add cli test for --location, fix some minor things, change dst from p…

…ositional params to kwarg, rebasing

.gitignore

@@ -4,6 +4,7 @@ private/
 *.py[cod]
 *$py.class
 
+cache
 issues.txt
 
 # C extensions

open_buildings/cli.py

@@ -39,19 +39,19 @@ def handle_comma_separated(ctx, param, value):
 def geocode(data: str):
     location = osmnx.geocode_to_gdf(data)
     wkt = box(*location.total_bounds)
-    geojson = json.dumps(mapping(wkt))
+    geojson = json.loads(json.dumps({"type": "Feature", "geometry": mapping(wkt)})) # turn geom tuple into list by (de-)serialising
     return geojson
 
 @main.command(name="get_buildings")
 @click.argument('geojson_input', type=click.File('r'), required=False)
-@click.argument('dst', type=str, default="buildings.json")
+@click.option('--dst', type=str, default="buildings.json", help='The path to write the output to. Can be a directory or file.')
+@click.option('--location', type=str, default=None, help='Use city or region name')
 @click.option('--source', default="overture", type=click.Choice(['google', 'overture']), help='Dataset to query, defaults to Overture')
 @click.option('--country_iso', type=str, default=None, help='A 2 character country ISO code to filter the data by.')
 @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.')
-@click.option('--location', type=str, default=None, help='Use city or region name')
-def get_buildings(geojson_input, location, dst, source, country_iso, silent, overwrite, verbose):
+@click.option('-v', '--verbose', default=False, is_flag=True, help='Print detailed logs with timestamps.')
+def get_buildings(geojson_input, dst, location, 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.
 
@@ -80,12 +80,12 @@ def get_buildings(geojson_input, location, dst, source, country_iso, silent, ove
 
     if geojson_input:
         geojson_data = json.load(geojson_input)
-    elif geocode:
+    elif location:
         geojson_data = geocode(location)
     else:
         geojson_data = json.load(click.get_text_stream('stdin'))
 
-    download_buildings(geojson_data, location, 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))

open_buildings/download_buildings.py

@@ -44,12 +44,6 @@ def geojson_to_wkt(data: dict) -> str:
     geometry = shape(data['geometry'])
     return geometry.wkt
 
-def geocode(data: str):
-    location = osmnx.geocode_to_gdf(data)
-    wkt = box(*location.total_bounds)
-    g2 = geojson.Feature(geometry=wkt)
-    return (g2)
-
 def quadkey_to_geojson(quadkey: str) -> dict:
     # Convert the quadkey to tile coordinates
     tile = mercantile.quadkey_to_tile(quadkey)
@@ -91,13 +85,10 @@ def quadkey(geojson_input):
 
 @cli.command()
 @click.argument('geojson_input', type=click.File('r'), required=False)
-@click.option('--location', type=str, help='geocode using a city or region')
-def WKT(geojson_input, location):
+def WKT(geojson_input):
     """Convert GeoJSON to Well Known Text."""
     if geojson_input:
         result = json.load(geojson_input)
-    elif location:
-        result = geocode(location)
     else:
         geojson_data = json.load(click.get_text_stream('stdin'))  
         result = geojson_to_wkt(geojson_data) 
@@ -142,7 +133,6 @@ def quad2json(quadkey_input):
 
 def download(
         geojson_data: Dict[str, Any],
-        location: Optional[str] = None,
         dst: Union[Path, str] = "buildings.json",
         source: Union[Source, str] = Source.OVERTURE,
         format: Optional[Union[Format, str]] = None, 

tests/test_open_buildings.py

@@ -6,7 +6,8 @@
 import re
 import subprocess
 
-from open_buildings.download_buildings import download, geojson_to_wkt, geojson_to_quadkey, quadkey_to_geojson, geocode
+from open_buildings.download_buildings import download, geojson_to_wkt, geojson_to_quadkey, quadkey_to_geojson
+from open_buildings.cli import geocode
 from open_buildings.settings import Source, Format, settings
 
 ###########################################################################
@@ -67,8 +68,7 @@ def test_quadkey_to_geojson():
 
 def test_geocode():
     """ Tests geocode() using a pre-established true value. """
-
-    assert json.loads(geocode('plymouth')) == {"type": "Polygon", "coordinates": [[[-4.0196056, 50.3327426], [-4.0196056, 50.4441737], [-4.2055324, 50.4441737], [-4.2055324, 50.3327426], [-4.0196056, 50.3327426]]]}
+    assert geocode('plymouth') == {"type": "Feature", "geometry": {"type": "Polygon", "coordinates": [[[-4.0196056, 50.3327426], [-4.0196056, 50.4441737], [-4.2055324, 50.4441737], [-4.2055324, 50.3327426], [-4.0196056, 50.3327426]]]}}
 
