Merge pull request #202 from alpheios-project/i75

I75

packages/client-adapters/src/adapters/tufts/engine/morpheusgrc.js

@@ -3,6 +3,11 @@ import * as Models from 'alpheios-data-models'
 
 let data = new ImportData(Models.GreekLanguageModel, 'morpheusgrc') // eslint-disable-line prefer-const
 
+// Morpheus uses 'irregular' as pofs for some pronouns, override with lemma
+// the dictionary entry's conjugation if it's available
+data.inflectionOverrides = {
+ [Models.Feature.types.part]: (i,ls) => i[Models.Feature.types.part].value === Models.Constants.TYPE_IRREGULAR && ls.filter( l => l.features[Models.Feature.types.part].value === Models.Constants.POFS_PRONOUN )
+}
 /*
 Below are value conversion maps for each grammatical feature to be parsed.
 Format:
@@ -17,4 +22,19 @@ data.addFeature(Models.Feature.types.gender).importer
 data.addFeature(Models.Feature.types.declension).importer
  .map('1st & 2nd', [[Models.Constants.ORD_1ST, 1], [Models.Constants.ORD_2ND, 2]])
 
+data.setPropertyParser(function (propertyName, propertyValue, inputElem) {
+ let propertyValues = []
+ if (propertyName === 'decl') {
+ propertyValues = propertyValue.split('&').map((p) => p.trim())
+ } else if (propertyName === 'comp' && propertyValue === 'positive') {
+ propertyValues = []
+ } else if (propertyName === 'pofs' && propertyValue === 'irregular' &&
+ inputElem.hdwd && inputElem.hdwd.$ === 'τίς' ) {
+ propertyValues = [ Models.Constants.POFS_PRONOUN ]
+ } else {
+ propertyValues = [propertyValue]
+ }
+ return propertyValues
+})
+
 export default data

packages/client-adapters/src/adapters/tufts/engine/sedra.js

@@ -11,7 +11,7 @@ data.setMeaningParser(function (meaning, targetWord) {
  return new Models.Definition(meaningText, lang, 'text/plain', targetWord)
 })
 
-data.setPropertyParser(function (propertyName, propertyValue) {
+data.setPropertyParser(function (propertyName, propertyValue, inputElem) {
  let propertyValues = []
  if (propertyName === 'paradigm') {
  // state has some extra "" around values

packages/client-adapters/src/adapters/tufts/engine/whitakers.js

@@ -5,7 +5,7 @@ const data = new ImportData(Models.LatinLanguageModel, 'whitakerLat')
 
 // Whitaker's has weird inflection data for conjugation, we prefer
 // the dictionary entry's conjugation if it's available
-data.inflectionOverrides = [Models.Feature.types.conjugation]
+data.inflectionOverrides = { [Models.Feature.types.conjugation]: (i,ls) => true }
 
 /*
 Below are value conversion maps for each grammatical feature to be parsed.
@@ -25,7 +25,7 @@ data.addFeature(Models.Feature.types.gender).importer
 data.addFeature(Models.Feature.types.tense).importer
  .map('future_perfect', Models.Constants.TENSE_FUTURE_PERFECT)
 
-data.setPropertyParser(function (propertyName, propertyValue) {
+data.setPropertyParser(function (propertyName, propertyValue, inputElem) {
  let propertyValues = []
  if (propertyName === 'decl') {
  propertyValues = propertyValue.split('&').map((p) => p.trim())

packages/client-adapters/src/adapters/tufts/lib.js

@@ -47,7 +47,7 @@ class ImportData {
 
  // may be overridden by specific engine use via setPropertyParser - default just returns the property value
  // as a list
- this.parseProperty = function (propertyName, propertyValue) {
+ this.parseProperty = function (propertyName, propertyValue, inputElem) {
  let propertyValues = []
  if (propertyName === 'decl') {
  propertyValues = propertyValue.split('&').map((p) => p.trim())
@@ -195,10 +195,10 @@ class ImportData {
  if (Array.isArray(inputItem)) {
  // There are multiple values of this feature
  for (const e of inputItem) {
- values.push(...this.parseProperty(inputName, e.$))
+ values.push(...this.parseProperty(inputName, e.$, inputElem))
  }
  } else {
- values = this.parseProperty(inputName, inputItem.$)
+ values = this.parseProperty(inputName, inputItem.$, inputElem)
  }
  // `values` is always an array as an array is a return value of `parseProperty`
  if (values.length > 0) {
@@ -233,11 +233,11 @@ class ImportData {
  console.warn('Mutiple feature values with mismatching attribute value', inputElem)
  }
  featureName = e[attributeName]
- values.push(...this.parseProperty(inputName, e.$))
+ values.push(...this.parseProperty(inputName, e.$, inputElem))
  }
  } else {
  featureName = inputItem[attributeName]
- values = this.parseProperty(inputName, inputItem.$)
+ values = this.parseProperty(inputName, inputItem.$, inputElem)
  }
  // `values` is always an array as an array is a return value of `parseProperty`
  if (values.length > 0) {
@@ -257,7 +257,8 @@ class ImportData {
  * @param {Lemma[]} lemmas the lemma objects
  */
  overrideInflectionFeatureIfRequired (featureType, inflection, lemmas) {
- if (this.inflectionOverrides.includes(featureType)) {
+ if (this.inflectionOverrides[featureType] &&
+ this.inflectionOverrides[featureType](inflection,lemmas)) {
  for (const lemma of lemmas.filter(l => l.features[featureType])) {
  inflection.addFeature(lemma.features[featureType])
  }

