cwrc / NERVE

Named Entity Recognition Vetting Environment
GNU General Public License v2.0
3 stars 3 forks source link

Result missing lemmaAttribute (tei: key attribute; cwrc/orlando: standard name) #89

Open jefferya opened 4 years ago

jefferya commented 4 years ago

Parts

Error:

Ms. <persName>Atwood</persName> won the Booker prize. 

Desired

Ms. <persName key="Atwood" resp="NERVE">Atwood</persName> won the Booker prize. 

Functionality removed from EncoderNER.java by

https://github.com/cwrc/NERVE/commit/b869f176d7aad199662f1b43e4307152f427f490#diff-6fffdaa24107651091aa9f57333e773c

Investigate fix by adding back code: https://github.com/cwrc/NERVE/blob/2e99fdeb3ea72ebd047b40711fa636754b7b3272/Scriber/src/main/java/ca/sharcnet/nerve/scriber/encoder/servicemodules/EncoderNER.java#L92-L95

jefferya commented 4 years ago

@ilovan @ajmacdonald do you know which side, NERVE or CWRC-Writer adds the resp attribute?

ajmacdonald commented 4 years ago

@jefferya I believe NERVE should add that in its response. The attribute would then get removed once the user accepts the entity in CWRC-Writer.

jefferya commented 4 years ago

@ajmacdonald @ilovan

Can this attribute be added via the context file in the defaults key? For example:

"TITLE": {
            "name"           : "TITLE",
            "lemmaAttribute" : "REG",
            "linkAttribute"  : "REF",
            "idAttribute"    : "annotationID",
            "defaults"       : {"RESP": "NERVE"}
        }
ilovan commented 4 years ago

I don't see why not but probably @ajmacdonald is a better judge of that

ajmacdonald commented 4 years ago

The resp attribute is now added to the defaults. https://github.com/cwrc/CWRC-WriterBase/commit/4a054114a2ffadd543c3426fd920c4323153c75a