Merge pull request #984 from spencermountain/dev

Dev

changelog.md

@@ -12,11 +12,23 @@ While all _Major_ releases should be reviewed, our only _large_ releases are **v
 - deprecate #Participle for #PastParticiple
 - run chunker only when necessary
 - cleanup verb+noun metadata in json()
+- remove confidence numbers
 -->
 
-<!-- #### 14.7.2 [Unreleased]
+<!-- #### 14.7.3 [Unreleased]
 -->
 
+#### 14.8.0 
+- **[fix]** - tagging fixes
+- **[new]** - add Person .presumedMale(), .presumedFemale() methods
+- **[new]** - add Pronoun class, .refersTo()
+- **[new]** - add Noun.references()
+- **[new]** - .nouns('spencer') shorthand as an if-match
+- **[change]** - "[do] you .." etc now #QuestionWord
+- **[new]** - add #Hyphenated tag
+- **[fix]** - improved Auxiliary verb tagging
+- **[update]** - dependencies
+
 #### 14.7.1 
 - **[fix]** - concat fix
 - **[change]** - tagging fixes

data/lexicon/adjectives/adjectives.js

@@ -241,7 +241,6 @@ export default [
   'unappealing',
   'unbecoming',
   'unassuming',
-  'cunning',
   'foregoing',
   'ongoing',
   'easygoing',
@@ -265,15 +264,14 @@ export default [
   'stylish',
   'squeamish',
   'loutish',
-  'lush',
+
 
   'mammoth',
   'north',
   'south',
   'punjabi',
   'mini',
   'a priori',
-  'back',
   'ok',
   'brisk',
 
@@ -305,8 +303,6 @@ export default [
   'royal',
 
   // -el
-  'rebel',
-  'parallel',
   'civil',
   'all',
   'overall',
@@ -338,8 +334,6 @@ export default [
 
   'dim',
   'longterm',
-  'premium',
-  'humdrum',
   'ad infinitum',
   'urban',
   'suburban',
@@ -357,7 +351,6 @@ export default [
   'proven',
   'frozen',
 
-  'sovereign',
   'foreign',
   'benign',
   'main',
@@ -372,7 +365,6 @@ export default [
   'unknown',
   'gung ho',
   'macho',
-  'pro',
   'in vitro',
   'de facto',
   'ipso facto',
@@ -412,7 +404,6 @@ export default [
   'sui generis',
   'gratis',
   'for keeps',
-  'excess',
   'bogus',
   'dangerous',
   'offbeat',
@@ -531,7 +522,6 @@ export default [
   'past',
   'best',
   'honest',
-  'racist',
   'populist',
   'against',
   'most',
@@ -665,5 +655,9 @@ export default [
   'golden',
   'bare',
   'subpar',
+  'ingrown',
+  'agreeable',
+  'vile',
+  'win win',
 ]
 

data/lexicon/index.js

@@ -5,7 +5,7 @@ import lex from './misc.js'
 import demonyms from './nouns/demonyms.js'
 import organizations from './nouns/organizations.js'
 import possessives from './nouns/possessives.js'
-import professions from './nouns/professions.js'
+import actors from './nouns/actors.js'
 import pronouns from './nouns/pronouns.js'
 import singulars from './nouns/singulars.js'
 import sportsTeams from './nouns/sportsTeams.js'
@@ -60,7 +60,6 @@ import personDate from './switches/person-date.js'
 import personVerb from './switches/person-verb.js'
 import personPlace from './switches/person-place.js'
 import unitNoun from './switches/unit-noun.js'
-import honorificNoun from './switches/honorific-noun.js'
 
 
 //add-in the generic, flat word-lists
@@ -69,7 +68,7 @@ const data = [
   [demonyms, 'Demonym'],
   [organizations, 'Organization'],
   [possessives, 'Possessive'],
-  [professions, 'Actor'],
+  [actors, 'Actor'],
   [pronouns, 'Pronoun'],
   [singulars, 'Singular'],
   [sportsTeams, 'SportsTeam'],
@@ -92,10 +91,10 @@ const data = [
   [dates, 'Date'],
   [durations, 'Duration'],
 
-  [femaleNames, 'FemaleName'],
+  [femaleNames, 'FemaleName'], //5kb
   [firstnames, 'FirstName'],
-  [lastnames, 'LastName'],
-  [maleNames, 'MaleName'],
+  [lastnames, 'LastName'], //3k
+  [maleNames, 'MaleName'], //5kb
   [people, 'Person'],
 
   [adjectives, 'Adjective'],
@@ -111,9 +110,9 @@ const data = [
   [modals, 'Modal'],
   [verbs, 'Verb'],
   [participles, 'Participle'],
-  [phrasals, 'PhrasalVerb'], //5kb
+  [phrasals, 'PhrasalVerb'], //3kb
 
