WMTSCapabilities.createLayer and Format.WMTS issues

[woodbri]

I am trying to get WMTS layer to be created from the WMTSGetCapabilities document. It seems to be reading the document ok and creating the layer but the layers does not work correctly.

Here is the getCapabilities document:

 http://imaptools.com:8080/geocache/wmts?service=wmts&request=getcapabilities

Here is a simple web page I'm using for testing:

 http://imaptools.com:8080/gc4.html

This has 3 layers defined:

1. WMTS in spherical mercator explicitly defining all the parameters. This layer works fine as far as I can tell. It is for reference and to show that the service is up and running correctly.
2. This attempts to read the getCapabilites document and create a layer identical to the above based on the document. It seems that this layer is in WGS84, degrees, and resolutions, etc if I inspect the map.layer[ 1 ].* properties, instead of the values from the requested matrixSet.
3. This attempts to read the getCapabilites document and create a layer that is WGS84 based on the document, and I figured it had at least some chance of working given the defaults found in 2. And map.zoom==1, shows two broken tiles with TILEROW=-1

It seems that OpenLayers.Format.WMTSCapabilities() is not creating a appropriate layer definition from the getCapabilities document.

In OpenLayers.Format.WMTSCapabilities:148 we have:

            layer = new OpenLayers.Layer.WMTS(
                OpenLayers.Util.applyDefaults(config, {
                    url: capabilities.operationsMetadata.GetTile.dcp.http.get,
                    name: layerDef.title,
                    style: style,
                    matrixIds: matrixSet.matrixIds
                })
            );

This is not setting:

layer.resolutions[]
layer.projection
layer.units
layer.maxExtents

from the capabilities document information so these are defaulting to EPSG:4326 values that are not valid for sphericalMercator layers.

Another possible conclusion is that the OpenLayers.Layer.WMTS() constructor is not populating these items from the matrixIds if all the information is in there.

Yes, I know that I can set these items manually via the "config" argument, but this seems to defeat the whole purpose of defining the layer via the capabilities document in the first place.

And digging into this a little more, it looks like OpenLayers/Format/OWSCommon/v1.js does not extract the BoundingBox tags at all:

            "BoundingBox": function(node, obj) {
                // FIXME: We consider that BoundingBox is the same as WGS84BoundingBox
                // LowerCorner = "min_x min_y"
                // UpperCorner = "max_x max_y"
                // It should normally depend on the projection
                this.readers['ows']['WGS84BoundingBox'].apply(this, [node, obj]);
            },

According to the standard you can have zero or more of these:

<ows:BoundingBox crs="urn:ogc:def:crs:EPSG::4326">
<ows:LowerCorner>-180.000000 -90.000000</ows:LowerCorner>
<ows:UpperCorner>180.000000 90.000000</ows:UpperCorner>
</ows:BoundingBox>

mod-geocache emits one for each supported crs on a layer and each TileMatrixSet has a SupportedCRS tag that points back to the appropriate BoundingBox. So the BoundingBox tags should get extracted as an array either like:

capabilities.contents.layer[n].bounds{crs} = OpenLayers.Bounds

or

capabilities.contents.layer[n].bounds[0..n] = {
  crs: crs,
  bounds: OpenLayers.Bounds
}

Then it would be possible to fix OpenLayers.Format.WMTSCapabilities.createLayer() to extract the appropriate information from the capabilities document and create a working layer.