IO / Read and write KML file with geopandas and fiona driver

kml_io.py

import fiona 
import geopandas as gpd

# Enable fiona driver
gpd.io.file.fiona.drvsupport.supported_drivers['KML'] = 'rw'

# Read file
df = gpd.read_file(path, driver='KML')

# Write file
with fiona.drivers():
    # Might throw a WARNING - CPLE_NotSupported in b'dataset sample_out.kml does not support layer creation option ENCODING'
    df.to_file('sample_out.kml', driver='KML')
    
# Drop Z dimension of polygons that occurs often in kml 
import shapely
import fiona 
import geopandas as gpd

# Enable fiona driver
fiona.supported_drivers['KML'] = 'rw'

# Read file
df = gpd.read_file(path, driver='KML')

# Write file
with fiona.drivers():
    # Might throw a WARNING - CPLE_NotSupported in b'dataset sample_out.kml does not support layer creation option ENCODING'
    df.to_file('sample_out.kml', driver='KML') # You may change driver to GeoJSON

# Drop Z dimension of polygons that occurs often in kml - * expansion handles xy anx xyz cases
import shapely
df = df.set_geometry(
    df.geometry.map(
        lambda polygon: shapely.ops.transform(lambda x, y, *_: (x, y), polygon)
    )
)
# Orient the shape counter clockwise to be Vega conform
df = df.set_geometry(
    df.geometry.map(lambda geom: shapely.geometry.polygon.orient(geom))
)
    

Hi @mazzma12 - this is helpful and really demonstrates how Fiona does all of the heavy lifting for ogr. However I created a KML file from a GPKG using this method and ran into an issue - it seems somewhere along the line the field names get screwed up (the KML <name> was taken from the feature 'date' field for example). If I do it the hard way using ogr the naming is correct. Have you seen this and do you know a way around it? Or is this a question for the Fiona team?

I have really a few experiences with Linestring, sorry... It works fine on Polygon and Point type for me. The only constraint is that it does not work with nested KML as you could have in Google Earth.