Re: MT Confidence

Thanks for the fast feedback Yves. I thought annotatorsRef was element only
thanks for clearing that up. I will update the output this evening.

Thanks,
Leroy


On 11 December 2012 12:57, Yves Savourel <ysavourel@enlaso.com> wrote:

> Hi Leroy, all,
>
> There were a few issue with MT confidence.
>
> I've committed the changes in the input files and attached the expected
> output files.
>
> Here are the issues I've found:
>
> 1-- Several file were using mtConfidenceScore and its-mt-confidence-score
> instead of mtConfidence and its-mt-confidence-score as in the specification.
>
> 2-- Several files had output with attribute nodes that were missing the
> annotatorsRef values. (That value applies to elements and attributes). The
> output values were also wrong in several cases. The result of annotatorsRef
> must be the aggregation of all the references for all data categories. So,
> for example, in the span[1] node of mtconfidence4html.html it should be
> annotatorsRef="disambiguation|file:///tools.xml#T2
> mt-confidence|file:///tools.xml#T1" (the annotator for disambiguation is
> inherited from p).
>
> 3-- mtconfidence2html.html had a confidence set on the p element, but no
> annotatorsRef.
>
> 4-- mtconfidence3html.html had extra its-annotators-ref in the output.
>
> 5-- mtconfidence4html.html had a title text that didn't reflect what the
> test was doing. I've changed it to avoid confusion.
>
> 6-- mtconfidence1xml.xml, mtconfidence2xml.xml, mtconfidence5xml.xml and
> mtconfidence6xml.xml had their outputs with the attributes id and
> its:annotatorsRef in the wrong order.
>
> 7-- mtconfidence3xml.xml had the attribute its:version missing in its
> output.
>
> cheers,
> -yves
>
>
>
>