Lexical Inventory

Task List Summer 2014

Data Sets

Stems

stem [urn identifier]
stem  [string representation]
hasCategory [ controlled vocabulary ]
hasLanguage [iso language code]
hasStatus [controlled vocabulary (e.g. proposed, accepted, etc)
submittedBy userURI
acceptedBy userURI
createdAt datetime
approvedAt datetime

Form analyses

form analysis [urn identifier]
hasAnalysis [json per lexicon.xsd - complete analysis]
attestedAt* [cts urn]
hasLanguage [iso language]
hasStem* [stem urn]
isInstanceOf [lexicalentity urn]
hasStatus [controlled vocab]
submittedBy userURI
acceptedBy userURI
createdAt [datetime]
approvedAt [datetime]

Vote Annotations

annotationIdentifier [urn]
formIdentifier hasStem stemIdentifier
voteType [ implicit (by selecting a form) vs explicit (voting yes or no) ]
hasVote [yes|no]
submittedBy userURI

Work Assignments

Robert & Gernot

• defining controlled vocabuary
• controlled vocabulary service
• llt form builder service
• json representation of lexicon.xsd
• review ui in arethusa (reached through sosol)

Bridget (or Adam)

• update greek lexical entity collection to match latin format (with triples for links to morpheus lemmas and dictionaries, etc.)
• urn identifier sequence creator
• update Morphology service json output
• sosol api to crud operations on triple store (bypassing sosol identifier/git)
• board review of external identifier (bypassing sosol identifier/git)
• deploy issues
  • decide on triple store (fuseki, MarkLogic?, or..)
  • deploy on services.perseids.org
  • backup

Old Notes

Steps to populate a base inventory:

1. Extract all distinct lemmas (and any available short defs) from

   1. Lewis & Short Lexicon (via Alpheios index)
   2. Perseus hib_lemmas table
   3. Philologic DB
   4. Latin Language Toolkit DB

2. Normalize lemmas to:

   1. strip vowel length (Question: do we want only to strip vowel length or all accents?)
   2. retain case
   3. strip hyphens ? (Question: I don't actually know if the hyphens need to be stripped or if there is something more subtle going on with them. I think Helma was going to be looking at these)

3. Submit each lemma to morpheus as a check to see if it parsed ? We know Morpheus is not a definitive authority here so I'm not sure whether we want to do this at this point.