Update identification for ISO 19115

[roomthily]

Based on priorities and harvest results, but basically how to catch Service Metadata as the main document (vs. a standard MI/MD with a Service block) and, possibly, Data Series handling.

We will be tagging anything generated with the newer *-1 incorrectly with our basic namespace check.

[roomthily]

Changed it a bit to remove the regex (ugh) to avoid some unfortunate false positives.

[roomthily]

Better with this: 6f138ab but see the other two open issues.

[roomthily]

I am going to treat 19119 (Service Metadata) records as just some version of 19115?. The SV_ServiceIdentification element can be a sibling of an MD_DataIdentification element (and within the main Mx_Metadata record) and there's no unique combination of namespacing and elements to identify an MD with a Service vs an MD as a container for a service (container in that it only includes the Service block).

I suppose we could go with "has SV_ServiceIdentification where no sibling == Mx_Metadata" but I have very few real world examples to know if that is a reasonable assumption.

Update: another example where we have the two metadata structures (and they are in different identificationInfo blocks):
http://coastwatch.pfeg.noaa.gov/erddap/metadata/iso19115/xml/erdCalCOFIlrvcntpos_iso19115.xml

[roomthily]

See c9e49ab and f5947a5