 @pytest.mark.integration
 @pytest.mark.flaky(reruns=NUM_RERUNS)
@@ -150,7 +150,7 @@ def test_cli_get_buildings_from_file_to_directory(aoi: Dict[str, Any], tmp_path:
     input_path = tmp_path.joinpath("input.json")
     with open(input_path, "w") as f:
         json.dump(aoi, f)
-    subprocess.run(["ob", "get_buildings", str(input_path), str(tmp_path), "--country_iso", "SC"])
+    subprocess.run(["ob", "get_buildings", str(input_path), "--dst", str(tmp_path), "--country_iso", "SC"], check=True)
     output_path = tmp_path.joinpath("buildings.json") # default file name
     assert os.path.exists(output_path)
     assert os.path.getsize(output_path) != 0
@@ -164,7 +164,7 @@ def test_cli_get_buildings_from_stdin_to_directory(aoi: Dict[str, Any], tmp_path
     Verifies that a log message with timestamp gets written to stdout. 
     """
     # we can't use pipes (e.g. f"echo {json.dumps(aoi)} | ...") in subprocess.run, instead we pass the json as stdin using the input/text arguments,
-    process = subprocess.run([ "ob", "get_buildings", "-", str(tmp_path), "--country_iso", "SC"], input=json.dumps(aoi), text=True,check=True, capture_output=True)
+    process = subprocess.run([ "ob", "get_buildings", "-", "--dst", str(tmp_path), "--country_iso", "SC"], input=json.dumps(aoi), text=True, check=True, capture_output=True)
     dt_regex = re.compile(r"^\[[0-9]{4}(-[0-9]{2}){2} ([0-9]{2}:){2}[0-9]{2}\] ") # match timestamp format e.g. "[2023-10-18 19:08:24]"
     assert dt_regex.search(process.stdout) # ensure that stdout contains at least one timestamped message
     output_path = tmp_path.joinpath("buildings.json") # default file name
@@ -180,7 +180,7 @@ def test_cli_get_buildings_from_stdin_to_file_silent(aoi: Dict[str, Any], tmp_pa
     """
     output_path = tmp_path.joinpath("test123.json")
     # we can't use pipes (e.g. f"echo {json.dumps(aoi)} | ...") in subprocess.run, instead we pass the json as stdin using the input/text arguments,
-    process = subprocess.run(["ob", "get_buildings", "-", str(output_path), "--silent", "--country_iso", "SC"], input=json.dumps(aoi), text=True, check=True, capture_output=True)
+    process = subprocess.run(["ob", "get_buildings", "-", "--dst", str(output_path), "--silent", "--country_iso", "SC"], input=json.dumps(aoi), text=True, check=True, capture_output=True)
     assert process.stdout == "" # assert that nothing gets printed to stdout
     assert process.stderr == "" # assert that nothing gets printed to stdout
     assert os.path.exists(output_path)
@@ -198,8 +198,19 @@ def test_cli_get_buildings_from_stdin_to_file_overwrite_false(aoi: Dict[str, Any
     with open(output_path, "w") as f:
         f.write("Foo bar")
     # we can't use pipes (e.g. f"echo {json.dumps(aoi)} | ...") in subprocess.run, instead we pass the json as stdin using the input/text arguments,
-    process = subprocess.run(["ob", "get_buildings", "-", str(output_path), "--country_iso", "SC"], input=json.dumps(aoi), text=True, check=True, capture_output=True)
+    process = subprocess.run(["ob", "get_buildings", "-", "--dst", str(output_path), "--country_iso", "SC"], input=json.dumps(aoi), text=True, check=True, capture_output=True)
     assert os.path.exists(output_path)
     with open(output_path, "r") as f:
         assert f.read() == "Foo bar" # verify that the file still has the same content as before
-    assert "exists" in process.stdout # verify that the user has been warned about the existing file
+    assert "exists" in process.stdout # verify that the user has been warned about the existing file
+
+@pytest.mark.integration
+@pytest.mark.flaky(reruns=NUM_RERUNS)
+def test_cli_get_buildings_geocode(tmp_path: Path):
+    """ 
+    Tests the geocoding functionality, implemented as the argument "location".
+    """
+    output_path = tmp_path.joinpath("geocode_test.json")
+    subprocess.run(["ob", "get_buildings", "--dst", str(output_path), "--location", "oxford uk", "--country_iso", "UK"], check=True)
+    assert os.path.exists(output_path)
+    assert os.path.getsize(output_path) != 0