Getting GeoSeries with all the points and their coordinates from GeoSeries that has multiple types of geometries in GeoPandas

Question

I'm writing a program that reads two MultiLineString shapefiles and intersects them, then processes the result in various ways. The intersection normally will give me a GeoSeries of points, like this:

n1 = geopandas.read_file('file1.shp')
n2 = geopandas.read_file('file2.shp')

internodes = n1.intersection(n2)
inex = internodes.explode(index_parts=True).droplevel(0)
  
>>>inex
0    POINT (-49.09646 -13.26470)
1    POINT (-57.73859 -10.66091)
2     POINT (-61.73795 -8.95471)
dtype: geometry
<class 'geopandas.geoseries.GeoSeries'>

But sometimes two or more segments from the MultiLineStrings will overlap in their totality, and this will give me a GeoSeries with both Points and LineStrings:

>>>inex
0    LINESTRING (-53.17200 -8.94200, -45.42200 -17.162)
1                          POINT (-59.89331 -11.36051)
dtype: geometry
<class 'geopandas.geoseries.GeoSeries'>

This leads to errors in my program since it expects a GeoSeries with only points.

What's the best method to get a new GeoSeries (or maybe a GeoDataFrame) that contains only the points and coordinates of this GeoSeries? Using the example above, I want to obtain something like this:

>>>inex
0    POINT (-53.17200 -8.94200)
1    POINT (-57.73859 -10.66091)
2     POINT (-59.89331 -11.36051)
dtype: geometry
<class 'geopandas.geoseries.GeoSeries'>

I've tried using GeoSeries.extract_unique_points() and GeoSeries.get_coordinates() as mentioned here and here, but then I get an error message that says 'GeoSeries' object has no attribute 'extract_unique_points' or 'GeoSeries' object has no attribute 'get_coordinates' .

I've also tried converting the GeoSeries to a GeoDataFrame and using extract_xy() from the GemGIS package, but as stated here, currently GeoDataFrames with multiple types of geometries are not supported.

Answer 1

GeoSeries.extract_unique_points() and GeoSeries.get_coordinates() should indeed do the trick.

If you get an error that these functions don't exist, probably you are using an old version of GeoPandas. The current version is 0.14.4.

Use the following command to check the version of GeoPandas and its dependencies:

import geopandas

print(geopandas.show_versions())

Answer 2

You can create a column of all geometry coordinates and explode by it:

import geopandas as gpd

layer1 = gpd.read_file(r"C:\GIS\GIStest\layer1.geojson") #The black lines in the plot
layer2 = gpd.read_file(r"C:\GIS\GIStest\layer2.geojson") #Yellow lines

inter = gpd.overlay(layer1, layer2, how="intersection", keep_geom_type=False) #Blue points and line
# print(inter)
# 0  LINESTRING (13.30732 58.29968, 13.30924 58.30205)
# 1                          POINT (13.31158 58.30220)
# 2                          POINT (13.31282 58.30228)
# 3                          POINT (13.31409 58.29808)

ax = layer1.plot(figsize=(10,10), linewidth=15, color="black", zorder=1)
layer2.plot(ax=ax, linewidth=10, color="yellow", zorder=2)
inter.plot(ax=ax, color="skyblue", zorder=3, linewidth=3)

inter["coords"] = inter.geometry.apply(lambda x: x.coords) #Extract all geometry coordinates
inter = inter.explode("coords") #Explode by the coordinate sequence

#print(inter)
# 0  LINESTRING (13.30732 58.29968, 13.30924 58.30205)    (13.307323135774453, 58.2996779551392)
# 0  LINESTRING (13.30732 58.29968, 13.30924 58.30205)   (13.309244098550504, 58.30204903170543)
# 1                          POINT (13.31158 58.30220)  (13.311582383432421, 58.302200356511705)
# 2                          POINT (13.31282 58.30228)   (13.31282270444919, 58.302280625318566)
# 3                          POINT (13.31409 58.29808)   (13.314094087572387, 58.29808463744689)