Personally I think I'd prefer a separate method, like toGeoJSON.gpxMetadata() -> metadata object rather than putting properties on FeatureCollection. I still think that putting properties on a FeatureCollection is technically allowed but practically dangerous: as soon as you send that GeoJSON through a transformation step like Turf.js, the extra properties will be removed. They won't be accessible via most GeoJSON editors or through map libraries.

I think it just makes more sense to have a separate process to get metadata, and if people strongly want to combine that metadata with their FeatureCollection object, they can do so with { ...featureCollection, ...metadata }.

Based on: https://gis.stackexchange.com/a/415412

You can insert a Polygon feature without any coordinates and set your properties as you like:

{
 "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "geometry": {
        "type": "Polygon",
        "coordinates": [          
        ]
      },
      "properties": {
        "description": "This is the geometry for..."
      }
    },
    {
      // Other features
    }
  ]
}

Maybe something like this would be even more easier to store and therefore portable (also when it comes to snapshot testing):

{
  "type": "FeatureCollection",
  "features": [
    {
      "type": "Feature",
      "geometry": {
        "type": "Polygon",
        "coordinates": [ ]     // eventually populated by https://www.topografix.com/GPX/1/1/#type_boundsType
      },
      "properties": {
        "_gpxType": "metadata" // everything else related to https://www.topografix.com/GPX/1/1/#type_metadataType
        "name": "...",
        "desc": "...",
        "author": {
            "name": "...",
            "email": "",
            "link": {
               "href": "...",
               "text": "...",
               "type": "...",
            }
        },
        "copyright": {
          "author": "...",
          "year": "...",
          "license": "..."
        },
        "time": "...",
        "keywords": "...",
        "extensions": ??
    },
    {
      // Other features
    }
  ]
}

Not really sure if these property names make sense:

• "_gpxType": "metadata" --> similarly to what was already done to distinguish <rte> and <trk> tags
• "extensions": ?? --> probably a more complex structure (right now, I don't remember what's the current gpx extensions specfication..)

but that's just to give you a quick idea.

Related info:

• togeojson/lib/__snapshots__/index.test.ts.snap

  Lines 12768 to 12811 in d395dde

  	exports[`toGeoJSON > ground_overlay.kml 1`] = `
  	{
  	"features": [
  	{
  	"geometry": {
  	"coordinates": [
  	[
  	[
  	14.600668902543442,
  	37.91801270483731,
  	],
  	[
  	15.35770894852841,
  	37.92006947469352,
  	],
  	[
  	15.358941332345658,
  	37.466463107960706,
  	],
  	[
  	14.60190128636069,
  	37.464406338104496,
  	],
  	[
  	14.600668902543442,
  	37.91801270483731,
  	],
  	],
  	],
  	"type": "Polygon",
  	},
  	"properties": {
  	"@geometry-type": "groundoverlay",
  	"description": "Overlay shows Mount Etna erupting
  	on July 13th, 2001.",
  	"icon": "https://developers.google.com/kml/documentation/images/etna.jpg",
  	"name": "Large-scale overlay on terrain",
  	},
  	"type": "Feature",
  	},
  	],
  	"type": "FeatureCollection",
  	}
  	`;

• togeojson/lib/gpx.ts

  Lines 181 to 191 in d395dde

  	/**
  	*
  	* Convert a GPX document to GeoJSON. The first argument, `doc`, must be a GPX
  	* document as an XML DOM - not as a string. You can get this using jQuery's default
  	* `.ajax` function or using a bare XMLHttpRequest with the `.response` property
  	* holding an XML DOM.
  	*
  	* The output is a JavaScript object of GeoJSON data, same as `.kml` outputs, with the
  	* addition of a `_gpxType` property on each `LineString` feature that indicates whether
  	* the feature was encoded as a route (`rte`) or track (`trk`) in the GPX document.
  	*/

👋 Raruto