chore: import upstream snapshot with attribution

plugins/speech/README.md

@@ -0,0 +1,88 @@
+<div align="center">
+  <img src="https://cloud.githubusercontent.com/assets/399657/23590290/ede73772-01aa-11e7-8915-181ef21027bc.png" />
+
+  <div>pronounciation metadata plugin for <a href="https://github.com/spencermountain/compromise/">compromise</a></div> 
+
+  <!-- npm version -->
+  <a href="https://npmjs.org/package/compromise-speech">
+    <img src="https://img.shields.io/npm/v/compromise-speech.svg?style=flat-square" />
+  </a>
+  
+  <!-- file size -->
+  <a href="https://unpkg.com/compromise-speech/builds/compromise-speech.min.js">
+    <img src="https://badge-size.herokuapp.com/spencermountain/compromise/master/plugins/speech/builds/compromise-speech.min.js" />
+  </a>
+
+  <div align="center">
+    <code>npm install compromise-speech</code>
+  </div>
+</div>
+
+<!-- spacer -->
+<img height="30px" src="https://user-images.githubusercontent.com/399657/68221862-17ceb980-ffb8-11e9-87d4-7b30b6488f16.png"/>
+
+### Syllables
+Tokenize a word approximately by it's pronounced syllables. We use a custom rule-based alortithm for this.
+```js
+import plg from 'compromise-speech'
+nlp.extend(plg)
+
+let doc = nlp('seventh millenium. white collar')
+doc.syllables()
+[ [ 'se', 'venth', 'mil', 'le', 'nium' ], [ 'white', 'col', 'lar' ] ]
+```
+
+Alternatively, if you'd like to combine syllable information with other properties, you can use `.compute('syllables')`
+```js
+let doc = nlp('edmonton oilers')
+doc.compute('syllables')
+doc.json()
+/*[{
+  "terms": [
+    {
+      "normal": "edmonton",
+      "syllables": ["ed", "mon", "ton"]
+    },
+    {
+      "normal": "oilers",
+      "syllables": ["oi", "lers"]
+    }
+  ]
+}]
+*/
+```
+
+### SoundsLike
+Generate an expected pronounciation for the word, as a normalized string.
+We use the [metaphone implementation](https://en.wikipedia.org/wiki/Metaphone), which is a simple rule-based pronounciation system.
+
+```js
+import plg from 'compromise-speech'
+nlp.extend(plg)
+
+let doc = nlp('seventh millenium. white collar')
+doc.soundsLike()
+// [ [ 'sefenth', 'milenium' ], [ 'wite', 'kolar' ] ]
+```
+
+Alternatively, if you'd like to combine syllable information with other properties, you can use `.compute('syllables')`
+```js
+let doc = nlp('edmonton oilers')
+doc.compute('soundsLike')
+doc.json()
+/*[{
+  "terms": [
+    {
+      "normal": "edmonton",
+      "soundsLike": "etmonton"
+    },
+    {
+      "normal": "oilers",
+      "soundsLike": "oilers"
+    }
+  ]
+}]
+*/
+```
+
+MIT

