Hello! I've noticed a possible missing entity type in COARSE-METO in

Annotation Issue (coarse.meto-fine.comp) about hipe-2022-data HOT 1 OPEN

hipe-eval commented on September 10, 2024

Annotation Issue (coarse.meto-fine.comp)

from hipe-2022-data.

Comments (1)

e-maud commented on September 10, 2024

Many thanks for spotting.

After some investigations, here are some (strange) elements, for information and memo.

The mention is in document EXP-1908-01-21-a-i0053 line 2130 (in HIPE-2022-data file and in CLEF-HIPE-2020-internal file).
In INCEpTION, the mention appears correctly annotated:

In the exported annotations in EXP-1908-01-21-a-i0053.xmi (CLEF-HIPE-2020-internal), the annotation is not there. See lines 1075 and after (permalink), where only comp.title|comp.name|comp.functionare exported:

    <custom:ImpressoNamedEntity xmi:id="12992" sofa="1" begin="2104" end="2106" is_NIL="false" literal="false" noisy_ocr="false" unsolvable="false" unsolvable_linking="false" value="comp.title"/>
    <custom:ImpressoNamedEntity xmi:id="13005" sofa="1" begin="2107" end="2123" is_NIL="false" literal="false" noisy_ocr="false" unsolvable="false" unsolvable_linking="false" value="comp.name"/>
    <custom:ImpressoNamedEntity xmi:id="13018" sofa="1" begin="2125" end="2152" is_NIL="false" literal="false" noisy_ocr="false" unsolvable="false" unsolvable_linking="false" value="comp.function"/>
    <custom:ImpressoNamedEntity xmi:id="13031" sofa="1" begin="2146" end="2152" is_NIL="false" literal="false" noisy_ocr="false" unsolvable="false" unsolvable_linking="false" value="loc.adm.reg" wikidata_id="http://www.wikidata.org/entity/Q12745"/>

We will correct this in the current data and log the change, but since it seems to come from the xmi2IOB step, I wonder if other (complex) entities are also lost in the way, and whether it is worth checking the export code.

@mromanello What do you think?

from hipe-2022-data.

Related Issues (6)

Recommend Projects