Add support for bbox filtering when reading generated parquet files

[songololo]

When opening an overturemaps-py generated geoparquet file from geopandas it isn't possible to filter by bounding box:

For example, this:

gpd.read_parquet(str(input_path.resolve()), bbox=bounds_buff.bounds)

Throws this:

ValueError: Specifying 'bbox' not supported for this Parquet file (it should either have a bbox covering column or use 'point' encoding).

It would be extremely useful if generated geoparquet files supported this functionality to assist with targeted gpd.read_parquet reads for a given bbox.

[songololo]

Pull request proposed for fixing this behaviour: #41