Missing element values with unwrap()

[laurianvm]

unwrap() (or copy()) of this example_xml:

<Element>
    <ChildA>value_a</ChildA>
     <ChildB namespace="I-am-an-attribute">value_b</ChildB>
</Element>

gives:

{
   "Element":{
      "ChildA":"value_a",
      "ChildB":{
         "@namespace":"I-am-an-attribute"
      }
   }
}

while you would expect something like this:

{
   "Element":{
      "ChildA":"value_a",
      "ChildB":{
         "@namespace":"I-am-an-attribute",
         "#value":"value_b"
      }
   }
}

or, when implementing 'flattening' as mentioned in vliz-be-opsci/pysubyt#4, something like this:

{
   "Element":{
      "ChildA":"value_a",
      "ChildB": "value_b",
      "ChildB.@namespace":"I-am-an-attribute"
      }
   }

[marc-portier]

retake on that last idea - I don´t think we will have the luxury to start this "flattening" half way down

meaning the final thing will be more like:

{ 
    "Element": " <ChildA>value_a</ChildA><ChildB namespace="I-am-an-attribute">value_b</ChildB>", 
    "Element.ChildA":"value_a",
    "Element.ChildB": "value_b",
    "Element.ChildB.@namespace":"I-am-an-attribute"
}

this could indeed be produced through some xdict.flatten()

additionally we would need to have some notation for nested lists: choosing now for **path.element[index] over **path.element.index.

having the xmlasdict- wrapper approac this means this .flatten() becomes available at any level you like. this means one could first navigate down xdb = xdict.ChildB and only then call flatten() which should just return

{ 
    "@namespace": "I-am-an-attribute"
} 

we could consider an extra 'flatten(self_prefix='_') to produce

{ 
    "_": value_b
    "_.@namespace": "I-am-an-attribute"
} 

which in tern might be the smart thing to do in order to resolve this in a recursive way!

[marc-portier]

Just came across https://github.com/bramstein/xsltjson which might give some general advise on how they solved things.
Specially their reference to the Badgerfish convention seems to be relevant here.