packages/client-adapters/src/adapters/tufts/transform-adapter.js

@@ -206,7 +206,7 @@ class TransformAdapter {
  for (const f of this.config.featuresArrayAll) {
  try {
  mappingData.mapFeature(inflection, inflectionJSON, ...f, this.config.allowUnknownValues)
- mappingData.overrideInflectionFeatureIfRequired(f[1], inflection, lemmas)
+ mappingData.overrideInflectionFeatureIfRequired(Feature.types[f[1]], inflection, lemmas)
  } catch (e) {
  // quietly continue
  }
@@ -216,7 +216,7 @@ class TransformAdapter {
  for (const f of this.config.attributeBasedFeatures) {
  try {
  mappingData.mapFeatureByAttribute(inflection, inflectionJSON, ...f, this.config.allowUnknownValues)
- mappingData.overrideInflectionFeatureIfRequired(f[1], inflection, lemmas)
+ mappingData.overrideInflectionFeatureIfRequired(Feature.types[f[1]], inflection, lemmas)
  } catch (e) {
  // quietly continue
  }

packages/client-adapters/tests/adapters/tufts/morpheusgrc.test.js

@@ -0,0 +1,21 @@
+/* eslint-env jest */
+import { Constants } from 'alpheios-data-models'
+import Morpheus from '@clAdapters/adapters/tufts/engine/morpheusgrc'
+
+describe('morpheusgrc.test.js', () => {
+
+ beforeEach(() => {
+ })
+ afterEach(() => {
+ jest.resetModules()
+ })
+ afterAll(() => {
+ jest.clearAllMocks()
+ })
+
+ it('1 Morpheusgrc Engine - sets pofs for irregular pronoun ti/s', () => {
+ let mockInput = { hdwd: { $: "τίς"} }
+ let parsed = Morpheus.parseProperty('pofs', 'irregular',mockInput)
+ expect(parsed).toEqual([Constants.POFS_PRONOUN])
+ })
+})

packages/client-adapters/tests/adapters/tufts/tufts-example.test.js

@@ -468,4 +468,44 @@ describe('tufts-example.test.js', () => {
  expect(homonym.lexemes[0].lemma.features.source.value).not.toEqual(expect.stringMatching(/from sedra.bethmardutho.org/))
  })
 
+ it('24 TuftsExample - correct results for greek τίς', async () => {
+ let word = 'τίς'
+ let res = Fixture.getFixtureRes({
+ langCode: 'grc', adapter: 'tufts', word: word
+ })
+ let adapter = new AlpheiosTuftsAdapter({
+ category: 'morphology',
+ adapterName: 'tufts',
+ method: 'getHomonym',
+ sourceData: res
+ })
+ let homonym = await adapter.getHomonym(Constants.LANG_GREEK,word)
+ expect(homonym.lexemes.length).toEqual(2)
+ expect(homonym.lexemes[0].lemma.features['part of speech'].value).toEqual('pronoun')
+ expect(homonym.lexemes[1].lemma.features['part of speech'].value).toEqual('pronoun')
+ expect(homonym.lexemes[0].inflections[0]['part of speech'].value).toEqual("pronoun")
+ expect(homonym.lexemes[0].inflections[1]['part of speech'].value).toEqual("pronoun")
+ expect(homonym.lexemes[1].inflections[0]['part of speech'].value).toEqual("pronoun")
+ expect(homonym.lexemes[1].inflections[1]['part of speech'].value).toEqual("pronoun")
+ })
+
+ it('25 TuftsExample - correct results for greek τίνος', async () => {
+ let word = 'τίνος'
+ let res = Fixture.getFixtureRes({
+ langCode: 'grc', adapter: 'tufts', word: word
+ })
+ let adapter = new AlpheiosTuftsAdapter({
+ category: 'morphology',
+ adapterName: 'tufts',
+ method: 'getHomonym',
+ sourceData: res
+ })
+ let homonym = await adapter.getHomonym(Constants.LANG_GREEK,word)
+ expect(homonym.lexemes.length).toEqual(2)
+ expect(homonym.lexemes[0].lemma.features['part of speech'].value).toEqual('pronoun')
+ expect(homonym.lexemes[1].lemma.features['part of speech'].value).toEqual('pronoun')
+ expect(homonym.lexemes[0].inflections[0]['part of speech'].value).toEqual("pronoun")
+ expect(homonym.lexemes[1].inflections[0]['part of speech'].value).toEqual("pronoun")
+ })
+
 })