Closed jesper-friis closed 2 years ago
It turns out that the XML specification require that the name part of an IRI must start with an alphabetic character. Serialising EMMO to rdfxml using Owlready2 now produces invalid XML. Protege manage to produce valid XML by converting an IRI like
<emmo:967080e5_2f42_4eb2_a3a9_c58143e835f9>
to
<emmo2:e5_2f42_4eb2_a3a9_c58143e835f9>
where the namespace "emmo2" is "http://emmo.info/emmo#967080". But this completely messes up everything.
This PR readds the initial EMMO_ to all IRIs using
EMMO_
find -name '*.ttl' | xargs sed -i 's/\([#:]\)\([[:xdigit:]]\{8\}_[[:xdigit:]]\{4\}_[[:xdigit:]]\{4\}_[[:xdigit:]]\{4\}_[[:xdigit:]]\{12\}\)/\1EMMO_\2/g'
It turns out that the XML specification require that the name part of an IRI must start with an alphabetic character. Serialising EMMO to rdfxml using Owlready2 now produces invalid XML. Protege manage to produce valid XML by converting an IRI like
to
where the namespace "emmo2" is "http://emmo.info/emmo#967080". But this completely messes up everything.
This PR readds the initial
EMMO_
to all IRIs using