LineString([(3, 1), (3, 2)])]),
],
}
)
assert_geodataframe_equal(result1, expected1)
def filter_poly(
poly_pieces_path, markup_path,
pieces_info_path, original_image_path,
image_pieces_path, mask_pieces_path,
pxl_size_threshold, pass_chance
):
original_image = rs.open(original_image_path)
geojson_markup = gp.read_file(markup_path)
geojson_markup = geojson_markup.to_crs(original_image.crs)
pieces_info = pd.read_csv(pieces_info_path)
for i in tqdm(range(len(pieces_info))):
poly_piece_name = pieces_info['piece_geojson'][i]
start_x = pieces_info["start_x"][i]
start_y = pieces_info["start_y"][i]
x, y = original_image.transform * (start_x + 1, start_y + 1)
filename, _ = os.path.splitext(poly_piece_name)
try:
poly_piece = gp.read_file(os.path.join(poly_pieces_path, poly_piece_name))
except fiona.errors.DriverError:
print('Polygon is not found.')
remove_piece(
filename, poly_pieces_path,
image_pieces_path, mask_pieces_path
)
continue
if random() < pass_chance:
continue
intersection = gp.overlay(geojson_markup, poly_piece, how='intersection')
adjacency_list = compose_adjacency_list(intersection['geometry'])
components = get_components(intersection['geometry'], adjacency_list)
multi_polys = []
for component in components:
multi_polys.append(MultiPolygon(poly for poly in component))
png_file = os.path.join(mask_pieces_path, filename + '.png')
if len(multi_polys) == 0 or (imageio.imread(png_file)).sum() < 255 * pxl_size_threshold:
remove_piece(
filename, poly_pieces_path,
image_pieces_path, mask_pieces_path
)
def find_neighbors_in_shape_file(paths, existing_neighbors):
"""This function finds the neighbors in the shape file. Somehow, max-p cannot figure out the correct neighbors and
some clusters are physically neighbors but they are not considered as neighbors. This is where this function comes
in.
:param folder_names = The names of all the folders created for output.
:param existing_neighbors = The neighbors matrix that is created by using w and knn. The new neighbors are to be
added to this matrix.
:
"""
df = gpd.read_file(paths["parts_max_p"] + 'max_p_combined.shp')
df["NEIGHBORS"] = None
for index, cluster_number in df.iterrows():
# get 'not disjoint' countries
import pdb; pdb.set_trace()
neighbors = df[~df.geometry.disjoint(cluster_number.geometry.buffer(0.005))].CL.tolist()
df1 = df
df1.crs = {'init': 'epsg:4326'}
df1 = df1.to_crs({'init': 'epsg:32662'})
df2 = cluster_number.to_frame().T
df2 = gpd.GeoDataFrame(df2, geometry='geometry')
df2.crs = {'init': 'epsg:4326'}
df2 = df2.to_crs({'init': 'epsg:32662'})
df2.geometry = df2.geometry.buffer(100) # in m
test = gpd.overlay(df1, df2, how='intersection')
test['area'] = test['geometry'].area / 10 ** 6 # in km²
test = test[test['area'] > 0.01] # avoids that neighbors share only a point or a very small area
neighbors2 = test.CL_1.tolist()
neighbors = neighbors2
# remove own name from the list
neighbors = [cl_no for cl_no in neighbors if cluster_number.CL != cl_no]
# add names of neighbors as NEIGHBORS value
df.at[index, "NEIGHBORS"] = ','.join(str(n) for n in neighbors)
