added geocoding function

open_buildings/cli.py

@@ -8,7 +8,7 @@
 from open_buildings.download_buildings import download as download_buildings
 from open_buildings.overture.add_columns import process_parquet_files
 from open_buildings.overture.partition import process_db
-from open_buildings.settings import Source
+from settings import Source
 from datetime import datetime, timedelta
 from tabulate import tabulate
 import boto3  # Required for S3 operations
@@ -42,7 +42,8 @@ def handle_comma_separated(ctx, param, value):
 @click.option('-s', '--silent', is_flag=True, default=False, help='Suppress all print outputs.')
 @click.option('--overwrite', default=False, is_flag=True, help='Overwrite the destination file if it already exists.')
 @click.option('--verbose', default=False, is_flag=True, help='Print detailed logs with timestamps.')
-def get_buildings(geojson_input, dst, source, country_iso, silent, overwrite, verbose):
+@click.option('--geocode', default=False, is_flag=True, help='Use city or region name')
+def get_buildings(geojson_input, geocode, dst, source, country_iso, silent, overwrite, verbose):
     """Tool to extract buildings in common geospatial formats from large archives of GeoParquet data online. GeoJSON
     input can be provided as a file or piped in from stdin. If no GeoJSON input is provided, the tool will read from stdin.
 
@@ -74,7 +75,7 @@ def get_buildings(geojson_input, dst, source, country_iso, silent, overwrite, ve
     else:
         geojson_data = json.load(click.get_text_stream('stdin'))
 
-    download_buildings(geojson_data, source=source, generate_sql=False, dst=dst, silent=silent, overwrite=overwrite, verbose=verbose, country_iso=country_iso)
+    download_buildings(geojson_data, generate_sql=False, dst=dst, silent=silent, overwrite=overwrite, verbose=verbose, country_iso=country_iso)
 
 @google.command('benchmark')
 @click.argument('input_path', type=click.Path(exists=True))
@@ -162,6 +163,12 @@ def benchmark(
 @click.option(
     '--overwrite', is_flag=True, help="Whether to overwrite any existing output files."
 )
+@click.option(
+    '--geocode',
+    type=str,
+    is_flag=True, 
+    help="Use city or region name instead of geojson"
+)
 @click.option(
     '--process',
     type=click.Choice(['duckdb', 'pandas', 'ogr']),

open_buildings/download_buildings.py

@@ -1,7 +1,7 @@
 import json
 import click
 from math import tan, cos, log, pi
-from shapely.geometry import shape
+from shapely.geometry import shape, box
 from typing import Dict, Any, Union
 import mercantile 
 import duckdb
@@ -15,7 +15,7 @@
 import subprocess
 from shapely import wkb
 import shutil
-
+import osmnx
 from open_buildings.settings import Source, Format, settings
 
 def geojson_to_quadkey(data: dict) -> str:
@@ -43,6 +43,11 @@ def geojson_to_wkt(data: dict) -> str:
     geometry = shape(data['geometry'])
     return geometry.wkt
 
+def geocode_to_wkt(data: str):
+    location = osmnx.geocode_to_gdf(data)
+    wkt = box(*location.total_bounds)
+    return wkt
+
 def quadkey_to_geojson(quadkey: str) -> dict:
     # Convert the quadkey to tile coordinates
     tile = mercantile.quadkey_to_tile(quadkey)
@@ -81,49 +86,53 @@ def quadkey(geojson_input):
         geojson_data = json.load(click.get_text_stream('stdin'))
 
     result = geojson_to_quadkey(geojson_data)
+    print()
     click.echo(result)
 
 @cli.command()
 @click.argument('geojson_input', type=click.File('r'), required=False)
-def WKT(geojson_input):
-    """Convert GeoJSON to Well Known Text."""
+@click.option('--geocode', type=str, is_flag=True, help='geocode using a city or region')
+def WKT(geojson_input, geocode_input):
+    """Convert GeoJSON to Well Known Text."""          
     if geojson_input:
         geojson_data = json.load(geojson_input)
     else:
-        geojson_data = json.load(click.get_text_stream('stdin'))
-
-    result = geojson_to_wkt(geojson_data)
+        geojson_data = json.load(click.get_text_stream('stdin'))   
+    if geocode_input: 
+        result = geocode_to_wkt(geocode_input)
+    else:
+        result = geojson_to_wkt(geojson_data)
     click.echo(result)
 
