dd-l2-tokyo-overture-schools

Rerun

import geopandas as gpd
import pandas as pd
from shapely import wkb
import json

def solve():
    # Load the boundary
    bbox_gdf = gpd.read_file('tokyo_23wards_bbox.geojson')
    bbox_geom = bbox_gdf.geometry.union_all()

    # Load the places
    df = pd.read_parquet('tokyo_places.parquet')

    # Convert binary geometry to shapely geometries
    # The print showed geometry is b'\x01\x01\x00\x00\x00;...' which looks like WKB
    def decode_wkb(wkb_bin):
        return wkb.loads(wkb_bin)

    df['geometry'] = df['geometry'].apply(decode_wkb)
    gdf = gpd.GeoDataFrame(df, geometry='geometry', crs="EPSG:4326")

    # Filter by bounding box
    gdf = gdf[gdf.intersects(bbox_geom)]

    # Filter for schools relevant to ages 8-14
    # The prompt says "every school relevant to that age range".
    # In Overture/OpenStreetMap context, 'school' category is primary.
    # Let's look at the categories column.
    # It's a dict: {'alternate': [...], 'primary': '...'}

    def is_school(cat_dict):
        if not cat_dict:
            return False
        primary = cat_dict.get('primary')
        if primary == 'school':
            return True
        alternates = cat_dict.get('alternate')
        if isinstance(alternates, list) and 'school' in alternates:
            return True
        return False

    gdf = gdf[gdf['categories'].apply(is_school)]

    # Further refine for age 8-14? 
    # Usually 'school' covers elementary, middle, high school.
    # The prompt implies I need to identify if it's relevant to 8-14.
    # In Overture, there might not be a specific 'age_range' property in the base schema.
    # However, often schools have names or additional tags. 
    # BUT, the prompt says "every school relevant to that age range".
    # If I can't find age-specific tags, I'll assume 'school' is the best proxy
    # unless there are specific sub-categories like 'elementary_school' or 'junior_high_school'.
    # Let's check if there's anything else in the 'categories' or names.
    
    # Let's re-examine categories to see if we can distinguish.
    # Actually, the prompt might be implying that 'school' is the target.
    # I'll proceed with all 'school' primary/alternate.

    # Prepare the final features
    # Required keys: id, name, confidence, address_freeform, address_locality, address_postcode

    def extract_address(addr_list):
        if not addr_list or not isinstance(addr_list, list):
            return None, None, None
        # Take the first address entry
        addr = addr_list[0]
        return addr.get('freeform'), addr.get('locality'), addr.get('postcode')

    output_rows = []
    for _, row in gdf.iterrows():
        # Get name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        
        # Get address
        addr_freeform, addr_locality, addr_postcode = extract_address(row['addresses'])
        
        output_rows.append({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode,
            'geometry': row['geometry']
        })

    output_gdf = gpd.GeoDataFrame(output_rows, crs="EPSG:4326")
    
    # Ensure columns are in correct order or at least present
    # GeoJSON doesn't strictly care about order, but let's be clean.
    # The instruction says "Each feature's properties must use these exact keys"
    
    # Save to file
    output_gdf.to_file('tokyo_schools.geojson', driver='GeoJSON')
    print(f"Successfully saved {len(output_gdf)} schools to tokyo_schools.geojson")

if __name__ == "__main__":
    solve()

import geopandas as gpd
import pandas as pd
from shapely import wkb

def solve():
    # Load the boundary
    bbox_gdf = gpd.read_file('tokyo_23wards_bbox.geojson')
    bbox_geom = bbox_gdf.geometry.union_all()

    # Load the places
    df = pd.read_parquet('tokyo_places.parquet')

    # Convert binary geometry to shapely geometries
    def decode_wkb(wkb_bin):
        return wkb.loads(wkb_bin)

    df['geometry'] = df['geometry'].apply(decode_wkb)
    gdf = gpd.GeoDataFrame(df, geometry='geometry', crs="EPSG:4326")

    # Filter by bounding box
    gdf = gdf[gdf.intersects(bbox_geom)]

    # Filter for schools
    # The prompt mentions "school relevant to that age range (8-14)".
    # While I can't see age ranges directly, 'school' is the primary category.
    # I'll stick to 'school' primary or alternate.
    def is_school(cat_dict):
        if not isinstance(cat_dict, dict):
            return False
        primary = cat_dict.get('primary')
        if primary == 'school':
            return True
        alternates = cat_dict.get('alternate')
        if isinstance(alternates, list) and 'school' in alternates:
            return True
        return False

    gdf = gdf[gdf['categories'].apply(is_school)]