# Making the w.neighbors dictionary for replacing it in max_p_algorithm_2.
neighbors_corrected = dict()
for index, row in df.iterrows():
neighbors_for_one = row['NEIGHBORS'].split(',')
neighbors_int = list()
for neighbor in neighbors_for_one:
if neighbor:
neighbors_int.append(int(neighbor))
neighbors_corrected[index] = neighbors_int
for value in existing_neighbors[index]:
if value not in neighbors_corrected[index]:
neighbors_corrected[index].append(value)
neighbors_corrected[index] = sorted(neighbors_corrected[index])
return neighbors_corrected
def drawSectors(center, radius, sectors, start, steps):
end = 360 + start # end of circle in degrees
# prepare parameters
if start > end:
start = start - 360
else:
pass
step_angle_width = (end - start) / steps
sector_width = (end - start) / sectors
steps_per_sector = int(math.ceil(steps / sectors))
features = []
for x in xrange(0, int(sectors)):
segment_vertices = []
# first the center and first point
segment_vertices.append(polar_point(center, 0, 0))
segment_vertices.append(
polar_point(center, start + x * sector_width, radius))
# then the sector outline points
for z in xrange(1, steps_per_sector):
segment_vertices.append(
(polar_point(center,
start + x * sector_width + z * step_angle_width,
radius)))
# then again the center point to finish the polygon
segment_vertices.append(
polar_point(center, start + x * sector_width + sector_width,
radius))
segment_vertices.append(polar_point(center, 0, 0))
# create feature
features.append(Polygon(segment_vertices))
polys2 = gpd.GeoSeries(features)
global df2
df2 = gpd.GeoDataFrame({'geometry': polys2, 'id': range(sectors)})
df2.to_file("./output/sectors.shp")
global res #shapefile
res = gpd.overlay(
df2, df1,
how='intersection') #res_union=gpd.overlay(df1,df2,how='union')
res.to_file("./output/result.shp")
def image_to_walls(img):
"""From an image, retrieve the walls.
Aimed at usage for the structure input images.
Args:
img (np.array): Input image from structure dataset.
Returns:
geodataframe with the walls as a single polygon.
"""
gray_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
contours, _ = cv2.findContours(gray_img, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
poly_list = [transform_contour_to_polygon(_contour) for _contour in contours]
poly_list = [x for x in poly_list if x]
rooms = gpd.GeoDataFrame(geometry=poly_list)
rooms["area"] = rooms.area
rooms = rooms.sort_values("area", ascending=False)
rooms = rooms.tail(-1).reset_index(
drop=True
) # drop the largest area area outside of the floor
# find floors:
floors = (
gpd.GeoDataFrame(geometry=[rooms.unary_union]).explode().reset_index(drop=True)
)
# find best overlapping areas so they can be removed.
best_matches = {}
for i, _floor in floors.iterrows():
_floor = _floor.geometry
best_overlap = 0
for j, _room in rooms.iterrows():
_room = _room.geometry
if _floor.intersects(_room):
overlap = _floor.intersection(_floor).area / _floor.area * 100
if overlap > best_overlap:
best_matches[i] = j
best_overlap = overlap
rooms = rooms.drop(best_matches.values(), axis=0)
res_union = gpd.overlay(floors, rooms, how="difference")
res_union.geometry = res_union.scale(yfact=-1, origin=(0, 0))
return res_union
def buffer_point_polygon_overlay(df,
buff_dist=2000,
method='difference',
oid_fld='NewID',
erase_shp_files=[]):
"""Generate <num_points> random points within a geometry.
Parameters
---------------------------------
df: a GeoPandas GeoDataFrame
This is a GeoDataFrame which contains the point geometries.
buff_dist: number of points to generate within the polygon
See above.
