This is on production, I tested with the summa_la_a corpus (Summa theologiae)
While entering a CQL type query, the <s>, <i> and <p> structures are not recognized as existing structures; on the other hand, <h> is recognized (so it's not one-letter structures that are problematic). For <s> and <p>, it could be related to the fact that there are structures with longer names that start with these letters (<seg> and <pars>); for <i>, however, there is no such structure, only an attribute named id.
Possibly related: https://github.com/ufal/lindat-kontext/issues/195
This is on production, I tested with the
summa_la_a
corpus (Summa theologiae) While entering a CQL type query, the<s>
,<i>
and<p>
structures are not recognized as existing structures; on the other hand,<h>
is recognized (so it's not one-letter structures that are problematic). For<s>
and<p>
, it could be related to the fact that there are structures with longer names that start with these letters (<seg>
and<pars>
); for<i>
, however, there is no such structure, only an attribute namedid
. Possibly related: https://github.com/ufal/lindat-kontext/issues/195