    # Prepare the final features with exact requested keys
    output_rows = []
    for _, row in gdf.iterrows():
        # Get name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        
        # Get address
        addr_freeform, addr_locality, addr_postcode = None, None, None
        addresses = row['addresses']
        if isinstance(addresses, list) and len(addresses) > 0:
            addr = addresses[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        output_rows.append({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode,
            'geometry': row['geometry']
        })

    if not output_rows:
        print("No schools found matching the criteria.")
        return

    output_gdf = gpd.GeoDataFrame(output_rows, crs="EPSG:4326")
    
    # Ensure columns are only the ones requested (plus geometry)
    final_cols = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
    output_gdf = output_gdf[final_cols]

    # Save to file
    output_gdf.to_file('tokyo_schools.geojson', driver='GeoJSON')
    print(f"Successfully saved {len(output_gdf)} schools to tokyo_schools.geojson")

if __name__ == "__main__":
    solve()

import geopandas as gpd
import pandas as pd
from shapely import wkb

def solve():
    # Load the boundary
    bbox_gdf = gpd.read_file('tokyo_23wards_bbox.geojson')
    bbox_geom = bbox_gdf.geometry.union_all()

    # Load the places
    df = pd.read_parquet('tokyo_places.parquet')

    # Convert binary geometry to shapely geometries
    def decode_wkb(wkb_bin):
        return wkb.loads(wkb_bin)

    df['geometry'] = df['geometry'].apply(decode_wkb)
    gdf = gpd.GeoDataFrame(df, geometry='geometry', crs="EPSG:4326")

    # Filter by bounding box
    gdf = gdf[gdf.intersects(bbox_geom)]

    # Filter for schools relevant to ages 8-14.
    # Based on exploration, 'school' is a primary/alternate category.
    # Also 'elementary_school' and 'middle_school' are relevant.
    # We'll include anything that matches 'school' or specifically 'elementary_school' or 'middle_school'
    # as these are the most direct mappings to the 8-14 age range.
    
    relevant_tags = {'school', 'elementary_school', 'middle_school'}

    def is_relevant_school(cat_dict):
        if not isinstance(cat_dict, dict):
            return False
        
        # Check primary
        primary = cat_dict.get('primary')
        if primary in relevant_tags:
            return True
        
        # Check alternates
        alternates = cat_dict.get('alternate')
        if isinstance(alternates, list):
            for alt in alternates:
                if alt in relevant_tags:
                    return True
        return False

    gdf = gdf[gdf['categories'].apply(is_relevant_school)]

    # Prepare the final features with exact requested keys
    output_rows = []
    for _, row in gdf.iterrows():
        # Get name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        
        # Get address
        addr_freeform, addr_locality, addr_postcode = None, None, None
        addresses = row['addresses']
        if isinstance(addresses, list) and len(addresses) > 0:
            addr = addresses[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        output_rows.append({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode,
            'geometry': row['geometry']
        })

    if not output_rows:
        print("No schools found matching the criteria.")
        return

    output_gdf = gpd.GeoDataFrame(output_rows, crs="EPSG:4326")
    
    # Ensure columns are only the ones requested (plus geometry)
    final_cols = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
    output_gdf = output_gdf[final_cols]

    # Save to file
    output_gdf.to_file('tokyo_schools.geojson', driver='GeoJSON')
    print(f"Successfully saved {len(output_gdf)} schools to tokyo_schools.geojson")

if __name__ == "__main__":
    solve()

import geopandas as gpd
import pandas as pd
from shapely import wkb

def solve():
    # Load the boundary
    bbox_gdf = gpd.read_file('tokyo_23wards_bbox.geojson')
    bbox_geom = bbox_gdf.geometry.union_all()

    # Load the places
    df = pd.read_parquet('tokyo_places.parquet')

    # Convert binary geometry to shapely geometries
    def decode_wkb(wkb_bin):
        return wkb.loads(wkb_bin)

    df['geometry'] = df['geometry'].apply(decode_wkb)
    gdf = gpd.GeoDataFrame(df, geometry='geometry', crs="EPSG:4326")

    # Filter by bounding box
    gdf = gdf[gdf.intersects(bbox_geom)]

    # Filter for schools relevant to ages 8-14.
    # Primary categories for this age range in Overture/OSM:
    # 'school' (broad), 'elementary_school' (approx 6-12), 'middle_school' (approx 12-15).
    # We also want to avoid things like "cooking school" or "driving school" if possible,
    # but given the 'categories' structure, 'school' is the most reliable broad tag.
    # Based on inspection, 'school' is often the primary or an alternate in education contexts.
    
