korap / korap-xml-tei Goto Github PK

This is just to note that the <metadata> element doesn't seem to do anything (and I guess this statement can be strengthened to say that the element is misguiding). It is required by text.rng though.
If you'd like me to, I can remove the requirement in the schema, so that old files are still going to be valid, but new files won't need it -- with millions of files, that might maybe help a bit?

The error occurs e.g. if there is not enough memory to start the Java VM. The result can be something like:

<?xml version="1.0" encoding="UTF-8"?>
<?xml-model href="span.rng"
            type="application/xml"
            schematypens="http://relaxng.org/ns/structure/1.0"?>
<layer docid="AAZ22_JAN.00001"
       xmlns="http://ids-mannheim.de/ns/KorAP"
       version="KorAP-0.4">
  <spanList>
    <span id="t_0" from="Error" to="occurred" />
    <span id="t_1" from="during" to="initialization" />
    <span id="t_2" from="of" to="VM" />
  </spanList>
</layer>
``

Error message:

Maybe empty textSigle => skipping this text ...
data=

I'm looking at a fragment of the Gingko corpus header from https://github.com/KorAP/KorAP-XML-Krill/blob/master/t/real/corpus/Gingko/ATZ07/JAN/00001/header.xml :

<?xml version="1.0" encoding="UTF-8"?>
<?xml-model href="header.rng"
            type="application/xml"
            schematypens="http://relaxng.org/ns/structure/1.0"?>
<!DOCTYPE idsCorpus PUBLIC "-//IDS//DTD IDS-XCES 1.0//EN"
          "http://corpora.ids-mannheim.de/idsxces1/DTD/ids.xcesdoc.dtd">
<idsHeader TEIform="teiHeader" pattern="text" status="new" type="text" version="1.0">
    <fileDesc>
        <titleStmt>
            <textSigle>ATZ07/JAN.00001</textSigle>
            <t.title assemblage="external">ATZ07/JAN.00001 ATZ - Automobiltechnische Zeitschrift, Januar 2007, Nr. 109(1), S. 10-15; Ein neues Energiemanagement-Konzept für das elektrische Bordnetz</t.title>
        </titleStmt>

And it seems to me that both schema statements are ornaments, because

• header.rng requires that the root element should be <header>
• while the DTD requires that the root element should be <idsCorpus>

Fixing the root element in the DTD doesn't move us forward, because then lots of errors crop up that suggest that the DTD hasn't been updated while the XML has evolved.

Removing both schema associations would make the result truer, in a way ;-)
<!DOCTYPE idsHeader> also works, but, naturally, it doesn't do much.

When using --inline-tokens, the inline token annotations are stored in the target foundry#layer, but also in the structures foundry#layer. This is not very helpful, as the annotations are stored as attributes in the structure foundry and hard to reach. They should be ignored by default.