Documentation compiled with version 7.0.0-pr6

Matching and filtering morphological properties

We’ve loaded a LatinCorpus of the Fabulae of Hyginus. (See an example of how to do that.)

Filtering tokens for a morphological value

For morphological filtering and querying, we need the tabulae library.

import edu.holycross.shot.tabulae._

We can filter tokens for a classified value using the morphologyMatches method together with member values of the MorphologicalValue object.

val imperfectTokens = hyginus.tokens.filter(_.morphologyMatches(MorphologicalValue.imperfect))

Lots of imperfects in Hyginus!

imperfectTokens.size
// res0: Int = 690

Filtering for a list of morphological values

We can also work with Vectors of classified values. We can separate out imperfect subjunctives from imperfect indicatives, for example, using a Vector listing both a mood value and a tense value.

val imperfectIndicative = Vector(MorphologicalValue.imperfect, MorphologicalValue.indicative)
val imperfectSubjunctive = Vector(MorphologicalValue.imperfect, MorphologicalValue.subjunctive)

The andMorphMatches method applies logical and to require that the token matches all the listed properties.

val imperfectIndicativeTokens =  hyginus.tokens.filter(t => t.andMorphMatches(imperfectIndicative))
val imperfectSubjunctiveTokens = hyginus.tokens.filter(t => t.andMorphMatches(imperfectSubjunctive))

So our result is that imperfect subjunctives are roughly four times as frequent as imperfect indicatives. (Teach the imperfect subjunctive early and often.)

imperfectIndicativeTokens.size
// res1: Int = 136
imperfectSubjunctiveTokens.size
// res2: Int = 554

Chaining filters with anding and oring

As with any filtering operation in Scala, we can chain results together to apply further filters. To find all forms that are either present indicative passive, or impferfect indicative passive, we could apply a parallel orMorphMathces method first, to select all tokens that are either in the perfect or imperfect tense:

or-ing the requirements:

val perfectOrImperfect = hyginus.tokens.filter(t => t.orMorphMatches(Vector(MorphologicalValue.perfect, MorphologicalValue.imperfect)))

(And they are legion: more than 3600.)

perfectOrImperfect.size
// res3: Int = 3657

And then applying to that set of tokens a second filter anding a requirement that the tokens be indicative mood and passive voice.

val indicativePassive = Vector(MorphologicalValue.indicative, MorphologicalValue.passive)
val perfectOrimperfectIndicativePassive =  perfectOrImperfect.filter(t => t.andMorphMatches(indicativePassive))

Considerable fewer of those!

perfectOrimperfectIndicativePassive.size
// res4: Int = 65