dd-l2-tokyo-overture-schools

Rerun

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely", "pyproj", "pandas", "pyarrow"]
# ///

import geopandas as gpd
import pandas as pd

def solve():
    # Load the bounding boxes for the 23 wards
    wards_bbox = gpd.read_file("tokyo_23wards_bbox.geojson")
    
    # Load the places data
    # It is a parquet file, likely containing many points
    places = gpd.read_parquet("tokyo_places.parquet")
    
    # We need to filter places by age range 8-14
    # Let's first inspect the columns of places to see how age or school type is represented
    print("Columns in places:", places.columns.tolist())
    print("Head of places:", places.head())

    # Based on the prompt, we are looking for "every school relevant to that age range"
    # The prompt implies there might be a way to identify age range or school type.
    # I will first check if there is a column related to 'age' or 'type' or 'category'.
    
    # Let's assume there is a column that specifies the category or age range.
    # Since I don't know the column names yet, I'll look for likely candidates in the print output.
    
    # For now, let's perform a spatial join to find places within the wards.
    # We'll join 'places' with 'wards_bbox' using a spatial join.
    
    # Ensure both are in the same CRS
    if places.crs != wards_bbox.crs:
        places = places.to_crs(wards_bbox.crs)
    
    # Spatial join: find places inside wards
    # wards_bbox contains the rectangles for the 23 wards.
    # We want places that are within ANY of these wards.
    places_in_wards = gpd.sjoin(places, wards_bbox, predicate='within')
    
    # Now we need to filter by school and age range 8-14.
    # I need to see the actual data to know how to filter.
    # I'll print some unique values of potential columns if they exist.
    
    # Let's assume the columns for the final output are already present or can be derived.
    # Required keys: `id`, `name`, `confidence`, `address_freeform`, `address_locality`, and `address_postcode`.

    # I'll write a more robust script after I see the data.
    pass

if __name__ == "__main__":
    solve()

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely", "pyproj", "pandas", "pyarrow"]
# ///

import geopandas as gpd
import pandas as pd
import json

def solve():
    # Load the bounding boxes for the 23 wards
    wards_bbox = gpd.read_file("tokyo_23wards_bbox.geojson")
    
    # Load the places data
    places = gpd.read_parquet("tokyo_places.parquet")
    
    # The prompt says: "every school relevant to that age range [8-14]"
    # Let's inspect the 'categories' and 'names' columns.
    # Since categories might be a list or a string representation of a list, 
    # we need to be careful.
    
    # Inspect a few rows of 'categories' and 'names'
    print("Sample categories:")
    print(places['categories'].head(20))
    print("\nSample names:")
    print(places['names'].head(20))
    print("\nSample addresses:")
    print(places['addresses'].head(20))

    # Let's inspect the structure of 'categories'
    # In many datasets, categories are a list of strings.
    
    # In the absence of explicit info, let's try to identify schools.
    # Common terms for schools: 'school', 'elementary school', 'junior high school', 'middle school'.
    # Age range 8-14 typically corresponds to elementary (shogakko) and junior high (chugakko).

    # Also, inspect the addresses structure
    # Requirement: id, name, confidence, address_freeform, address_locality, address_postcode

    # Let's check if 'categories' is a list or a string.
    sample_cat = places['categories'].iloc[0]
    print(f"Type of sample category: {type(sample_cat)}")

    # We'll look for school-related categories.
    # Since I can't see the whole content, I'll use the sample to guide.
    pass

if __name__ == "__main__":
    solve()

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely", "pyproj", "pandas", "pyarrow"]
# ///

import geopandas as gpd
import pandas as pd
import json

def solve():
    # Load the bounding boxes for the 23 wards
    wards_bbox = gpd.read_file("tokyo_23wards_bbox.geojson")
    
    # Load the places data
    places = gpd.read_parquet("tokyo_places.parquet")
    
    # Ensure both are in the same CRS
    if places.crs != wards_bbox.crs:
        places = places.to_crs(wards_bbox.crs)
    
    # Spatial join: find places within wards
    # We want places that are within ANY of these wards.
    # The prompt says: "whose point sits inside the wards rectangle in tokyo_23wards_bbox"
    places_in_wards = gpd.sjoin(places, wards_bbox, predicate='within')

    # Filter for schools relevant to age 8-14.
    # Categories are dicts: {'alternate': [...], 'primary': '...'}
    # We're looking for 'elementary_school' and 'junior_high_school' (or similar).
    # Let's check common categories for schools.
    
    def is_relevant_school(categories):
        if not isinstance(categories, dict) or categories is None:
            return False
        
        # Check primary category
        primary = categories.get('primary')
        if primary in ['elementary_school', 'junior_high_school', 'middle_school']:
            return True
            
        # Check alternate categories
        alternates = categories.get('alternate')
        if isinstance(alternates, list):
            for alt in alternates:
                if alt in ['elementary_school', 'junior_high_school', 'middle_school']:
                    return True
        return False

