search

scaife-viewer / beyond-translation-site

Site used to iterate on translation alignments within the Scaife Viewer ecosystem
3 stars 4 forks source link

data question: why does klisih not appear as a vocab word in Il. 1.185? #57

Closed gregorycrane closed 2 years ago

gregorycrane commented 2 years ago

https://beyond-translation-gagdt-dev.herokuapp.com/reader/urn:cts:greekLit:tlg0012.tlg001.perseus-grc2:1.182-1.192?mode=alignments&rs=urn%3Acite2%3Ascaife-viewer%3Aalignment.v1%3Ailiad-word-alignment-parrish-998078bc3bab42978b47fa8e8b852cae

Are we still using Giuseppe's lemmatization here? NB: some editions combine klisihn + de and present this as one word, rather than two.

image

image

image

jacobwegner commented 2 years ago

@gregorycrane When the user makes a selection, we do a lookup for the lemma against Morpheus (in the Morphology widget):

image

Morpheus is returning κλισι-ην, so that is what we end up using to query the dictionaries.

I've got a few other query-by-lemma tweaks I'm working on this week; let's touch base on Tuesday to see if any of them help this edge case.

jacobwegner commented 2 years ago

Next steps:

1) Prefer the n attributes

      <div xml:id="clisie-cunliffe-lex" type="textpart" n="κλισία">
        <head>
          <foreign xml:lang="greek">κλισίη</foreign>
        </head>

2) Fallback to Morpheus for a lemma

jacobwegner commented 2 years ago

I've switched Cunliffe over to using the n attributes for headwords; I've also created #61 to help deal with the "fallback" issue (where Morpheus differs from our "curated" lemmas)

jacobwegner commented 2 years ago

View on the preview instance:

image