TitleCased numbers in ProperNouns

[thegoatherder]

In v13, Any #TextValue such as five or twelve is correctly tagged, regardless of the casing of the string.
In v14, the tagger fails to add #TextValue for any word-number with a first uppercase letter - instead the word gets tagged as #ProperNoun

e.g. The Twelve Days of Christmas
Got:
Twelve => #ProperNoun
Want:
Twelve => #TextValue #Value

Possibly related #992 ?

Example

const nlp = require('compromise')
const nlpDates = require('compromise-dates')
const nlpNumbers = require('compromise-numbers')
nlp.plugin(nlpDates)
nlp.plugin(nlpNumbers)

const text = [
  // uppercase #TextValue fails to tag as #TextValue
  'I watched the Twelve Days of Christmas',
  'I put him on the Two-Week Wait cohort',
  'I saw him Five years ago',
  // lowercase #TextValue correctly tagged as #TextValue
  'I watched the twelve days of christmas',
  'I put him on the two-week wait cohort',
  'I saw him five years ago',
]

text.forEach((t) => {
  const doc = nlp(t)
  const result = doc.has('#Value') ? 'MATCH SUCCESS' : 'MATCH FAIL'
  console.log(result)
  console.log(doc.debug())
})

v14 output

MATCH FAIL

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'watched'  - Verb, PastTense
  │ 'the'      - Determiner
  │ 'Twelve'   - ProperNoun, Noun
  │ 'Days'     - Noun
  │ 'of'       - Preposition, Date
  │ 'Christmas'  - Date, Noun, Holiday


View { ptrs: undefined }
MATCH FAIL

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'put'      - Verb, PresentTense, Infinitive
  │ 'him'      - Noun, Pronoun
  │ 'on'       - Preposition
  │ 'the'      - Determiner
  │ 'Two'      - ProperNoun, Noun
  │ 'Week'     - Date, Noun, Duration
  │ 'Wait'     - Noun, Singular, ProperNoun
  │ 'cohort'   - Noun, Singular


View { ptrs: undefined }
MATCH FAIL

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'saw'      - Verb, PastTense
  │ 'him'      - Noun, Pronoun
  │ 'Five'     - ProperNoun, Noun
  │ 'years'    - Date, Noun, Duration
  │ 'ago'      - Date


View { ptrs: undefined }
MATCH SUCCESS

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'watched'  - Verb, PastTense
  │ 'the'      - Determiner
  │ 'twelve'   - Value, TextValue, Cardinal, Date
  │ 'days'     - Date, Noun, Duration
  │ 'of'       - Preposition, Date
  │ 'christmas'  - Date, Noun, Holiday


View { ptrs: undefined }
MATCH SUCCESS

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'put'      - Verb, PresentTense, Infinitive
  │ 'him'      - Noun, Pronoun
  │ 'on'       - Preposition, Date
  │ 'the'      - Determiner, Date
  │ 'two'      - Value, TextValue, Cardinal, Date
  │ 'week'     - Date, Noun, Duration
  │ 'wait'     - Noun, Singular
  │ 'cohort'   - Noun, Singular


View { ptrs: undefined }
MATCH SUCCESS

  ┌─────────
  │ 'I'        - Noun, Pronoun
  │ 'saw'      - Verb, PastTense
  │ 'him'      - Noun, Pronoun
  │ 'five'     - Value, TextValue, Cardinal, Date, DateShift
  │ 'years'    - Date, Noun, Duration, DateShift
  │ 'ago'      - Date, DateShift

v13 output

MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'watched'  - PastTense, Verb
  ｜ 'the'      - Determiner
  ｜ 'Twelve'   - TextValue, Value, Cardinal, ProperNoun, Noun, Date
  ｜ 'Days'     - Duration, Date, Noun, Plural
  ｜ 'of'       - Date
  ｜ 'Christmas'  - Holiday, Date, Noun

Doc$1 {
  list: [ Phrase$3 { start: 'i-gsfbh1W', length: 7, isA: 'Phrase' } ]
}
MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'put'      - Infinitive, PresentTense, Verb
  ｜ 'him'      - Pronoun, Noun, Singular
  ｜ 'on'       - Date
  ｜ 'the'      - Determiner, Date
  ｜ 'Two'      - TextValue, Value, Cardinal, ProperNoun, Noun, Singular, Date
  ｜ 'Week'     - Duration, Date, Noun, Singular
  ｜ 'Wait'     - ProperNoun, Noun, Singular
  ｜ 'cohort'   - Noun, Singular

Doc$1 {
  list: [ Phrase$3 { start: 'i-okIEpXA', length: 9, isA: 'Phrase' } ]
}
MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'saw'      - PastTense, Verb
  ｜ 'him'      - Pronoun, Noun, Singular
  ｜ 'Five'     - TextValue, Value, Cardinal, ProperNoun, Noun, Date, DateShift
  ｜ 'years'    - Duration, Date, Noun, Plural, DateShift
  ｜ 'ago'      - Date, DateShift

Doc$1 {
  list: [ Phrase$3 { start: 'i-fzB8oMG', length: 6, isA: 'Phrase' } ]
}
MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'watched'  - PastTense, Verb
  ｜ 'the'      - Determiner
  ｜ 'twelve'   - TextValue, Value, Cardinal, Date
  ｜ 'days'     - Duration, Date, Noun, Plural
  ｜ 'of'       - Date
  ｜ 'christmas'  - Holiday, Date, Noun

Doc$1 {
  list: [ Phrase$3 { start: 'i-kLi0Yga', length: 7, isA: 'Phrase' } ]
}
MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'put'      - Infinitive, PresentTense, Verb
  ｜ 'him'      - Pronoun, Noun, Singular
  ｜ 'on'       - Date
  ｜ 'the'      - Determiner, Date
  ｜ 'two'      - TextValue, Value, Cardinal, Date
  ｜ 'week'     - Duration, Date, Noun, Singular
  ｜ 'wait'     - Infinitive, PresentTense, Verb
  ｜ 'cohort'   - Noun, Singular

Doc$1 {
  list: [ Phrase$3 { start: 'i-rDNGLHS', length: 9, isA: 'Phrase' } ]
}
MATCH SUCCESS
=====
  -----
  ｜ 'I'        - Pronoun, Noun, Singular
  ｜ 'saw'      - PastTense, Verb
  ｜ 'him'      - Pronoun, Noun, Singular
  ｜ 'five'     - TextValue, Value, Cardinal, Date, DateShift
  ｜ 'years'    - Duration, Date, Noun, Plural, DateShift
  ｜ 'ago'      - Date, DateShift

Doc$1 {
  list: [ Phrase$3 { start: 'i-VCEua12', length: 6, isA: 'Phrase' } ]
}

[spencermountain]

tricky one - 'the Twelve Days of Christmas' is a ProperNoun, just like 'Visa One Express' or whatever.

for your purpose, if you want to avoid this, you may want to just add doc.match('(one|two|..twelve)').tag('Value').

not sure what to do on this

[thegoatherder]

I’m not sure I understand why they have to be mutually exclusive? Can’t it be a #Value, #TextValue and #ProperNoun all at the same time? That would appear semantically correct to me.

[thegoatherder]

@spencermountain the strings were correctly labelled both #TextValue, #Value and #ProperNoun in v13. Can’t we just follow the same logic in v14? Why can’t a #ProperNoun also be a #Value?