method: see GeoPandas overlay doc for how=keyword
num_points_fld: field containing number of points to generate within associated geometry
oid_fld: field containing value to assign each geometry created within a village
erase_shp_files: list containing paths to additional shapefiles for which to erase from buffered household geometries
Usage Notes
---------------------------------
"""
# Replace the dataframe geometries with buffers
df['geometry'] = df['geometry'].buffer(buff_dist)
new_df = df.copy()
# Iteratively erase the additional geometries
for erase_shp in erase_shp_files:
# load the new shape file and make sure it has the same spatial reference
temp_df = gpd.read_file(erase_shp)
print(temp_df.crs, df.crs)
assert temp_df.crs['init'] == df.crs['init']
# erase the geometry from the buffered point dataframe
new_df = gpd.overlay(new_df, temp_df, how=method)
# ensure same CRS
new_df.crs = df.crs
return new_df
def calculate_kde(clean_sur,years):
contour_collection=[]
for year in years:
#get year data
kdev = str_to_class("kdeclus"+(str(year)))
year_data = str(kdev.objects.filter(surname=clean_sur).values('kde'))
#prepare data
idx = [int(x) for x in year_data.split(';')[0][21:].split(',')]
kdx = [int(x) for x in year_data.split(';')[1][:-5].split(',')]
#pd dataframe
kdf = pd.DataFrame({'gid':idx,'val':kdx})
kdf = kdf[(kdf['val'] <= level)]
#add values to grid
kde = pd.merge(gridc,kdf,on='gid',how='inner')
coord = [[int(x[1]),int(x[0])] for x in (list(zip(kde.x,kde.y)))]
cs,lbls = dbscan(coord,eps=2000)
kde = kde.copy()
kde['group'] = lbls
kde = kde[(kde['group'] >= 0)]
#group to concave points
contourp = to_concave_points(kde,coord)
#clip
contours = gpd.GeoSeries([Polygon(contour) for contour in contourp if len(contour) >= 3])
contours = gpd.GeoDataFrame({'geometry': contours})
contours.crs = from_epsg(27700)
clp_prj = gpd.overlay(uk,contours,how='intersection')
#smooth and project
clp_prj['geometry'] = clp_prj.geometry.buffer(10000,join_style=1).buffer(-10000,join_style=1)
clp_prj['geometry'] = clp_prj['geometry'].to_crs(epsg=4326)
#to json
contourprj = clp_prj.to_json()
#add to collection
data = []
data.append(year)
data.append(contourprj)
contour_collection.append(data)
#return
return(contour_collection)
def allocate_population_to_raster(self):
pop_shp = gpd.read_file(self.filepath.root_tmp_path + 'pop_shp.shp')
in_rst_fn = self.filepath.root_work_path + 'dem_aggr_rst.tif'
rst_shp = self.polygonize_raster_layer(in_rst_fn)
# rst_shp = gpd.read_file(self.filepath.root_work_path + 'rst_shp.shp')
res_intersection = gpd.overlay(rst_shp, pop_shp, how='intersection')
# res_intersection.to_file(self.filepath.root_work_path + 'intersection_shp.shp')
# res_intersection = gpd.read_file(self.filepath.root_work_path + 'intersection_shp.shp')
original_size = pop_shp.area[0]
res_intersection['TOT_P'] = res_intersection['TOT_P'] * (
res_intersection.area / original_size)
if in_rst_fn not in self.raster_metadata:
self.load_raster_metadata(in_rst_fn)
out_rst_fn = self.filepath.root_work_path + 'pop_rst.tif'
with rasterio.open(out_rst_fn, 'w+',
**self.raster_metadata[in_rst_fn]) as out_rst:
out_rst_data = out_rst.read(1)
shapes = ((geom, value) for geom, value in zip(
res_intersection.geometry, res_intersection.shape[0] * [0])
if features.is_valid_geom(geom))
burned = features.rasterize(shapes=shapes,
fill=0,
out=out_rst_data,
transform=out_rst.transform,
all_touched=True,
merge_alg=MergeAlg.replace)
out_rst.write_band(1, burned)