plugins/speech/builds/compromise-speech.cjs

@@ -0,0 +1,386 @@
+(function (global, factory) {
+  typeof exports === 'object' && typeof module !== 'undefined' ? module.exports = factory() :
+  typeof define === 'function' && define.amd ? define(factory) :
+  (global = typeof globalThis !== 'undefined' ? globalThis : global || self, global.compromiseSpeech = factory());
+})(this, (function () { 'use strict';
+
+  //individual manipulations of the text
+  const transformations = {
+    dedup: (s) => {
+      return s.replace(/([^c])\1/g, '$1')
+    },
+    dropInitialLetters: (s) => {
+      if (s.match(/^(kn|gn|pn|ae|wr)/)) {
+        return s.substring(1, s.length - 1)
+      }
+      return s
+    },
+    dropBafterMAtEnd: (s) => {
+      return s.replace(/mb$/, 'm')
+    },
+    cchange: (s) => {
+      s = s.replace(/([^s]|^)(c)(h)/g, '$1x$3').trim();
+      s = s.replace(/cia/g, 'xia');
+      s = s.replace(/c([iey])/g, 's$1');
+      return s.replace(/c/g, 'k')
+    },
+    dchange: (s) => {
+      s = s.replace(/d(ge|gy|gi)/g, 'j$1');
+      return s.replace(/d/g, 't')
+    },
+    dropG: (s) => {
+      // Drop 'G' if followed by 'H' and 'H' is not at the end or before a vowel. 
+      s = s.replace(/gh(^$|[^aeiou])/g, 'h$1');//eslint-disable-line
+      // Drop 'G' if followed by 'N' or 'NED' and is at the end.
+      return s.replace(/g(n|ned)$/g, '$1')
+    },
+    changeG: (s) => {
+      s = s.replace(/gh/g, 'f');
+      s = s.replace(/([^g]|^)(g)([iey])/g, '$1j$3');
+      s = s.replace(/gg/g, 'g');
+      return s.replace(/g/g, 'k')
+    },
+    dropH: (s) => {
+      return s.replace(/([aeiou])h([^aeiou]|$)/g, '$1$2')
+    },
+    changeCK: (s) => {
+      return s.replace(/ck/g, 'k')
+    },
+    changePH: (s) => {
+      return s.replace(/ph/g, 'f')
+    },
+    changeQ: (s) => {
+      return s.replace(/q/g, 'k')
+    },
+    changeS: (s) => {
+      return s.replace(/s(h|io|ia)/g, 'x$1')
+    },
+    changeT: (s) => {
+      s = s.replace(/t(ia[^n]|io)/g, 'x$1');
+      // return s.replace(/th/, '0')
+      return s
+    },
+    dropT: (s) => {
+      return s.replace(/tch/g, 'ch')
+    },
+    changeV: (s) => {
+      return s.replace(/v/g, 'f')
+    },
+    changeWH: (s) => {
+      return s.replace(/^wh/, 'w')
+    },
+    dropW: (s) => {
+      return s.replace(/w([^aeiou]|$)/g, '$1')
+    },
+    changeX: (s) => {
+      s = s.replace(/^x/, 's');
+      return s.replace(/x/g, 'ks')
+    },
+    dropY: (s) => {
+      return s.replace(/y([^aeiou]|$)/g, '$1')
+    },
+    changeZ: (s) => {
+      return s.replace(/z/, 's')
+    },
+    dropVowels: (s) => {
+      return s //.charAt(0) + s.substr(1, s.length).replace(/[aeiou]/g, '');
+    },
+  };
+  var m = transformations;
+
+  //a js version of the metaphone (#1) algorithm
+
+  const metaphone = function (s) {
+    s = m.dedup(s);
+    s = m.dropInitialLetters(s);
+    s = m.dropBafterMAtEnd(s);
+    s = m.changeCK(s);
+    s = m.cchange(s);
+    s = m.dchange(s);
+    s = m.dropG(s);
+    s = m.changeG(s);
+    s = m.dropH(s);
+    s = m.changePH(s);
+    s = m.changeQ(s);
+    s = m.changeS(s);
+    s = m.changeX(s);
+    s = m.changeT(s);
+    s = m.dropT(s);
+    s = m.changeV(s);
+    s = m.changeWH(s);
+    s = m.dropW(s);
+    s = m.dropY(s);
+    s = m.changeZ(s);
+    s = m.dropVowels(s);
+    return s.trim()
+  };
+
+  var metaphone$1 = metaphone;
+
+  const soundsLike = function (view) {
+    view.docs.forEach(terms => {
+      terms.forEach(term => {
+        term.soundsLike = metaphone$1(term.normal || term.text);
+      });
+    });
+  };
+
+  var soundsLike$1 = soundsLike;
+
+  const starts_with_single_vowel_combos = /^(eu)/i;
+  const joining_consonant_vowel = /^[^aeiou]e([^d]|$)/;
+  const cvcv_same_consonant = /^([^aeiouy])[aeiouy]\1[aeiouy]/;
+  const cvcv_same_vowel = /^[^aeiouy]([aeiouy])[^aeiouy]\1/;
+  const cvcv_known_consonants = /^([tg][aeiouy]){2}/;
+  const only_one_or_more_c = /^[^aeiouy]+$/;
+
+  const ends_with_vowel$1 = /[aeiouy]$/;
+  const starts_with_consonant_vowel$1 = /^[^aeiouy]h?[aeiouy]/;
+
+  const ones = [
+    /^[^aeiou]?ion/,
+    /^[^aeiou]?ised/,
+    /^[^aeiou]?iled/,
+
+    // -ing, -ent
+    /[aeiou]n[gt]$/,
+
+    // -ate, -age
+    /\wa[gt]e$/,
+  ];
+
+  //suffix fixes
+  const postprocess = function (arr) {
+    //trim whitespace
+    arr = arr.map(function (w) {
+      return w.trim()
+    });
+    arr = arr.filter(function (w) {
+      return w !== ''
+    });
+    // if (arr.length > 2) {
+    //   return arr;
+    // }
+    let l = arr.length;
+    if (l > 1) {
+      let suffix = arr[l - 2] + arr[l - 1];
+      for (let i = 0; i < ones.length; i++) {
+        if (suffix.match(ones[i])) {
+          arr[l - 2] = arr[l - 2] + arr[l - 1];
+          arr.pop();
+        }
+      }
+    }
+
+    // since the open syllable detection is overzealous,
+    // sometimes need to rejoin incorrect splits
+    if (arr.length > 1) {
+      let first_is_open =
+        (arr[0].length === 1 || arr[0].match(starts_with_consonant_vowel$1)) &&
+        arr[0].match(ends_with_vowel$1);
+      let second_is_joining = arr[1].match(joining_consonant_vowel);
+
+      if (first_is_open && second_is_joining) {
+        let possible_combination = arr[0] + arr[1];
+        let probably_separate_syllables =
+          possible_combination.match(cvcv_same_consonant) ||
+          possible_combination.match(cvcv_same_vowel) ||
+          possible_combination.match(cvcv_known_consonants);