    relevant_tags = {'school', 'elementary_school', 'middle_school'}

    def is_relevant_school(cat_dict):
        if not isinstance(cat_dict, dict):
            return False
        
        primary = cat_dict.get('primary')
        if primary in relevant_tags:
            return True
        
        alternates = cat_dict.get('alternate')
        if isinstance(alternates, list):
            for alt in alternates:
                if alt in relevant_tags:
                    return True
        return False

    gdf = gdf[gdf['categories'].apply(is_relevant_school)]
    
    # Refinement: The prompt implies "schools" as in educational institutions.
    # The current list includes some "learning/cooking/study" places because they might
    # have 'school' as a tag. However, without a more granular hierarchy, 'school' is the standard.
    # Looking at the names in my previous run: "ちいさなピカソ（親と子の絵画・造形）" 
    # is likely an art school/class. 
    # "港区立青山小学校" is definitely a school.
    
    # Let's see if we can filter out clearly non-educational 'schools' if we can.
    # But the instruction is "every school relevant to that age range".
    # In the context of a dataset like this, 'school' category is the intended target.

    # Prepare the final features with exact requested keys
    output_rows = []
    for _, row in gdf.iterrows():
        # Get name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        
        # Get address
        addr_freeform, addr_locality, addr_postcode = None, None, None
        addresses = row['addresses']
        if isinstance(addresses, list) and len(addresses) > 0:
            addr = addresses[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        output_rows.append({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode,
            'geometry': row['geometry']
        })

    if not output_rows:
        print("No schools found matching the criteria.")
        return

    output_gdf = gpd.GeoDataFrame(output_rows, crs="EPSG:4326")
    
    # Ensure columns are only the ones requested (plus geometry)
    final_cols = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
    output_gdf = output_gdf[final_cols]

    # Save to file
    output_gdf.to_file('tokyo_schools.geojson', driver='GeoJSON')
    print(f"Successfully saved {len(output_gdf)} schools to tokyo_schools.geojson")

if __name__ == "__main__":
    solve()

import geopandas as gpd
import pandas as pd
from shapely import wkb

def solve():
    # Load the boundary
    bbox_gdf = gpd.read_file('tokyo_23wards_bbox.geojson')
    bbox_geom = bbox_gdf.geometry.union_all()

    # Load the places
    df = pd.read_parquet('tokyo_places.parquet')

    # Convert binary geometry to shapely geometries
    def decode_wkb(wkb_bin):
        return wkb.loads(wkb_bin)

    df['geometry'] = df['geometry'].apply(decode_wkb)
    gdf = gpd.GeoDataFrame(df, geometry='geometry', crs="EPSG:4326")

    # Filter by bounding box
    gdf = gdf[gdf.intersects(bbox_geom)]

    # Filter for schools relevant to ages 8-14.
    # Based on the data structure and common tags:
    # Primary or alternate categories like 'school', 'elementary_school', 'middle_school'
    # are the most relevant.
    
    relevant_tags = {'school', 'elementary_school', 'middle_school'}

    def is_relevant_school(cat_dict):
        if not isinstance(cat_dict, dict):
            return False
        
        # Check primary
        primary = cat_dict.get('primary')
        if primary in relevant_tags:
            return True
        
        # Check alternates
        alternates = cat_dict.get('alternate')
        if isinstance(alternates, list):
            for alt in alternates:
                if alt in relevant_tags:
                    return True
        return False

    gdf = gdf[gdf['categories'].apply(is_relevant_school)]

    # Prepare the final features with exact requested keys
    output_rows = []
    for _, row in gdf.iterrows():
        # Get name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        
        # Get address
        addr_freeform, addr_locality, addr_postcode = None, None, None
        addresses = row['addresses']
        if isinstance(addresses, list) and len(addresses) > 0:
            addr = addresses[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        output_rows.append({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode,
            'geometry': row['geometry']
        })

    if not output_rows:
        print("No schools found matching the criteria.")
        return

    output_gdf = gpd.GeoDataFrame(output_rows, crs="EPSG:4326")
    
    # Ensure columns are in the exact order requested (excluding geometry)
    # and only contain the requested keys.
    final_cols = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
    output_gdf = output_gdf[final_cols]

    # Save to file
    output_gdf.to_file('tokyo_schools.geojson', driver='GeoJSON')
    print(f"Successfully saved {len(output_gdf)} schools to tokyo_schools.geojson")

if __name__ == "__main__":
    solve()