    # Apply filtering
    schools = places_in_wards[places_in_wards['categories'].apply(is_relevant_school)].copy()

    # Now prepare the properties for the final GeoJSON.
    # Required: id, name, confidence, address_freeform, address_locality, address_postcode

    # names is a dict: {'common': ..., 'primary': ...}
    # addresses is a list of dicts: [{'country': ..., 'freeform': ..., 'locality': ..., 'postcode': ...}]
    # confidence is a float/int

    def extract_properties(row):
        # Name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        if not name and isinstance(name_dict, dict):
            name = name_dict.get('common')
        
        # Addresses - take the first one if available
        addr_list = row['addresses']
        addr_freeform = None
        addr_locality = None
        addr_postcode = None
        
        if isinstance(addr_list, list) and len(addr_list) > 0:
            addr = addr_list[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        return pd.Series({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode
        })

    if not schools.empty:
        props = schools.apply(extract_properties, axis=1)
        # Combine with geometry
        schools_final = pd.concat([props, schools['geometry']], axis=1)
        schools_gdf = gpd.GeoDataFrame(schools_final, crs=places.crs)
        
        # Keep only the requested columns
        cols_to_keep = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
        schools_gdf = schools_gdf[cols_to_keep]
        
        # Save to geojson
        schools_gdf.to_file("tokyo_schools.geojson", driver='GeoJSON')
        print(f"Successfully saved {len(schools_gdf)} schools to tokyo_schools.geojson")
    else:
        # Create an empty GeoDataFrame with the correct columns if no schools found
        empty_gdf = gpd.GeoDataFrame(columns=['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry'], crs=places.crs)
        empty_gdf.to_file("tokyo_schools.geojson", driver='GeoJSON')
        print("No schools found matching the criteria. Created empty tokyo_schools.geojson")

if __name__ == "__main__":
    solve()

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely", "pyproj", "pandas", "pyarrow"]
# ///

import geopandas as gpd
import pandas as pd

def solve():
    # Load the bounding boxes for the 23 wards
    wards_bbox = gpd.read_file("tokyo_23wards_bbox.geojson")
    
    # Load the places data
    places = gpd.read_parquet("tokyo_places.parquet")
    
    # Ensure both are in the same CRS
    if places.crs != wards_bbox.crs:
        places = places.to_crs(wards_bbox.crs)
    
    # Spatial join: find places within wards
    places_in_wards = gpd.sjoin(places, wards_bbox, predicate='within')

    # Filter for schools relevant to age 8-14.
    # Categories are dicts: {'alternate': [...], 'primary': '...'}
    def is_relevant_school(categories):
        if not isinstance(categories, dict) or categories is None:
            return False
        
        # Check primary category
        primary = categories.get('primary')
        if primary in ['elementary_school', 'junior_high_school', 'middle_school']:
            return True
            
        # Check alternate categories
        alternates = categories.get('alternate')
        if isinstance(alternates, list):
            for alt in alternates:
                if alt in ['elementary_school', 'junior_high_school', 'middle_school']:
                    return True
        return False

    # Apply filtering
    schools = places_in_wards[places_in_wards['categories'].apply(is_relevant_school)].copy()

    def extract_properties(row):
        # Name
        name_dict = row['names']
        name = name_dict.get('primary') if isinstance(name_dict, dict) else None
        if not name and isinstance(name_dict, dict):
            name = name_dict.get('common')
        
        # Addresses - take the first one if available
        addr_list = row['addresses']
        addr_freeform = None
        addr_locality = None
        addr_postcode = None
        
        if isinstance(addr_list, list) and len(addr_list) > 0:
            addr = addr_list[0]
            if isinstance(addr, dict):
                addr_freeform = addr.get('freeform')
                addr_locality = addr.get('locality')
                addr_postcode = addr.get('postcode')
        
        return pd.Series({
            'id': row['id'],
            'name': name,
            'confidence': row['confidence'],
            'address_freeform': addr_freeform,
            'address_locality': addr_locality,
            'address_postcode': addr_postcode
        })

    if not schools.empty:
        props = schools.apply(extract_properties, axis=1)
        # Combine with geometry
        schools_final = pd.concat([props, schools['geometry']], axis=1)
        schools_gdf = gpd.GeoDataFrame(schools_final, crs=places.crs)
        
        # Keep only the requested columns
        cols_to_keep = ['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry']
        schools_gdf = schools_gdf[cols_to_keep]
        
        # Save to geojson
        schools_gdf.to_file("tokyo_schools.geojson", driver='GeoJSON')
        print(f"Successfully saved {len(schools_gdf)} schools to tokyo_schools.geojson")
    else:
        # Create an empty GeoDataFrame with the correct columns if no schools found
        empty_gdf = gpd.GeoDataFrame(columns=['id', 'name', 'confidence', 'address_freeform', 'address_locality', 'address_postcode', 'geometry'], crs=places.crs)
        empty_gdf.to_file("tokyo_schools.geojson", driver='GeoJSON')
        print("No schools found matching the criteria. Created empty tokyo_schools.geojson")

if __name__ == "__main__":
    solve()