Extracting and tweaking data from Marble-Lexicon

[klosoter]

Extracting and tweaking data from Marble-Lexicon

[klosoter]

What we might need to do is the following:

• decide what to do with multiple entries (which might have multiple 'meanings' themselves)
• whether or not to preprocess the Marble data (like transforming {L:Dead Sea<SDBH:יָם־הַמֶּלַח>}; into Dead Sea (יָם־הַמֶּלַח))
• find out what else we want to extract and use from the marble lexicon. There are a lot of interesting things:
  • synonyms
  • contextual definitions/references
  • core/sub domains
  • base form/part of speech
  • references to other occurences in Scripture

[jonathanrobie]

In the instance data, I think we want:

• semantic domain
• gloss

Perhaps like this:

sdbh="semdom gloss", using the English gloss. If there are multiple entries, concatenate them with a delimiter.

[jonathanrobie]

A lot of the rest of this is at the lexicon level. We need to figure out what we are doing with Greek and Hebrew semantic lexicons. I need to talk to Reinier about what we have rights to.

[klosoter]

There were no issues when adding the semantic domain and glosses from Marble, so I'll close the issue.

The Marble data is structured as follows:
:::<gloss;gloss;gloss>

Multiple glosses and domains within one entry are separated by ";"
Multiple entries are separated by "|"

xml
<m USFMId="GEN 6:16!1" n="010060160011" lang="H" after="׀" lemma="6672 b" morph="Ncfsa" pos="noun" type="common" gender="feminine" number="singular" state="absolute" cherith-english="roof" cherith-chinese="窗户" marble-sense="צֹהַר:001005001:Constructions:covering;roof|צֹהַר:001005001:Constructions:skylight">צֹ֣הַר</m>