Problem with prefixes created for URIs containing %20

[Aleksander-Drozd]

The result of runnig this code

from rdflib import Namespace, Graph, BNode, Literal

graph = Graph()
namespace = Namespace('http://example.org/')
graph.bind('', namespace)
node = BNode()

graph.add((node, namespace['first%20name'], Literal('John')))
print(graph.serialize(format='turtle').decode())

is

@prefix : <http://example.org/> .
@prefix ns1: <http://example.org/first%20> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix xml: <http://www.w3.org/XML/1998/namespace> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

[] ns1:name "John" .

I expect that triple would be displayed as

[] :first%20name "John" .

[joernhees]

this is not a bug but a feature... (i agree that it's confusing, but it would break a lot of things in XML otherwise).

The reason for the behavior you see is that QNames and CURIEs restict the allowed amount of characters in the so called LocalPart a lot more than one would think. The reason is that they are intended to be valid XML attribute names. In other words: <foo :first%20name="something" /> wouldn't work, cause the % is not allowed in that place.

Here's the spec: https://www.w3.org/TR/REC-xml-names/#ns-qualnames
Clicking through you'll arrive at https://www.w3.org/TR/REC-xml/#NT-Name ...

hex(ord('%')) == '0x25' is simply not in the char ranges allowed by NameChar.

I'll close this for now, feel free to reopen.

[joernhees]

on second thought i'll re-open this as Turtle as of RDF 1.1 actually allows such chars in PrefixedName via PN_LOCAL.

I can only guess that this (as so many other issues) has a historical reason or is due to the interplay of N3, TTL, NT...

As the serialization is not really wrong (just a little more cryptic than it needs to be), i'll give this low prio... if anyone else wants to investigate the origin and how to solve this (in a backwards compatible way), feel free

[Aleksander-Drozd]

I will only add that in test files defined for applications implementing W3C CSV to RDF conversion recommendation, there are examples containing '%', e.g. https://github.com/w3c/csvw/blob/gh-pages/tests/test009.ttl

[aayush17002]

I would like to resolve the issue

[white-gecko]

@aayush17002 you are welcome. The best is, if you create a pull-request and refer to this issue in the pull-request description. We can then discuss if we want to introduce it in 5.0.0 or a later release.

[aayush17002]

I have provided a solution to solve the above problem. I would like you to review the solution and revert back.