from rdflib import Graph from rdflib import RDF, URIRef, Namespace XPUB = Namespace("http://xpub.nl/terms/") import sys # url = "https://gallery.constantvzw.org/index.php/Maison-des-arts-expo/" # g = Graph() # with open("index.rdfa.html") as fin: # data = fin.read() # g.parse(publicID="index.html", data=data, format="rdfa") # # g.parse(file=fin, format="rdfa", override_encoding="utf-8") g = Graph() g.parse("https://project.xpub.nl/index.rdfa.html", format="rdfa") context = { "@context": { "@language": "en", "xpub": "http://xpub.nl/terms/", "xpubproject": "https://project.xpub.nl/", "dc": "http://purl.org/dc/terms/", "student": { "@id": "xpub:student", "@lang": "en" }, "project": { "@id": "xpub:project", "@type": "@id" }, "thesis": { "@id": "xpub:thesis", "@type": "@id" }, "image": {"@id": "dc:image", "@type": "@id" }, "title": "dc:title", "date": {"@id": "dc:date", "@type": "http://www.w3.org/2001/XMLSchema#date"}, "hasPart": {"@id": "dc:hasPart", "@type": "@id"} } } print (g.serialize(format="json-ld", context=context, auto_compact=True))