 
 @click.command()
 @click.argument('geojson_input', type=click.File('r'), required=False)
 @click.option('--only-quadkey', is_flag=True, help='Include only the quadkey in the WHERE clause.')
 @click.option('--local', is_flag=True, help='Use local path for parquet files instead of the S3 URL.')
-def sql(geojson_input, only_quadkey, local):
-    """Generate an SQL query based on the input GeoJSON."""
+# def sql(geojson_input, only_quadkey, local):
+#     """Generate an SQL query based on the input GeoJSON."""
 
-    # Read the GeoJSON
-    if geojson_input:
-        geojson_data = json.load(geojson_input)
-    else:
-        geojson_data = json.load(click.get_text_stream('stdin'))
-
-    quadkey = geojson_to_quadkey(geojson_data)
-    wkt = geojson_to_wkt(geojson_data)
-
-    # Adjust the path in read_parquet based on the --local flag
-    path = '*.parquet' if local else 's3://us-west-2.opendata.source.coop/cholmes/overture/geoparquet-country-quad-2/*.parquet'
-    base_sql = f"select * from read_parquet('{path}')"
+#     # Read the GeoJSON
+#     if geojson_input:
+#         geojson_data = json.load(geojson_input)
+#     else:
+#         geojson_data = json.load(click.get_text_stream('stdin'))
+
+#     quadkey = geojson_to_quadkey(geojson_data)
+#     wkt = geojson_to_wkt(geojson_data)
+
+#     # Adjust the path in read_parquet based on the --local flag
+#     path = '*.parquet' if local else 's3://us-west-2.opendata.source.coop/cholmes/overture/geoparquet-country-quad-2/*.parquet'
+#     base_sql = f"select * from read_parquet('{path}')"
 
-    # Construct the WHERE clause based on the options
-    where_clause = f"WHERE quadkey LIKE '{quadkey}%'"
-    if not only_quadkey:
-        where_clause += f" AND\nST_Within(ST_GeomFromWKB(geometry), ST_GeomFromText('{wkt}'))"
+#     # Construct the WHERE clause based on the options
+#     where_clause = f"WHERE quadkey LIKE '{quadkey}%'"
+#     if not only_quadkey:
+#         where_clause += f" AND\nST_Within(ST_GeomFromWKB(geometry), ST_GeomFromText('{wkt}'))"
 
-    sql_query = f"{base_sql},\n{where_clause}"
-    full_sql_query = f"COPY ('{sql_query}' TO 'buildings.fgb' WITH (FORMAT GDAL, DRIVER 'FlatGeobuf')"
-    click.echo(full_sql_query) 
+#     sql_query = f"{base_sql},\n{where_clause}"
+#     full_sql_query = f"COPY ('{sql_query}' TO 'buildings.fgb' WITH (FORMAT GDAL, DRIVER 'FlatGeobuf')"
+#     click.echo(full_sql_query) 
 
 @cli.command()
 @click.argument('quadkey_input', type=str)
@@ -132,9 +141,9 @@ def quad2json(quadkey_input):
     result = quadkey_to_geojson(quadkey_input)
     click.echo(json.dumps(result, indent=2))
 
-
 def download(
-        geojson_data: Dict[str, Any], 
+        geojson_data: Dict[str, Any],
+        geocode: Optional[str] = None,
         dst: Union[Path, str] = "buildings.json",
         source: Union[Source, str] = Source.OVERTURE,
         format: Optional[Union[Format, str]] = None, 

requirements.txt

@@ -2,6 +2,7 @@ click
 duckdb
 pandas
 geopandas
+osmnx
 shapely
 openlocationcode
 tabulate