+
+        if (!probably_separate_syllables) {
+          arr[0] = arr[0] + arr[1];
+          arr.splice(1, 1);
+        }
+      }
+    }
+
+    if (arr.length > 1) {
+      let second_to_last_is_open =
+        arr[arr.length - 2].match(starts_with_consonant_vowel$1) &&
+        arr[arr.length - 2].match(ends_with_vowel$1);
+      let last_is_joining =
+        arr[arr.length - 1].match(joining_consonant_vowel) &&
+        ones.every(re => !arr[arr.length - 1].match(re));
+
+      if (second_to_last_is_open && last_is_joining) {
+        let possible_combination = arr[arr.length - 2] + arr[arr.length - 1];
+        let probably_separate_syllables =
+          possible_combination.match(cvcv_same_consonant) ||
+          possible_combination.match(cvcv_same_vowel) ||
+          possible_combination.match(cvcv_known_consonants);
+
+        if (!probably_separate_syllables) {
+          arr[arr.length - 2] = arr[arr.length - 2] + arr[arr.length - 1];
+          arr.splice(arr.length - 1, 1);
+        }
+      }
+    }
+
+    if (arr.length > 1) {
+      let single = arr[0] + arr[1];
+      if (single.match(starts_with_single_vowel_combos)) {
+        arr[0] = single;
+        arr.splice(1, 1);
+      }
+    }
+
+    if (arr.length > 1) {
+      if (arr[arr.length - 1].match(only_one_or_more_c)) {
+        arr[arr.length - 2] = arr[arr.length - 2] + arr[arr.length - 1];
+        arr.splice(arr.length - 1, 1);
+      }
+    }
+
+    return arr
+  };
+  var postProcess = postprocess;
+
+  //chop a string into pronounced syllables
+
+  const all_spaces = / +/g;
+  const ends_with_vowel = /[aeiouy]$/;
+  const starts_with_consonant_vowel = /^[^aeiouy]h?[aeiouy]/;
+  const starts_with_e_then_specials = /^e[sm]/;
+  const starts_with_e = /^e/;
+  const ends_with_noisy_vowel_combos = /(eo|eu|ia|oa|ua|ui)$/i;
+  const aiouy = /[aiouy]/;
+  const ends_with_ee = /ee$/;
+  // const whitespace_dash = /\s\-/
+
+  //method is nested because it's called recursively
+  const doWord = function (w) {
+    let all = [];
+    let chars = w.split('');
+    let before = '';
+    let after = '';
+    let current = '';
+    for (let i = 0; i < chars.length; i++) {
+      before = chars.slice(0, i).join('');
+      current = chars[i];
+      after = chars.slice(i + 1, chars.length).join('');
+      let candidate = before + chars[i];
+
+      //it's a consonant that comes after a vowel
+      if (before.match(ends_with_vowel) && !current.match(ends_with_vowel)) {
+        if (after.match(starts_with_e_then_specials)) {
+          candidate += 'e';
+          after = after.replace(starts_with_e, '');
+        }
+        all.push(candidate);
+        return all.concat(doWord(after))
+      }
+
+      //unblended vowels ('noisy' vowel combinations)
+      if (candidate.match(ends_with_noisy_vowel_combos)) {
+        //'io' is noisy, not in 'ion'
+        all.push(before);
+        all.push(current);
+        return all.concat(doWord(after)) //recursion
+      }
+
+      // if candidate is followed by a CV, assume consecutive open syllables
+      if (candidate.match(ends_with_vowel) && after.match(starts_with_consonant_vowel)) {
+        all.push(candidate);
+        return all.concat(doWord(after))
+      }
+    }
+    //if still running, end last syllable
+    if (w.match(aiouy) || w.match(ends_with_ee)) {
+      //allow silent trailing e
+      all.push(w);
+    } else if (w) {
+      let last = all.length - 1;
+      if (last < 0) {
+        last = 0;
+      }
+      all[last] = (all[last] || '') + w; //append it to the last one
+    }
+    return all
+  };
+
+  let syllables$2 = function (str) {
+    let all = [];
+    if (!str) {
+      return all
+    }
+    str = str.replace(/[.,?]/g, '');
+    str.split(all_spaces).map(s => {
+      all = all.concat(doWord(s));
+    });
+
+    // str.split(whitespace_dash).forEach(doWord)
+    all = postProcess(all);
+
+    //for words like 'tree' and 'free'
+    if (all.length === 0) {
+      all = [str];
+    }
+    //filter blanks
+    all = all.filter(s => s);
+
+    return all
+  };
+
+  // console.log(syllables('civilised'))
+
+  var getSyllables = syllables$2;
+
+  // const defaultObj = { normal: true, text: true, terms: false }
+
+  const syllables = function (view) {
+    view.docs.forEach(terms => {
+      terms.forEach(term => {
+        term.syllables = getSyllables(term.normal || term.text);
+      });
+    });
+  };
+
+  var syllables$1 = syllables;
+
+  var compute = {
+    soundsLike: soundsLike$1,
+    syllables: syllables$1
+  };
+
+  const api = function (View) {
+    /** */
+    View.prototype.syllables = function () {
+      this.compute('syllables');
+      let all = [];
+      this.docs.forEach(terms => {
+        let some = [];
+        terms.forEach(term => {
+          some = some.concat(term.syllables);
+        });
+        if (some.length > 0) {
+          all.push(some);
+        }
+      });
+      return all
+    };
+    /** */
+    View.prototype.soundsLike = function () {
+      this.compute('soundsLike');
+      let all = [];
+      this.docs.forEach(terms => {
+        let some = [];
+        terms.forEach(term => {
+          some = some.concat(term.soundsLike);
+        });
+        if (some.length > 0) {
+          all.push(some);
+        }
+      });
+      return all
+    };
+  };
+  var api$1 = api;
+
+  var plugin = {
+    api: api$1,
+    compute
+  };
+
+  return plugin;
+
+}));

