fio-l2-cairo-mixedgeom-split

Rerun

import geopandas as gpd

# Load the data
gdf = gpd.read_file('heritage_sites.geojson')

print("Initial Data Info:")
print(gdf.info())
print("\nGeometry Types found:")
print(gdf.geom_type.unique())
print("\nFirst few rows:")
print(gdf.head())

import geopandas as gpd

# Load the data
gdf = gpd.read_file('heritage_sites.geojson')

# Egypt National Grid (Egypt Red Belt / Egypt TM)
# EPSG:22992 is Egypt Red Belt
# However, usually, standard is EPSG:32636 (WGS 84 / UTM zone 36N) 
# or WGS 84 / UTM zone 35N depending on location.
# Let's search for a more standard Egypt National Grid or just use a common one.
# According to some sources EPSG:22992 is Egypt Red Belt.
# Let's try to find the most common one if not specified.
# Given the prompt says "Egypt's national grid", I will use EPSG:22992.
# If it's not available, I'll fall back to a standard UTM for Egypt.

target_crs = "EPSG:22992"

# To be safe, check if EPSG:22992 is available or just use UTM 36N (32636) which is widely used.
# Actually, I'll just try 22992.

# 1. Ensure all geometries are single-part
gdf['geometry'] = gdf.geometry.apply(lambda x: x if x.geom_type in ['Point', 'LineString', 'Polygon'] else x.explode(index_parts=True))

# Note: explode() might return a GeoSeries, need to make sure it stays as a GeoDataFrame or handle it.
# Let's refine the explosion process.

def ensure_single_part(gdf):
    exploded_list = []
    for _, row in gdf.iterrows():
        geom = row.geometry
        if geom.geom_type in ['Point', 'LineString', 'Polygon']:
            exploded_list.append(row)
        else:
            # Explode Multi-part
            exploded_geom = geom.explode(index_parts=False)
            for part in exploded_geom:
                new_row = row.copy()
                new_row.geometry = part
                exploded_list.append(new_row)
    return gpd.GeoDataFrame(exploded_list, crs=gdf.crs)

gdf_single = ensure_single_part(gdf)

# Re-project to Egypt's national grid
gdf_single = gdf_single.to_crs(target_crs)

# 2. Split into three layers
points = gdf_single[gdf_single.geometry.geom_type == 'Point']
lines = gdf_single[gdf_single.geometry.geom_type == 'LineString']
polygons = gdf_single[gdf_single.geometry.geom_type == 'Polygon']

# 3. Save to GPKG
output_file = 'heritage.gpkg'
with pd.io.gpd.GPDFile() as driver: # This is not how geopandas works for multiple layers
    pass

# Use the correct way to write multiple layers in geopandas
import pandas as pd # just in case

# We use the engine='pyogrio' if available or just standard driver
# Geopandas write_to_file supports 'layer' argument

points.to_file(output_file, layer='points', driver='GPKG')
lines.to_file(output_file, layer='lines', driver='GPKG')
polygons.to_file(output_file, layer='polygons', driver='GPKG')

print(f"Successfully created {output_file} with layers: points, lines, polygons")
print(f"Projected CRS: {gdf_single.crs}")

import geopandas as gpd

# Load the data
gdf = gpd.read_file('heritage_sites.geojson')

# Egypt National Grid (Egypt Red Belt / Egypt TM)
# EPSG:22992 is Egypt Red Belt.
target_crs = "EPSG:22992"

def ensure_single_part(gdf):
    # explode() on a GeoDataFrame is the easiest way to turn Multi-geometries into single-part
    # It handles all geom types and preserves columns.
    # index_parts=False means it doesn't create a multi-index.
    exploded_gdf = gdf.explode(index_parts=False)
    return exploded_gdf

# 1. Ensure all geometries are single-part
gdf_single = ensure_single_part(gdf)

# 2. Re-project to Egypt's national grid
gdf_single = gdf_single.to_crs(target_crs)

# 3. Split into three layers based on geometry type
# Note: explode() might leave us with Point, LineString, Polygon
# We need to make sure we categorize them correctly.
points = gdf_single[gdf_single.geometry.geom_type == 'Point']
lines = gdf_single[gdf_single.geometry.geom_type == 'LineString']
polygons = gdf_single[gdf_single.geometry.geom_type == 'Polygon']

# 4. Save to GPKG
output_file = 'heritage.gpkg'

# Writing layers
points.to_file(output_file, layer='points', driver='GPKG')
lines.to_file(output_file, layer='lines', driver='GPKG')
polygons.to_file(output_file, layer='polygons', driver='GPKG')

print(f"Successfully created {output_file} with layers: points, lines, polygons")
print(f"Projected CRS: {gdf_single.crs}")
print(f"Counts - Points: {len(points)}, Lines: {len(lines)}, Polygons: {len(polygons)}")