RDFa ordering not preserved on duplicated properties about extruct

[adityas114]

Hi, I would like to work on this. It would be great if you can assign me the issue.

I am thinking when the parsed output is returned from rdflib in rdfa.py, we can use the document object to fix the order in extruct itself. I have started working on it and I can raise a pull request soon.

[ivanprado]

@croqaz the order is important for values. Usually, some pages annotate several images but the most important one (the one with the biggest resolution or more relevant) is the first one. If a set is used the order is lost. Instead, it would better to get consistent ordering by having values in the order they appear in the page. That's unfortunately not the case for RDFa.

[croqaz]

Thank you for the explanation @ivanprado 👍

[kmike]

fwiw, extruct is using https://github.com/RDFLib/rdflib, not pyRdfa. This could be related (not 100% sure): RDFLib/rdflib#538.

[croqaz]

What would be the implications to use SET instead or LIST in the RDFa values?
Eg: The values would be converted to sets of frozen Dicts:

"http://ogp.me/ns#image": [
                {
                    "@value": "http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg"
                },
                {
                    "@value": "http://images.sk-static.com/SECONDARY_IMAGE.jpg"
                }
            ],

[ivanprado]

Hi @adityas114 . Nice if you can work on it 👍 . In order to contribute, just create a PR with the fix. There is no need to assign the issue.

[adityas114]

@ivanprado I have created a PR (#139) for this issue. Please have a look and let me know if it requires any changes.