plugins/speech/builds/compromise-speech.min.js

@@ -0,0 +1 @@
+var e,t;e=this,t=function(){var e={dedup:e=>e.replace(/([^c])\1/g,"$1"),dropInitialLetters:e=>e.match(/^(kn|gn|pn|ae|wr)/)?e.substring(1,e.length-1):e,dropBafterMAtEnd:e=>e.replace(/mb$/,"m"),cchange:e=>(e=(e=(e=e.replace(/([^s]|^)(c)(h)/g,"$1x$3").trim()).replace(/cia/g,"xia")).replace(/c([iey])/g,"s$1")).replace(/c/g,"k"),dchange:e=>(e=e.replace(/d(ge|gy|gi)/g,"j$1")).replace(/d/g,"t"),dropG:e=>(e=e.replace(/gh(^$|[^aeiou])/g,"h$1")).replace(/g(n|ned)$/g,"$1"),changeG:e=>(e=(e=(e=e.replace(/gh/g,"f")).replace(/([^g]|^)(g)([iey])/g,"$1j$3")).replace(/gg/g,"g")).replace(/g/g,"k"),dropH:e=>e.replace(/([aeiou])h([^aeiou]|$)/g,"$1$2"),changeCK:e=>e.replace(/ck/g,"k"),changePH:e=>e.replace(/ph/g,"f"),changeQ:e=>e.replace(/q/g,"k"),changeS:e=>e.replace(/s(h|io|ia)/g,"x$1"),changeT:e=>e=e.replace(/t(ia[^n]|io)/g,"x$1"),dropT:e=>e.replace(/tch/g,"ch"),changeV:e=>e.replace(/v/g,"f"),changeWH:e=>e.replace(/^wh/,"w"),dropW:e=>e.replace(/w([^aeiou]|$)/g,"$1"),changeX:e=>(e=e.replace(/^x/,"s")).replace(/x/g,"ks"),dropY:e=>e.replace(/y([^aeiou]|$)/g,"$1"),changeZ:e=>e.replace(/z/,"s"),dropVowels:e=>e},t=function(t){return t=e.dedup(t),t=e.dropInitialLetters(t),t=e.dropBafterMAtEnd(t),t=e.changeCK(t),t=e.cchange(t),t=e.dchange(t),t=e.dropG(t),t=e.changeG(t),t=e.dropH(t),t=e.changePH(t),t=e.changeQ(t),t=e.changeS(t),t=e.changeX(t),t=e.changeT(t),t=e.dropT(t),t=e.changeV(t),t=e.changeWH(t),t=e.dropW(t),t=e.dropY(t),t=e.changeZ(t),(t=e.dropVowels(t)).trim()};const a=/^(eu)/i,c=/^[^aeiou]e([^d]|$)/,n=/^([^aeiouy])[aeiouy]\1[aeiouy]/,l=/^[^aeiouy]([aeiouy])[^aeiouy]\1/,h=/^([tg][aeiouy]){2}/,o=/^[^aeiouy]+$/,i=/[aeiouy]$/,r=/^[^aeiouy]h?[aeiouy]/,g=[/^[^aeiou]?ion/,/^[^aeiou]?ised/,/^[^aeiou]?iled/,/[aeiou]n[gt]$/,/\wa[gt]e$/];var p=function(e){let t=(e=(e=e.map((function(e){return e.trim()}))).filter((function(e){return""!==e}))).length;if(t>1){let a=e[t-2]+e[t-1];for(let c=0;c<g.length;c++)a.match(g[c])&&(e[t-2]=e[t-2]+e[t-1],e.pop())}if(e.length>1){let t=(1===e[0].length||e[0].match(r))&&e[0].match(i),a=e[1].match(c);if(t&&a){let t=e[0]+e[1];t.match(n)||t.match(l)||t.match(h)||(e[0]=e[0]+e[1],e.splice(1,1))}}if(e.length>1){let t=e[e.length-2].match(r)&&e[e.length-2].match(i),a=e[e.length-1].match(c)&&g.every(t=>!e[e.length-1].match(t));if(t&&a){let t=e[e.length-2]+e[e.length-1];t.match(n)||t.match(l)||t.match(h)||(e[e.length-2]=e[e.length-2]+e[e.length-1],e.splice(e.length-1,1))}}if(e.length>1){let t=e[0]+e[1];t.match(a)&&(e[0]=t,e.splice(1,1))}return e.length>1&&e[e.length-1].match(o)&&(e[e.length-2]=e[e.length-2]+e[e.length-1],e.splice(e.length-1,1)),e};const u=/ +/g,s=/[aeiouy]$/,f=/^[^aeiouy]h?[aeiouy]/,d=/^e[sm]/,m=/^e/,y=/(eo|eu|ia|oa|ua|ui)$/i,$=/[aiouy]/,k=/ee$/,b=function(e){let t=[],a=e.split(""),c="",n="",l="";for(let e=0;e<a.length;e++){c=a.slice(0,e).join(""),l=a[e],n=a.slice(e+1,a.length).join("");let h=c+a[e];if(c.match(s)&&!l.match(s))return n.match(d)&&(h+="e",n=n.replace(m,"")),t.push(h),t.concat(b(n));if(h.match(y))return t.push(c),t.push(l),t.concat(b(n));if(h.match(s)&&n.match(f))return t.push(h),t.concat(b(n))}if(e.match($)||e.match(k))t.push(e);else if(e){let a=t.length-1;a<0&&(a=0),t[a]=(t[a]||"")+e}return t};var x=function(e){let t=[];return e?((e=e.replace(/[.,?]/g,"")).split(u).map(e=>{t=t.concat(b(e))}),t=p(t),0===t.length&&(t=[e]),t=t.filter(e=>e),t):t};return{api:function(e){e.prototype.syllables=function(){this.compute("syllables");let e=[];return this.docs.forEach(t=>{let a=[];t.forEach(e=>{a=a.concat(e.syllables)}),a.length>0&&e.push(a)}),e},e.prototype.soundsLike=function(){this.compute("soundsLike");let e=[];return this.docs.forEach(t=>{let a=[];t.forEach(e=>{a=a.concat(e.soundsLike)}),a.length>0&&e.push(a)}),e}},compute:{soundsLike:function(e){e.docs.forEach(e=>{e.forEach(e=>{e.soundsLike=t(e.normal||e.text)})})},syllables:function(e){e.docs.forEach(e=>{e.forEach(e=>{e.syllables=x(e.normal||e.text)})})}}}},"object"==typeof exports&&"undefined"!=typeof module?module.exports=t():"function"==typeof define&&define.amd?define(t):(e="undefined"!=typeof globalThis?globalThis:e||self).compromiseSpeech=t();