-  // switches
+  // switches - 10kb
   [adjGerund, 'Adj|Gerund'],
   [adjNoun, 'Adj|Noun'],
   [adjPast, 'Adj|Past'],
@@ -125,7 +124,6 @@ const data = [
   [personDate, 'Person|Date'],
   [personVerb, 'Person|Verb'],
   [unitNoun, 'Unit|Noun'],
-  [honorificNoun, 'Honorific|Noun'],
 ]
 for (let i = 0; i < data.length; i++) {
   const list = data[i][0]

data/lexicon/misc/adverbs.js

@@ -86,5 +86,6 @@ export default [
   'very',
   'well',
   'backward',
+  'hella',
 
 ]

data/lexicon/misc/conjunctions.js

@@ -40,9 +40,8 @@ export default [
   'or not',
   'otherwise',
   'plus',
-  'provided',
+  // 'provided',
   'provided that',
-  'provide that',
   'since',
   'so that',
   'supposing',

data/lexicon/misc/expressions.js

@@ -107,7 +107,7 @@ export default [
   'wtf',
   'ya',
   'yaa',
-  'yahoo',
+  // 'yahoo',
   'yay',
   'yeah',
   'yep',
@@ -121,5 +121,7 @@ export default [
   `d'oh`,
   'shoo',
   'there now',
-  'aye'
+  'aye',
+  'cmon',//come on
+  'excuse me',
 ]

data/lexicon/nouns/actors.js

@@ -0,0 +1,166 @@
+//professions 'lawyer' that aren't covered by verb.to_actor()
+
+export default [
+  'accountant',
+  'administrator',
+  'advisor',
+  'agent',
+  'aid',
+  'analyst',
+  'animator',
+  'announcer',
+  'architect',
+  'artist',
+  'assistant',
+  'attendant',
+  'bricklayer',
+  'broker',
+  'butcher',
+  'captain',
+  'caretaker',
+  'carpenter',
+  'ceo',
+  'cfo',
+  'cleaner',
+  'clerk',
+  'composer',
+  'cook',
+  'coordinator',
+  'cowboy',
+  'dancer',
+  'dealer',
+  'deputee',
+  'deputy',
+  'designer',
+  'detective',
+  'developer',
+  'dietician',
+  'director',
+  'doctor',
+  'driver',
+  'examiner',
+  'farmer',
+  'fighter',
+  'firefighter',
+  'fireman',
+  'fitter',
+  'gardener',
+  'guard',
+  'hairdresser',
+  'housekeeper',
+  'illustrator',
+  'inspector',
+  'installer',
+  'instructor',
+  'investigator',
+  'journalist',
+  'laborer',
+  'lawyer',
+  'leader',
+  'maker',
+  'mechanic',
+  'merchant',
+  'minister',
+  'musician',
+  'nurse',
+  'officer',
+  'operator',
+  'optometrist',
+  'organizer',
+  'photographer',
+  'physician',
+  'pilot',
+  'planner',
+  'player',
+  'plumber',
+  'policeman',
+  'politician',
+  'practitioner',
+  'president',
+  'producer',
+  'programmer',
+  'prospector',
+  'psychologist',
+  'receptionist',
+  'recruiter',
+  'reporter',
+  'representative',
+  'researcher',
+  'roofer',
+  'sailor',
+  'scientist',
+  'secretary',
+  'singer',
+  'soldier',
+  'specialist',
+  'spokesperson',
+  'supervisor',
+  'surgeon',
+  'technician',
+  'therapist',
+  'transcriptionist',
+  'underwriter',
+  'veterinarian',
+  'worker',
+  'writer',
+  'manager',
+  // 'engineer',
+
+
+  'admiral',
+  'ayatullah',
+  'baron',
+  'baroness',
+  'brigadier',
+  'chancellor',
+  'colonel',
+  'commander',
+  'congressman',
+  'congresswoman',
+  'constable',
+  'corporal',
+  'councillor',
+  'doctor',
+  'dutchess',
+  'excellency',
+  'field marshal',
+  'lieutenant',
+  // 'count',
+  'captain',
+  'judge',
+  'king',
+  'lady',
+  'lord',
+  'magistrate',
+  'marshal',
+  'mayor',
+  // 'miss',
+  'missus',
+  'mister',
+  'officer',
+  'pastor',
+  'president',
+  'prime minister',
+  'prince',
+  'princess',
+  'professor',
+  'queen',
+  'rabbi',
+  'rear admiral',
+  'first lady',
+  'reverend',
+  'sergeant',
+  'secretary',
+  'cheerleader',
+  'quarterback',
+  'running back',
+  'defenceman',
+  'goalie',
+  'goaltender',
+  'goalkeeper',
+  'sultan',
+  'taoiseach',
+  'vice admiral',
+  'sergeant major',
+  'lieutenant general',
+]

data/lexicon/nouns/possessives.js

@@ -1,6 +1,4 @@
 export default [
-  'anyone',
-  'anything',
   // 'her', //this one is check ambiguous
   // 'hers',
   // 'his',
@@ -10,11 +8,10 @@ export default [
   'none',
   'our',
   'ours',
-  'something',
-  'their',
-  'theirs',
-  'your',
-  'yours',
+  // 'their',
+  // 'theirs',
+  // 'your',
+  // 'yours',
 
 
 ]