RDF generation script for IANA Media Types

MediaType-rdf.py

import csv
from rdflib import Graph, Literal, URIRef, Namespace, RDF, RDFS

base_uri = "http://www.iana.org/assignments/media-types/"

DCT = Namespace("http://purl.org/dc/terms/")
DCTYPE = Namespace("http://purl.org/dc/dcmitype/")
DCAM = Namespace("http://purl.org/dc/dcam/")

g = Graph()
g.bind("dct", DCT)
g.bind("dcam", DCAM)

# get CSV of MIME type registry assignments from 
# http://www.iana.org/assignments/media-types/media-types.xml
mimetype_registries = ["application", "text", "video", "image"]

for mimetype_registry in mimetype_registries:
    with open(mimetype_registry+'.csv', newline='') as csvfile:
        reader = csv.reader(csvfile)

        # pop header row
        reader.__next__()

        for row in reader:

            if not row[1]:
                uri = base_uri + mimetype_registry + "/" + row[0]
            else:
                uri = base_uri + row[1]

            mimetype = g.resource(URIRef(uri))
            mimetype.add(RDF.type, DCT.FileFormat)
            mimetype.add(RDF.type, DCT.MediaTypeOrExtent)
            mimetype.add(RDF.value, Literal(row[0]))
            mimetype.add(DCAM.memberOf, DCT.IMT)

            if mimetype_registry == "image" and row[2]:
                mimetype.add(RDFS.label, Literal(row[2]))

g.serialize("mime.ttl", format="turtle")