plugins/speech/builds/compromise-speech.mjs

@@ -0,0 +1 @@
+var e={dedup:e=>e.replace(/([^c])\1/g,"$1"),dropInitialLetters:e=>e.match(/^(kn|gn|pn|ae|wr)/)?e.substring(1,e.length-1):e,dropBafterMAtEnd:e=>e.replace(/mb$/,"m"),cchange:e=>(e=(e=(e=e.replace(/([^s]|^)(c)(h)/g,"$1x$3").trim()).replace(/cia/g,"xia")).replace(/c([iey])/g,"s$1")).replace(/c/g,"k"),dchange:e=>(e=e.replace(/d(ge|gy|gi)/g,"j$1")).replace(/d/g,"t"),dropG:e=>(e=e.replace(/gh(^$|[^aeiou])/g,"h$1")).replace(/g(n|ned)$/g,"$1"),changeG:e=>(e=(e=(e=e.replace(/gh/g,"f")).replace(/([^g]|^)(g)([iey])/g,"$1j$3")).replace(/gg/g,"g")).replace(/g/g,"k"),dropH:e=>e.replace(/([aeiou])h([^aeiou]|$)/g,"$1$2"),changeCK:e=>e.replace(/ck/g,"k"),changePH:e=>e.replace(/ph/g,"f"),changeQ:e=>e.replace(/q/g,"k"),changeS:e=>e.replace(/s(h|io|ia)/g,"x$1"),changeT:e=>e=e.replace(/t(ia[^n]|io)/g,"x$1"),dropT:e=>e.replace(/tch/g,"ch"),changeV:e=>e.replace(/v/g,"f"),changeWH:e=>e.replace(/^wh/,"w"),dropW:e=>e.replace(/w([^aeiou]|$)/g,"$1"),changeX:e=>(e=e.replace(/^x/,"s")).replace(/x/g,"ks"),dropY:e=>e.replace(/y([^aeiou]|$)/g,"$1"),changeZ:e=>e.replace(/z/,"s"),dropVowels:e=>e};var a=function(a){return a=e.dedup(a),a=e.dropInitialLetters(a),a=e.dropBafterMAtEnd(a),a=e.changeCK(a),a=e.cchange(a),a=e.dchange(a),a=e.dropG(a),a=e.changeG(a),a=e.dropH(a),a=e.changePH(a),a=e.changeQ(a),a=e.changeS(a),a=e.changeX(a),a=e.changeT(a),a=e.dropT(a),a=e.changeV(a),a=e.changeWH(a),a=e.dropW(a),a=e.dropY(a),a=e.changeZ(a),(a=e.dropVowels(a)).trim()};const t=/^(eu)/i,c=/^[^aeiou]e([^d]|$)/,n=/^([^aeiouy])[aeiouy]\1[aeiouy]/,l=/^[^aeiouy]([aeiouy])[^aeiouy]\1/,h=/^([tg][aeiouy]){2}/,r=/^[^aeiouy]+$/,o=/[aeiouy]$/,g=/^[^aeiouy]h?[aeiouy]/,i=[/^[^aeiou]?ion/,/^[^aeiou]?ised/,/^[^aeiou]?iled/,/[aeiou]n[gt]$/,/\wa[gt]e$/];var p=function(e){let a=(e=(e=e.map((function(e){return e.trim()}))).filter((function(e){return""!==e}))).length;if(a>1){let t=e[a-2]+e[a-1];for(let c=0;c<i.length;c++)t.match(i[c])&&(e[a-2]=e[a-2]+e[a-1],e.pop())}if(e.length>1){let a=(1===e[0].length||e[0].match(g))&&e[0].match(o),t=e[1].match(c);if(a&&t){let a=e[0]+e[1];a.match(n)||a.match(l)||a.match(h)||(e[0]=e[0]+e[1],e.splice(1,1))}}if(e.length>1){let a=e[e.length-2].match(g)&&e[e.length-2].match(o),t=e[e.length-1].match(c)&&i.every(a=>!e[e.length-1].match(a));if(a&&t){let a=e[e.length-2]+e[e.length-1];a.match(n)||a.match(l)||a.match(h)||(e[e.length-2]=e[e.length-2]+e[e.length-1],e.splice(e.length-1,1))}}if(e.length>1){let a=e[0]+e[1];a.match(t)&&(e[0]=a,e.splice(1,1))}return e.length>1&&e[e.length-1].match(r)&&(e[e.length-2]=e[e.length-2]+e[e.length-1],e.splice(e.length-1,1)),e};const u=/ +/g,s=/[aeiouy]$/,d=/^[^aeiouy]h?[aeiouy]/,f=/^e[sm]/,m=/^e/,$=/(eo|eu|ia|oa|ua|ui)$/i,y=/[aiouy]/,k=/ee$/,E=function(e){let a=[],t=e.split(""),c="",n="",l="";for(let e=0;e<t.length;e++){c=t.slice(0,e).join(""),l=t[e],n=t.slice(e+1,t.length).join("");let h=c+t[e];if(c.match(s)&&!l.match(s))return n.match(f)&&(h+="e",n=n.replace(m,"")),a.push(h),a.concat(E(n));if(h.match($))return a.push(c),a.push(l),a.concat(E(n));if(h.match(s)&&n.match(d))return a.push(h),a.concat(E(n))}if(e.match(y)||e.match(k))a.push(e);else if(e){let t=a.length-1;t<0&&(t=0),a[t]=(a[t]||"")+e}return a};var x=function(e){let a=[];return e?((e=e.replace(/[.,?]/g,"")).split(u).map(e=>{a=a.concat(E(e))}),a=p(a),0===a.length&&(a=[e]),a=a.filter(e=>e),a):a};var b={api:function(e){e.prototype.syllables=function(){this.compute("syllables");let e=[];return this.docs.forEach(a=>{let t=[];a.forEach(e=>{t=t.concat(e.syllables)}),t.length>0&&e.push(t)}),e},e.prototype.soundsLike=function(){this.compute("soundsLike");let e=[];return this.docs.forEach(a=>{let t=[];a.forEach(e=>{t=t.concat(e.soundsLike)}),t.length>0&&e.push(t)}),e}},compute:{soundsLike:function(e){e.docs.forEach(e=>{e.forEach(e=>{e.soundsLike=a(e.normal||e.text)})})},syllables:function(e){e.docs.forEach(e=>{e.forEach(e=>{e.syllables=x(e.normal||e.text)})})}}};export{b as default};

plugins/speech/demo/index.html

@@ -0,0 +1,28 @@
+<!DOCTYPE html>
+<html>
+
+<head>
+
+  <script src="https://unpkg.com/compromise"></script>
+  <!-- <script src="https://unpkg.com/compromise-speech"></script> -->
+  <script src="../builds/compromise-speech.min.js"></script>
+
+</head>
+
+<body>
+  <p>compromise-speech demo:</p>
+  <ul>
+    <kbd>chocolate microscopes </kbd>
+    <pre id="res">loading</pre>
+  </ul>
+
+  <script defer>
+    nlp.plugin(compromiseSpeech) // window.compromiseSpeech
+    let txt = document.querySelector('kbd').innerText
+    let doc = nlp(txt)
+    let json = doc.syllables()
+    document.querySelector('#res').innerHTML = JSON.stringify(json, null, 2)
+  </script>
+</body>
+
+</html>

plugins/speech/index.d.ts

@@ -0,0 +1,13 @@
+import nlp from 'compromise'
+
+export interface SpeechMethods {
+  /** estimate spoken phenomes */
+  syllables(): String[][]
+  /** estimate pronounciation information */
+  soundsLike(): String[][]
+}
+
+/** extended compromise lib **/
+declare const nlpSpeech: nlp.TypedPlugin<SpeechMethods>
+
+export default nlpSpeech