out_rst_data = out_rst.read(1)
shapes = ((geom, value) for geom, value in zip(
res_intersection.geometry, res_intersection.TOT_P)
if features.is_valid_geom(geom))
burned = features.rasterize(shapes=shapes,
fill=0,
out=out_rst_data,
transform=out_rst.transform,
all_touched=True,
merge_alg=MergeAlg.add)
out_rst.write_band(1, burned)
return
def intersect_grid_with_habitat(self, grid_gdf, hab_gdf):
"""Returns list of geodataframes where grid_gdf intersects hab_gdf"""
with open('missing_joins.txt', 'w+') as f:
for _, grid_tile in grid_gdf.iterrows():
tile = gpd.GeoDataFrame({'OrthoID': [grid_tile['OrthoID']], 'geometry': [grid_tile['geometry']]})
tile.crs = {'init' :'epsg:32639'}
gdf_sub = gpd.overlay(tile, hab_gdf, how='intersection')
tile_folder = self.out_folder.joinpath(f'{grid_tile["OrthoID"]}')
if not tile_folder.exists():
tile_folder.mkdir(parents=True, exist_ok=True)
outfile = tile_folder.joinpath(f'{grid_tile["OrthoID"]}.shp')
gdf_sub.crs = {'init' :'epsg:32639'}
if not gdf_sub.empty:
gdf_sub.to_file(outfile)
else:
f.write(grid_tile['OrthoID'] + '\n')
def test_intersection(self):
df = overlay(self.polydf, self.polydf2, how="intersection")
assert df['BoroName'][0] is not None
assert df.shape == (68, 7)
def test_geoseries_warning(dfs):
df1, df2 = dfs
# Issue #305
with pytest.raises(NotImplementedError):
overlay(df1, df2.geometry, how="union")
def test_duplicate_column_name(dfs):
df1, df2 = dfs
df2r = df2.rename(columns={'col2': 'col1'})
res = overlay(df1, df2r, how="union")
assert ('col1_1' in res.columns) and ('col1_2' in res.columns)
def test_identity(self):
df = overlay(self.polydf, self.polydf2, how="identity")
assert df.shape == (154, 7)
def test_bad_how(dfs):
df1, df2 = dfs
with pytest.raises(ValueError):
overlay(df1, df2, how="spandex")
def test_duplicate_column_name(self):
polydf2r = self.polydf2.rename(columns={'value2': 'Shape_Area'})
df = overlay(self.polydf, polydf2r, how="union")
self.assertTrue('Shape_Area_2' in df.columns and 'Shape_Area' in df.columns)
def time_overlay(self, op):
overlay(self.countries, self.capitals, how=op)
def test_identity(self):
df = overlay(self.polydf, self.polydf2, how="identity")
self.assertEquals(df.shape, (154, 7))
def test_symmetric_difference(self):
df = overlay(self.polydf, self.polydf2, how="symmetric_difference")
self.assertEquals(df.shape, (122, 7))
def test_nonpoly(self):
with pytest.raises(TypeError):
overlay(self.pointdf, self.polydf, how="union")
def test_intersection(self):
df = overlay(self.polydf, self.polydf2, how="intersection")
self.assertIsNotNone(df['BoroName'][0])
self.assertEquals(df.shape, (68, 7))
def test_symmetric_difference(self):
df = overlay(self.polydf, self.polydf2, how="symmetric_difference")
assert df.shape == (122, 7)
def time_overlay(self, op):
overlay(self.df1, self.df2, how=op)
def test_difference(self):
df = overlay(self.polydf, self.polydf2, how="difference")
self.assertEquals(df.shape, (86, 7))
def test_bad_how(self):
with pytest.raises(ValueError):
overlay(self.polydf, self.polydf, how="spandex")
def test_difference(self):
df = overlay(self.polydf, self.polydf2, how="difference")
assert df.shape == (86, 7)
def f():
overlay(self.polydf, self.polydf2.geometry, how="union")
def test_union(self):
df = overlay(self.polydf, self.polydf2, how="union")
self.assertTrue(type(df) is GeoDataFrame)
self.assertEquals(df.shape, self.union_shape)
self.assertTrue('value1' in df.columns and 'Shape_Area' in df.columns)
