fix for implicit contractions output

changelog.md

@@ -11,6 +11,10 @@ While all _Major_ releases should be reviewed, our only two _large_ releases are
 <!-- #### [Unreleased] 
 -->
 
+#### 13.9.1 [Feb 2021]
+- **[fix]** - matches over a contraction 
+- **[new]** - add 'implicit' text output
+
 #### 13.9.0 [Feb 2021]
 - **[new]** - World.addConjugations() method
 - **[new]** - World.addPlurals() method

scratch.js

@@ -17,8 +17,20 @@ nlp.extend(require('./plugins/penn-tags/src'))
   done  - false
 */
 
-let doc = nlp(`haven't done`)
-doc.match(`have done`).debug()
+// let doc = nlp(`haven't done`)
+// doc.match(`have done`).debug()
+
+// let doc = nlp(`i haven't done it`)
+// let m = doc.match(`not done`)
+// console.log(`|${m.text()}|`)
+// m.debug()
+
+let doc = nlp(`is not foobar isn't`)
+doc = doc.terms().unique().debug()
+// console.log('|' + nlp(`isn't`).text('implicit') + '|')
+
+// console.log(`|${doc.text()}|`)
+
 // doc.termList().forEach(t => {
 //   console.log(t.text, t.isImplicit())
 // })

src/Doc/methods/transform/01-sort.js

@@ -131,7 +131,7 @@ exports.unique = function () {
   let list = [].concat(this.list)
   let obj = {}
   list = list.filter(p => {
-    let str = p.text('reduced').trim()
+    let str = p.text('reduced').trim() || p.text('implicit').trim()
     if (obj.hasOwnProperty(str) === true) {
       return false
     }

src/Phrase/methods/02-text.js

@@ -36,6 +36,13 @@ exports.text = function (options = {}, isFirst, isLast) {
         implicit: true,
         reduced: true,
       }
+    } else if (options === 'implicit') {
+      options = {
+        punctuation: true,
+        implicit: true,
+        whitespace: true,
+        trim: true,
+      }
     } else if (options === 'root') {
       options = {
         titlecase: false,
@@ -57,6 +64,10 @@ exports.text = function (options = {}, isFirst, isLast) {
     isFull = true
   }
   let text = terms.reduce((str, t, i) => {
+    // don't output intro space for a contraction-match  i'm good => "[am] good"
+    if (i === 0 && t.text === '' && t.implicit !== null && !options.implicit) {
+      return str
+    }
     options.last = isLast && i === terms.length - 1
     let showPre = true
     let showPost = true
@@ -71,6 +82,7 @@ exports.text = function (options = {}, isFirst, isLast) {
       }
     }
     let txt = t.textOut(options, showPre, showPost)
+    // console.log(terms)
     // if (options.titlecase && i === 0) {
     // txt = titleCase(txt)
     // }

src/Phrase/methods/05-json.js

@@ -14,6 +14,9 @@ exports.json = function (options = {}, world) {
   if (options.reduced) {
     res.reduced = this.text('reduced')
   }
+  if (options.implicit) {
+    res.implicit = this.text('implicit')
+  }
   if (options.root) {
     res.root = this.text('root')
   }

tests/match/match-contraction.test.js

@@ -10,7 +10,7 @@ test('match-contractions', function (t) {
   t.equal(m.text(), `haven't`, 'first-half-found')
 
   m = doc.match(`not done`)
-  t.equal(m.text(), ` done`, 'second-half-found')
+  t.equal(m.text(), `done`, 'second-half-found')
 
   m = doc.match(`haven't`)
   t.equal(m.text(), `haven't`, 'match-contraction')
@@ -80,7 +80,7 @@ test('contraction-optional', function (t) {
   t.equal(m.text(), `i'm`, `i am?`)
 
   m = doc.match(`am glad?`)
-  t.equal(m.text(), ` glad`, `am glad?`)
+  t.equal(m.text(), `glad`, `am glad?`)
 
   m = doc.match(`i am? glad`)
   t.equal(m.text(), `i'm glad`, `i am? glad`)

tests/output/text.test.js

@@ -71,6 +71,14 @@ test('text-reduced', function (t) {
   t.end()
 })
 
+test('text-implicit', function (t) {
+  let doc = nlp(`My dog isn't good, he's the best!`)
+  const str = 'My dog is not good, he is the best!'
+  t.equal(doc.json({ implicit: true })[0].implicit, str, 'json(implicit)')
+  t.equal(doc.text('implicit'), str, 'text(implicit): ')
+  t.end()
+})
+
 test('text-root', function (t) {
   let doc = nlp(`My dog LOVES pizza, and grapes!!`)
   const str = 'my dog love pizza and grape'

tests/unique.test.js

@@ -26,5 +26,6 @@ test('unique-normalize', function (t) {
   doc = nlp(`is not isn't`)
   doc = doc.terms().unique()
   t.equal(doc.text(), 'is not', 'normalize-contraction')
+  t.equal(doc.length, 2, 'implicit words are uniqued')
   t.end()
 })