Selecting and removing polygons that overlap within a single GeoDataFrame in Python

Question

I have a GeoDataFrame made of polygons- some overlapping, some not (see picture above). I'd like to create a shapefile or GeoPackage of only the polygon centroids that are not overlapping. There is a very clear way of how to do this in QGIS here, but I'd like to complete this in Python.

This is what I have tried thus far:

combined = point_gdf.overlay(point_gdf, how="union")
valid_buffers = combined.drop_duplicates()
return valid_buffers

How do you select the non-overlapping polygons from a single GeoDataFrame and create a new GeoPackage from that selection in Python?

Answer 1

Dissolve, create a cluster column, join this to the data frame, count, extract the single polygons:

import geopandas as gpd
df = gpd.read_file(r"C:\TEMP\random_polygons.shp")
diss = df.dissolve() #Dissolve into one big multipolygon
singleparts = gpd.GeoDataFrame(geometry=diss.apply(lambda x: [part for part in x.geometry.geoms], axis=1).explode(), crs=df.crs) #Multipart to singleparts
singleparts["cluster"] = range(singleparts.shape[0]) #Calculate a cluster/row number column

joined = gpd.sjoin(left_df=df, right_df=singleparts) #Join the cluster numbers to the df

#Count how many times the cluster number occurs. If more than once, there are other polygons overlapping current
joined = joined.merge(joined.groupby("cluster")["cluster"].count().rename("polycount"), left_on="cluster", right_index=True)

joined = joined.loc[joined["polycount"]==1] #Extract polygons with only one occurence

joined.to_file(r"C:\TEMP\random_polygons_SINGLES.shp")