loving build of wordnet in JSON.

no memory pointers, no python, no DSL, no guff. no crazy-framework stuff at all.

the data is zipped for github, but it automatically unzips when you first use it.

npm install wordnetjs

if you just want the JSON, unzip ./data.zip then you can just do your random shit. it's 6mb -> 32mb

it's the cutest way to use wordnet by a pretty wide margin.

API

wn= require("wordnetjs")
//generic lookup
wn.lookup('warrant')
//6 results
 
//pos-specific lookups
wn.verb('warrant') // (1 result)
wn.adjective('cheeky')
wn.adverb('slightly')
wn.noun('grape')
 
//sugar
wn.synonyms("perverse")
// [{id:"depraved.adjective.01"...}]
wn.antonyms("perverse")
// [{id:"docile.adjective.01"}]
wn.pos("swim")
//[ 'Verb', 'Noun' ]
 
//unique, alphabetical list of all words
wn.words((arr)=>{
  console.log(arr.filter((w)=> w.match(/cool/))
})
//[ 'air-cool', 'air-cooled', 'cool',  'cool down', ...
 

Liberties taken

Symmetric-relations

if the holonym of 'sausage' is 'sausage meat', the reverse (called a 'meronym') is almost always true.

As the beautiful George Miller explains, the meronym-holonym relations, and the hypernym-hyponym relations are symmetric with few exceptions. Ignoring these exceptions reduces the filesize by half, so I did it.

To go from 'sausage meat' to 'sausage', just query the opposite direction.

Antonyms on synsets

Adjective synsets in wordnet have no antonyms, but rather each individual word-sense has an antonym. This makes wordnet's antonym data really specific, but for most purposes, that's probably overdoing it. delete.

Kill 'Coordinate-terms'

Given wordnet is a graph, this is just redundant data. delete.

Reformatted glosses

Use only 1 gloss (description) per synset, and split it by semicolon-seperators.

Same-As links

Most Nouns include freebase ids and wikipedia titles. There were reconciled in a mostly-manual process by freebase in 2010.

Contents

117,657 synsets in total

82,113 Noun Synsets

{
  id: "candy cane.noun.01",
  lexname: "noun.food",
  syntactic_category: "Noun",
  description: "a hard candy in the shape of a rod (usually with stripes)",
  words: ["candy cane"],
  relationships: {
    type_of: ["candy.noun.01"],
    made_with: [],
    members: [],
    parts: [],
    instances: []
  },
  same_as: {
    freebase_topic: "/m/01hrm7",
    wikipedia_page: "Candy_cane"
  }
}

13,767 Verb Synsets

{
  id: "lean back.verb.01",
  lexname: "verb.motion",
  syntactic_category: "Verb",
  description: "move the upper body backwards and down",
  words: ["lean back", "recline"],
  assumes: [],
  causes: []
}

hypernym: the verb Y is a hypernym of the verb X if the activity X is a (kind of) Y (to perceive is an hypernym of to listen) troponym: the verb Y is a troponym of the verb X if the activity Y is doing X in some manner (to lisp is a troponym of to talk) entailment: the verb Y is entailed by X if by doing X you must be doing Y (to sleep is entailed by to snore)

18,156 Adjective Synsets

{
  id: "phantasmagoric.adjective.01",
  lexname: "adj.all",
  syntactic_category: "Adjective",
  description: "characterized by fantastic imagery and incongruous juxtapositions",
  words: ["phantasmagoric", "surreal", "phantasmagorical", "surrealistic"],
  similar: ["unrealistic.adjective.01"]
},

related nouns similar to participle of verb

3,621 Adverb Synsets

{
  id: "refreshingly.adverb.01",
  lexname: "adv.all",
  syntactic_category: "Adverb",
  description: "in a manner that relieves fatigue and restores vitality",
  words: ["refreshingly", "refreshfully"]
}

roll your own build

to build your own, get a freebase key and put it in ./build/build.js run 'npm install' then 'node ./build.js'