Add support for specifying the column used to join to the document to retrieve geometries

[jkerr5]

Right now, the column name used to join the document into the optic pipeline is hardcoded to DocId and this will come from the first view in the data sources array.

We need to make this configurable by requiring the column name to be specified in the geometry.source object in the layer descriptor. The new property should be called joinDocColumn and must be present if the geometry will be extracted via XPath (the xpath property is also present).

The code should then use that property when joining in the document such that the geometry can be extracted: https://github.com/marklogic-community/marklogic-geo-data-services/blob/master/src/main/ml-modules/services/geoQueryService.sjs#L1017

If the joinDocColumn is not present but the document is needed to extract the geometry, the fragment ID column from the "primary" data source (the first in the array) will be used.

The column that the document content is stored in is called doc. Perhaps this should be renamed to geometryDoc to make it more explicit and also reduce the changes of a collision with the name.

[jkerr5]

We need to relook at this to see if this is still an issue.