Laboratoire ICAR » Plateforme TXM

To reproduce the bug, use the PROFITEROLE-GOLD-V1-0 corpus (https://sharedocs.huma-num.fr/wl/?id=CzLWbeBSqR0lk4Yx0XqHSO7NebydM4ab) and use the following query:

#pivot:[pos="VERB"] & #clause:[cat!=("root"|"parataxis") & type="VFin"] & #clause >L #pivot & #clause >D #obj:[cat=("obj"|"ccomp"|"obj\:advneg"|"obj\:advmod")] & #clause >D #suj:[cat=("nsubj"|"csubj")] & #obj >L #objhead:[pos!="PRON"] & #suj >L #sujhead:[] & #sujhead .* #pivot   & #pivot   .* #objhead

Left context (inside a sentence) is not shown for one or several first occcurrences of certain texts. The pivot is not the first word of the text. The sententce is not necessaruly the first in the text. The console displays error messages as follows:

Error: trying to get word values from 129095 to 96307
Error: trying to get word values from 203229 to 42810
Error: trying to get word values from 155804 to 42901
Error: trying to get word values from 129095 to 42926

It looks like the first ID of the range is superior to the last...