In cboettig/emld: Ecological Metadata as Linked Data

This document outlines the conventions used in mapping EML XML into JSON-LD.

XML node names become JSON keys
XML node contents (either text/data or node names) become the corresponding JSON value to the key
Repeated elements become list-valued:

<url>http://example.com/1</url>
<url>http://example.com/2</url>
<url>http://example.com/3</url>

becomes:

"url": ["http://example.com/1", "http://example.com/2", "http://example.com/3"]

XML node attribute names are converted into JSON keys prefixed with #, unless the attribute name is id, in which case it uses the special JSON-LD prefix, @. For instance:

<dataset id="abc123", system="knb"><title>The Title<title></dataset>

becomes

"dataset": {
  "@id": "abc123", 
  "#system": "knb",
  "title": "The Title"
}

If a node contains text/data and an attribute, the node name is repeated to denote the value as well:

<url function="download">http://example.com</url>

becomes:

"url": {
  "#function"="download"
  "url": "http://example.com"
}

Contents of para and section (the two choices for components of any TextType content) are encoded as literal character strings and not parsed.
references are expanded:
Semantic annotations (EML 2.2) are expanded into native JSON-LD