plugins/speech/package.json

@@ -0,0 +1,44 @@
+{
+  "name": "compromise-speech",
+  "description": "plugin for nlp-compromise",
+  "version": "0.1.1",
+  "author": "Spencer Kelly <spencermountain@gmail.com> (http://spencermounta.in)",
+  "main": "./src/plugin.js",
+  "unpkg": "./builds/compromise-speech.min.js",
+  "module": "./builds/compromise-speech.mjs",
+  "type": "module",
+  "sideEffects": false,
+  "types": "./index.d.ts",
+  "exports": {
+    ".": {
+      "import": "./src/plugin.js",
+      "require": "./builds/compromise-speech.cjs",
+      "types": "./index.d.ts"
+    }
+  },
+  "repository": {
+    "type": "git",
+    "url": "git://github.com/spencermountain/compromise.git"
+  },
+  "homepage": "https://github.com/spencermountain/compromise/tree/master/plugins/speech",
+  "scripts": {
+    "test": "tape \"./tests/**/*.test.js\" | tap-dancer  --color always",
+    "testb": "TESTENV=prod tape \"./tests/**/*.test.js\" | tap-dancer  --color always",
+    "watch": "node --watch ./scratch.js",
+    "build": "rollup -c --silent"
+  },
+  "eslintIgnore": [
+    "builds/*.js"
+  ],
+  "files": [
+    "builds/",
+    "src/",
+    "index.d.ts"
+  ],
+  "peerDependencies": {
+    "compromise": ">=14.0.0"
+  },
+  "dependencies": {},
+  "devDependencies": {},
+  "license": "MIT"
+}

plugins/speech/scratch.js

@@ -0,0 +1,17 @@
+/* eslint-disable no-console, no-unused-vars */
+import nlp from '../../src/three.js'
+
+import speechPlugin from './src/plugin.js'
+nlp.plugin(speechPlugin)
+// nlp.verbose(true)
+// nlp.verbose('date')
+
+let txt = ''
+txt = 'seventh millenium. white collar'
+
+// let doc = nlp(txt).compute(['soundsLike', 'syllables'])
+// console.dir(doc.json()[0], { depth: 5 })
+
+const doc = nlp('calgary')
+doc.compute('soundsLike')
+console.log(JSON.stringify(doc.json()[0], null, 2))

plugins/speech/src/api.js

@@ -0,0 +1,33 @@
+const api = function (View) {
+  /** */
+  View.prototype.syllables = function () {
+    this.compute('syllables')
+    const all = []
+    this.docs.forEach(terms => {
+      let some = []
+      terms.forEach(term => {
+        some = some.concat(term.syllables)
+      })
+      if (some.length > 0) {
+        all.push(some)
+      }
+    })
+    return all
+  }
+  /** */
+  View.prototype.soundsLike = function () {
+    this.compute('soundsLike')
+    const all = []
+    this.docs.forEach(terms => {
+      let some = []
+      terms.forEach(term => {
+        some = some.concat(term.soundsLike)
+      })
+      if (some.length > 0) {
+        all.push(some)
+      }
+    })
+    return all
+  }
+}
+export default api

plugins/speech/src/compute/index.js

@@ -0,0 +1,7 @@
+import soundsLike from './soundsLike/index.js'
+import syllables from './syllables/index.js'
+
+export default {
+  soundsLike,
+  syllables
+}

plugins/speech/src/compute/soundsLike/index.js

@@ -0,0 +1,11 @@
+import metaphone from './metaphone.js'
+
+const soundsLike = function (view) {
+  view.docs.forEach(terms => {
+    terms.forEach(term => {
+      term.soundsLike = metaphone(term.normal || term.text)
+    })
+  })
+}
+
+export default soundsLike

plugins/speech/src/compute/soundsLike/metaphone.js

@@ -0,0 +1,32 @@
+//a js version of the metaphone (#1) algorithm
+//adapted from the work of Chris Umbel
+// https://github.com/NaturalNode/natural/blob/master/lib/natural/phonetics/metaphone.js
+
+import m from './transformations.js'
+
+const metaphone = function (s) {
+  s = m.dedup(s)
+  s = m.dropInitialLetters(s)
+  s = m.dropBafterMAtEnd(s)
+  s = m.changeCK(s)
+  s = m.cchange(s)
+  s = m.dchange(s)
+  s = m.dropG(s)
+  s = m.changeG(s)
+  s = m.dropH(s)
+  s = m.changePH(s)
+  s = m.changeQ(s)
+  s = m.changeS(s)
+  s = m.changeX(s)
+  s = m.changeT(s)
+  s = m.dropT(s)
+  s = m.changeV(s)
+  s = m.changeWH(s)
+  s = m.dropW(s)
+  s = m.dropY(s)
+  s = m.changeZ(s)
+  s = m.dropVowels(s)
+  return s.trim()
+}
+
+export default metaphone

plugins/speech/src/compute/soundsLike/transformations.js

@@ -0,0 +1,83 @@
+//individual manipulations of the text
+const transformations = {
+  dedup: (s) => {
+    return s.replace(/([^c])\1/g, '$1')
+  },
+  dropInitialLetters: (s) => {
+    if (s.match(/^(kn|gn|pn|ae|wr)/)) {
+      return s.substring(1, s.length - 1)
+    }
+    return s
+  },
+  dropBafterMAtEnd: (s) => {
+    return s.replace(/mb$/, 'm')
+  },
+  cchange: (s) => {
+    s = s.replace(/([^s]|^)(c)(h)/g, '$1x$3').trim()
+    s = s.replace(/cia/g, 'xia')
+    s = s.replace(/c([iey])/g, 's$1')
+    return s.replace(/c/g, 'k')
+  },
+  dchange: (s) => {
+    s = s.replace(/d(ge|gy|gi)/g, 'j$1')
+    return s.replace(/d/g, 't')
+  },
+  dropG: (s) => {
+    // Drop 'G' if followed by 'H' and 'H' is not at the end or before a vowel. 
+    s = s.replace(/gh(^$|[^aeiou])/g, 'h$1')//eslint-disable-line
+    // Drop 'G' if followed by 'N' or 'NED' and is at the end.
+    return s.replace(/g(n|ned)$/g, '$1')
+  },
+  changeG: (s) => {
+    s = s.replace(/gh/g, 'f')
+    s = s.replace(/([^g]|^)(g)([iey])/g, '$1j$3')
+    s = s.replace(/gg/g, 'g')
+    return s.replace(/g/g, 'k')
+  },
+  dropH: (s) => {
+    return s.replace(/([aeiou])h([^aeiou]|$)/g, '$1$2')
+  },
+  changeCK: (s) => {
+    return s.replace(/ck/g, 'k')
+  },
+  changePH: (s) => {
+    return s.replace(/ph/g, 'f')
+  },
+  changeQ: (s) => {
+    return s.replace(/q/g, 'k')
+  },
+  changeS: (s) => {
+    return s.replace(/s(h|io|ia)/g, 'x$1')
+  },
+  changeT: (s) => {
+    s = s.replace(/t(ia[^n]|io)/g, 'x$1')
+    // return s.replace(/th/, '0')
+    return s
+  },
+  dropT: (s) => {
+    return s.replace(/tch/g, 'ch')
+  },
+  changeV: (s) => {
+    return s.replace(/v/g, 'f')
+  },
+  changeWH: (s) => {
+    return s.replace(/^wh/, 'w')
+  },
+  dropW: (s) => {
+    return s.replace(/w([^aeiou]|$)/g, '$1')
+  },
+  changeX: (s) => {
+    s = s.replace(/^x/, 's')
+    return s.replace(/x/g, 'ks')
+  },
+  dropY: (s) => {
+    return s.replace(/y([^aeiou]|$)/g, '$1')
+  },
+  changeZ: (s) => {
+    return s.replace(/z/, 's')
+  },
+  dropVowels: (s) => {
+    return s //.charAt(0) + s.substr(1, s.length).replace(/[aeiou]/g, '');
+  },
+}
+export default transformations

