lemmas used for dictionary entries

[gregorycrane]

I know we have discussed this but I just wanted to reiterate that we need to use an available treebank first when mapping forms to lemmas. It looks like Iliad 1 does not use the treebank (it has the wrong lemma for

https://beyond-translation.perseus.org/reader/urn:cts:greekLit:tlg0012.tlg001.perseus-grc2:1.1-1.20?mode=dictionary-entries&entryUrn=urn%3Acite2%3AexploreHomer%3Aentries.atlas_v1%3A2.6316

It shoiuld be Chryses the person.

[jacobwegner]

Hi @gregorycrane. I'll be doing some work on Beyond Translation this weekend and can take a closer look. I suspect that one layer of this is that there are separate "buckets" for token-level data...the treebank data will be used if you're in Syntax Trees mode (and when we have multiple treebanks available, it will use the selected treebank):

https://beyond-translation.perseus.org/reader/urn:cts:greekLit:tlg0012.tlg001.perseus-grc2:1.11?mode=syntax-trees&entryUrn=urn%3Acite2%3AexploreHomer%3Aentries.atlas_v1%3A2.6327

Looking at the other "bucket" (the lemmatization data used for Iliad in the interlinear display), I see a separate issue:

https://beyond-translation.perseus.org/reader/urn:cts:greekLit:tlg0012.tlg001.perseus-grc2:1.11?mode=interlinear&entryUrn=urn%3Acite2%3AexploreHomer%3Aentries.atlas_v1%3A2.6316

The lemma is shown as Χρύσης, but when we click the word, we're using Morpheus to do a lookup, which returns two results:

1. Χρύση
2. Χρύση

I'll see if I can implement a more intelligent "fallback" from "curated" lemmas back to Morpheus, and where we're currently just using the first result from Morpheus, we should probably support clicking a result in the Morpheus widget to set the lemma.