Identify Feature in XML Mode forcing Accept header to application/xml when Resolving external

[sgrellet]

Environment

• Plugin version: 1.3.1
• QGIS version: 3.16.2-Hannover Hannover, f1660f9da5
• Operating System: Windows 10
• Python version: 3.7.0 (v3.7.0:1bf9cc5093, Jun 27 2018, 04:59:51) [MSC v.1914 64 bit (AMD64)]

Description

• in a GML file, when appending content (Resolve external), the Accept header is forced to application/xml.
• this may cause issue in content negociation as most of the time we resolve (note: for now the plugin is in "XML" paradigm)
  • to application/gml+xml for features
  • to application/rdf+xml for vocabularies

To Reproduce

Steps to reproduce the behavior:
Attached file: BoreholeView_Instance.zip

1. Load the attached file in the Load Wizard > File/URL
2. Loading options > 'Load in XML mode'
3. Next and Finish
4. epos:status > Resolve external > Embedded

-> Python error
"XML parsing error" : The external resource is not a well formed XML

Expected behavior

• the epos:status attribute content should be enriched (thanks to its URI) by the content available at https://data.geoscience.earth/ncl/BoreholeStatus/drillingCompleted

---

Additional context

• QGIS 'Network Logger' (F12) shows an Accept header set to application/xml

• querying that URI in a application such as PostMan with Accept header set to application/xml leads to fetching the wrong serialization from a linked data registry. it should be application/rdf+xml (in an XML paradigm)

• but we cannot fix it to application/rdf+xml as it will have a side impact when we ask for application/gml+xml

• idea

  • either ask the client in the GUI
  • or try various Accept values (like what normal web client do) : application/xml, application/rdf+xml

[Guts]

It seems that the server use an URL query parameter format to determine which output is expected. For example for rdf/xml: https://data.geoscience.earth/ncl/BoreholeStatus/drillingCompleted?_format=rdf

@sgrellet is there a way to get this kind of information (accepted content to use) from original XML?

cURL samples:

curl -H "Accept: application/rdf+xml" -X GET https://data.geoscience.earth/ncl/BoreholeStatus/drillingCompleted

works, but:

curl -H "Accept: application/xml,application/rdf+xml" -X GET https://data.geoscience.earth/ncl/BoreholeStatus/drillingCompleted

doesn't work. It returns HTML: text/html. It should returns one of the requested ones if it supports it.

BTW, the server doesn't answer with OPTIONS.

[Guts]

The server doesn't answer with OPTIONS.

[sgrellet]

@Guts : sorry for the latency.

no need to pass a 'format' parameter, just real conneg

in a 'XML' view of the world, "application/rdf+xml" should be your choice.
It makes sense the server does not know what yo do with application/xml as there is no obvious XML serialization for the linked data content available (note : I wrote XML" not "RDF/XML").

What do you mean by ' original XML' in your message above ?