plugins/speech/src/compute/syllables/index.js

@@ -0,0 +1,13 @@
+import getSyllables from './syllables.js'
+
+// const defaultObj = { normal: true, text: true, terms: false }
+
+const syllables = function (view) {
+  view.docs.forEach(terms => {
+    terms.forEach(term => {
+      term.syllables = getSyllables(term.normal || term.text)
+    })
+  })
+}
+
+export default syllables

plugins/speech/src/compute/syllables/postProcess.js

@@ -0,0 +1,107 @@
+const starts_with_single_vowel_combos = /^(eu)/i
+const joining_consonant_vowel = /^[^aeiou]e([^d]|$)/
+const cvcv_same_consonant = /^([^aeiouy])[aeiouy]\1[aeiouy]/
+const cvcv_same_vowel = /^[^aeiouy]([aeiouy])[^aeiouy]\1/
+const cvcv_known_consonants = /^([tg][aeiouy]){2}/
+const only_one_or_more_c = /^[^aeiouy]+$/
+
+const ends_with_vowel = /[aeiouy]$/
+const starts_with_consonant_vowel = /^[^aeiouy]h?[aeiouy]/
+
+const ones = [
+  /^[^aeiou]?ion/,
+  /^[^aeiou]?ised/,
+  /^[^aeiou]?iled/,
+
+  // -ing, -ent
+  /[aeiou]n[gt]$/,
+
+  // -ate, -age
+  /\wa[gt]e$/,
+]
+
+//suffix fixes
+const postprocess = function (arr) {
+  //trim whitespace
+  arr = arr.map(function (w) {
+    return w.trim()
+  })
+  arr = arr.filter(function (w) {
+    return w !== ''
+  })
+  // if (arr.length > 2) {
+  //   return arr;
+  // }
+  const l = arr.length
+  if (l > 1) {
+    const suffix = arr[l - 2] + arr[l - 1]
+    for (let i = 0; i < ones.length; i++) {
+      if (suffix.match(ones[i])) {
+        arr[l - 2] = arr[l - 2] + arr[l - 1]
+        arr.pop()
+      }
+    }
+  }
+
+  // since the open syllable detection is overzealous,
+  // sometimes need to rejoin incorrect splits
+  if (arr.length > 1) {
+    const first_is_open =
+      (arr[0].length === 1 || arr[0].match(starts_with_consonant_vowel)) &&
+      arr[0].match(ends_with_vowel)
+    const second_is_joining = arr[1].match(joining_consonant_vowel)
+
+    if (first_is_open && second_is_joining) {
+      const possible_combination = arr[0] + arr[1]
+      const probably_separate_syllables =
+        possible_combination.match(cvcv_same_consonant) ||
+        possible_combination.match(cvcv_same_vowel) ||
+        possible_combination.match(cvcv_known_consonants)
+
+      if (!probably_separate_syllables) {
+        arr[0] = arr[0] + arr[1]
+        arr.splice(1, 1)
+      }
+    }
+  }
+
+  if (arr.length > 1) {
+    const second_to_last_is_open =
+      arr[arr.length - 2].match(starts_with_consonant_vowel) &&
+      arr[arr.length - 2].match(ends_with_vowel)
+    const last_is_joining =
+      arr[arr.length - 1].match(joining_consonant_vowel) &&
+      ones.every(re => !arr[arr.length - 1].match(re))
+
+    if (second_to_last_is_open && last_is_joining) {
+      const possible_combination = arr[arr.length - 2] + arr[arr.length - 1]
+      const probably_separate_syllables =
+        possible_combination.match(cvcv_same_consonant) ||
+        possible_combination.match(cvcv_same_vowel) ||
+        possible_combination.match(cvcv_known_consonants)
+
+      if (!probably_separate_syllables) {
+        arr[arr.length - 2] = arr[arr.length - 2] + arr[arr.length - 1]
+        arr.splice(arr.length - 1, 1)
+      }
+    }
+  }
+
+  if (arr.length > 1) {
+    const single = arr[0] + arr[1]
+    if (single.match(starts_with_single_vowel_combos)) {
+      arr[0] = single
+      arr.splice(1, 1)
+    }
+  }
+
+  if (arr.length > 1) {
+    if (arr[arr.length - 1].match(only_one_or_more_c)) {
+      arr[arr.length - 2] = arr[arr.length - 2] + arr[arr.length - 1]
+      arr.splice(arr.length - 1, 1)
+    }
+  }
+
+  return arr
+}
+export default postprocess

