⭐ parse.xml

[arlimus]

Add support for parsing XML files, either by providing the file or by providing the XML contents directly.

Using inline data:

> parse.xml(content: '<root />').params
parse.xml.params: {
  root: {}
}

Via file example.xml:

<root>
  <box>
    <hello a="1"/>
  </box>
  <box>
    <world b="2">
      <c>3</c>
      4
    </world>
  </box>
  <box>🌎</box>
</root>

Running this in MQL:

> parse.json('example.xml').params
parse.xml.params: {
  root: {
    box: [
      0: {
        hello: {
          @a: "1"
        }
      }
...

Since the conversion from XML to a flat (JSON-like) structure isn't standardized, we took an approach very similar to other formatters (e.g. xq, jsonformatter). This means:

• single items are added as a flat child element, e.g. the first <box/> element has a flat field hello above
• once there are multiple elements, they are turned into a list, e.g. root.box is an array above
• attributes are added as @attribute fields in the element
• if an element only has text contents, the value is just set to the text, e.g. <root>val</root> is turned into {"root": "val"}
• if an element has a mix of child elements and text contents, all text contents are added to the __text child field

Future consideration are to add access to the entire XML structure, which is more complex to traverse, but may be required for some use-cases. We will also handle streaming data separately (for all of these formats)

[github-actions]

Test Results

3 705 tests  +11   3 701 ✅ +11   1m 51s ⏱️ +4s
  400 suites ± 0       4 💤 ± 0 
   30 files   ± 0       0 ❌ ± 0 

Results for commit d3f4a33. ± Comparison against base commit da56dae.

♻️ This comment has been updated with latest results.

[afiune]

One nit and one comment that needs to get in before merging (map re-initialization), besides that, this looks good to me!

[afiune]

This always re-initializes the map.

Suggested change

	x.attributes = map[string]string{}
	if x.attributes == nil {
	x.attributes = map[string]string{}
	}

[afiune]

nit I wonder if we should separate the text content into a slice to simplify this code, maybe adding the slice to the xmlElem and append it in UnmarshalXML() for the xml.CharData case.