我正在处理带有geopandas
的GIS数据,并将其存储在各种Django
模型中。 geopandas
在后台使用shapely
,而Django
在后台使用。
以下是一些代码:
import geopandas as gpd
from django.contrib.gis.db import models
class MyModel(models.Model):
geometry = models.PolygonField()
name = models.CharField(max_length=255, null=False, unique=True)
some_property = models.IntegerField()
gdf = gpd.read_file("some_data.geojson")
# ...do some stuff w/ gdf...
for data in gdf.to_dict("records"):
name = data.pop("name")
MyModel.objects.create_or_update(
name=name,
defaults=data,
)
以上操作将失败,并显示以下错误:
TypeError:无法将MyModel SpatialProxy(POLYGON)设置为以下类型的值:
除非我添加一些讨厌的代码,例如:
from django.contrib.gis.geos import fromstr, Polygon
data["geometry"] = Polygon(fromstr(str(data["geometry"])))
有什么方法可以避免这种情况,并直接从shapely
映射到Django
?
修改:
以下是一些值:
>> data["geometry"]
<shapely.geometry.polygon.Polygon object at 0x7fb374f41908>
>> str(data["geometry"])
'POLYGON ((-4.337076919429241 53.41842814531255, -4.336698521348041 53.4182242737367, ....))'
>> fromstr(str(data["geometry"]))
<Polygon object at 0x7fb3733d158e>
答案 0 :(得分:1)
您的解决方案似乎没有您想像的那么棘手。
由于您的data['geometry']
字段返回了WKT
字符串表示形式('POLYGON ((-4.337076919429241 53.41842814531255, ... ))
),因此您可以避免执行fromstr
步骤并将其直接传递给to a GEOSGeometry:
from django.contrib.gis.geos import GEOSGeometry
polygon = GEOSGeometry('POLYGON ((-4.337076919429241 53.41842814531255, ... ))')
您还可以添加一些错误处理,而不必担心解决方案崩溃了:):
for data in gdf.to_dict("records"):
name = data.pop("name")
geometry_str = data.pop('geometry')
try:
geometry = GEOSGeometry(geometry_str)
except (TypeError, ValueError) as exc:
# If the geometry_str is not a valid WKT, EWKT or HEXEWKB string
# or is None then either continue, break or do something else.
# I will go with continue here.
continue
if geometry.geom_type != 'Polygon':
# If the created geometry is not a Polygon don't pass it on MyModel
continue
MyModel.objects.create_or_update(
name=name, geometry=geometry,
defaults=data,
)
答案 1 :(得分:0)
更改此行
gdf = gpd.read_file(“ some_data.geojson”)
到
gdf = gpd.read_file(“ some_data.geojson”,driver =“ GeoJSON”)