plugins/speech/src/compute/syllables/syllables.js

@@ -0,0 +1,90 @@
+//chop a string into pronounced syllables
+import postProcess from './postProcess.js'
+
+const all_spaces = / +/g
+const ends_with_vowel = /[aeiouy]$/
+const starts_with_consonant_vowel = /^[^aeiouy]h?[aeiouy]/
+const starts_with_e_then_specials = /^e[sm]/
+const starts_with_e = /^e/
+const ends_with_noisy_vowel_combos = /(eo|eu|ia|oa|ua|ui)$/i
+const aiouy = /[aiouy]/
+const ends_with_ee = /ee$/
+// const whitespace_dash = /\s\-/
+
+//method is nested because it's called recursively
+const doWord = function (w) {
+  const all = []
+  const chars = w.split('')
+  let before = ''
+  let after = ''
+  let current = ''
+  for (let i = 0; i < chars.length; i++) {
+    before = chars.slice(0, i).join('')
+    current = chars[i]
+    after = chars.slice(i + 1, chars.length).join('')
+    let candidate = before + chars[i]
+
+    //it's a consonant that comes after a vowel
+    if (before.match(ends_with_vowel) && !current.match(ends_with_vowel)) {
+      if (after.match(starts_with_e_then_specials)) {
+        candidate += 'e'
+        after = after.replace(starts_with_e, '')
+      }
+      all.push(candidate)
+      return all.concat(doWord(after))
+    }
+
+    //unblended vowels ('noisy' vowel combinations)
+    if (candidate.match(ends_with_noisy_vowel_combos)) {
+      //'io' is noisy, not in 'ion'
+      all.push(before)
+      all.push(current)
+      return all.concat(doWord(after)) //recursion
+    }
+
+    // if candidate is followed by a CV, assume consecutive open syllables
+    if (candidate.match(ends_with_vowel) && after.match(starts_with_consonant_vowel)) {
+      all.push(candidate)
+      return all.concat(doWord(after))
+    }
+  }
+  //if still running, end last syllable
+  if (w.match(aiouy) || w.match(ends_with_ee)) {
+    //allow silent trailing e
+    all.push(w)
+  } else if (w) {
+    let last = all.length - 1
+    if (last < 0) {
+      last = 0
+    }
+    all[last] = (all[last] || '') + w //append it to the last one
+  }
+  return all
+}
+
+const syllables = function (str) {
+  let all = []
+  if (!str) {
+    return all
+  }
+  str = str.replace(/[.,?]/g, '')
+  str.split(all_spaces).map(s => {
+    all = all.concat(doWord(s))
+  })
+
+  // str.split(whitespace_dash).forEach(doWord)
+  all = postProcess(all)
+
+  //for words like 'tree' and 'free'
+  if (all.length === 0) {
+    all = [str]
+  }
+  //filter blanks
+  all = all.filter(s => s)
+
+  return all
+}
+
+// console.log(syllables('civilised'))
+
+export default syllables

plugins/speech/src/plugin.js

@@ -0,0 +1,7 @@
+import compute from './compute/index.js'
+import api from './api.js'
+
+export default {
+  api,
+  compute
+}

plugins/speech/tests/_lib.js

@@ -0,0 +1,15 @@
+import build from '../../../builds/one/compromise-one.mjs'
+import src from '../../../src/one.js'
+import plgBuild from '../builds/compromise-speech.mjs'
+import plg from '../src/plugin.js'
+let nlp;
+
+if (process.env.TESTENV === 'prod') {
+  console.warn('== production build test 🚀 ==')  // eslint-disable-line
+  nlp = build
+  nlp.plugin(plgBuild)
+} else {
+  nlp = src
+  nlp.plugin(plg)
+}
+export default nlp

plugins/speech/tests/soundsLike.test.js

@@ -0,0 +1,21 @@
+import test from 'tape'
+import nlp from './_lib.js'
+
+test('soundsLike-tests', function (t) {
+  const arr = [
+    ['phil collins', 'fil kolins'],
+    ['Philadelphia freedom', 'filatelfia fretom'],
+    ['Shine on me', 'shine on me'],
+    ['peace of mind', 'pease of mint'],
+    ['Yes I do', 'yes i to'],
+    ['Zapped me', 'sapet me'],
+    ['Right between the eyes', 'rit betwen the eyes'],
+    // ['But the times have changed', 'but the times hafe kshanjet'],
+  ]
+  arr.forEach((a) => {
+    const doc = nlp(a[0])
+    const str = doc.soundsLike({})[0].join(' ')
+    t.equal(str, a[1], a[0])
+  })
+  t.end()
+})

plugins/speech/tests/syllables.test.js

@@ -0,0 +1,73 @@
+import test from 'tape'
+import nlp from './_lib.js'
+
+const words = [
+  'sud den ly',
+  'con sti pa tion',
+  'di a bo lic',
+  'fa ted',
+  'ge ne tic',
+  'hy giene',
+  'o ma ha',
+  'i mi ta ted',
+  'tree',
+  'ci vi lised',
+  'went',
+  'to ge ther',
+  'tog gle',
+  'move ment',
+  'mo ment',
+  'do nate',
+  'vi king',
+  'wat tage',
+  'con gre gate',
+  'some thing',
+  'sales man',
+  're sour ces',
+  'eve ry thing',
+  'eve ry bo dy',
+  'hole',
+  'ho ly',
+  'sec ret',
+  'cause',
+  'fate',
+  'fates',
+  'eu lo gy',
+  'e on',
+  'rea sons',
+  'and',
+  'hoc key',
+  'wild',
+  'fun',
+  'fun ny',
+  'hor ses',
+  'cal ga ry',
+  'tor na do',
+  'pen',
+  'stu dy',
+  'ja mai ca',
+  'tri ni dad',
+  'ca na da',
+  'i ran',
+  'i raq',
+  'sau di a ra bi a',
+  'clin ton',
+  'hub ris',
+  'tur key',
+  'cor si ca',
+  'dev iled',
+  // 'horse',
+  // 'chance',
+  // 'peo ple'
+]
+
+test('test length', function (t) {
+  words.forEach(sep => {
+    const str = sep.replace(/ /g, '')
+    const doc = nlp(str)
+    const have = doc.syllables()[0]
+    const want = sep.split(/ /g)
+    t.equal(have.length, want.length, str)
+  })
+  t.end()
+})