Skip to main content

A helper library for accessing and manipulating WordNets in the MultiWordNet

Project description

https://travis-ci.org/wmshort/multiwordnet.svg?branch=master

MultiWordNet

Background

A WordNet is a lexico-conceptual database for a language. In a WordNet, a language’s lexemes (nouns, verbs, adjectives, and adverbs) are grouped into sets of semantically related words called synsets (for “synonym sets”), which thus correspond to the senses that are lexicalized in the language. A WordNet also typically includes information about semantic relations (i.e., relations between synsets) and about lexical relations (i.e., relations between words).

Created by Stefano Minozzi between 2004 and 2008 as part of the Fondazione Bruno Kessler’s MultiWordNet, the original Latin WordNet contained 9,124 lemmas drawn from Riganti’s Lessico Fondamentale Latino. The University of Exeter’s TExtual Project aims to build on and expand Minozzi’s work by adding some 30,000 words, covering the Latin language from the archaic period to late antiquity (and somewhat beyond).

Installation

To get setup, all you need to do is compile the relevant SQLite databases:

from multiwordnet.db import compile compile('latin')

You will need to do the same for the English and Italian synset databases:

compile('english', 'synset') compile('italian', 'synset')

To make full use of the semantic data that is included in the MultiWordNet, you will also want to compile the list of common relations and semfield hierarchy:

compile('common')

Basic usage

from multiwordnet.wordnet import WordNet

LWN = WordNet('latin') for lemma in LWN.lemmas: # all the lemmas currently in the WordNet `` print(lemma.lemma, lemma.pos)`` abalieno = LWN.get_lemma('abalieno', 'v') # this gives you access to a single lemma abalieno.synonyms # all lemmas that share a synset with 'abalieno' abalieno.antonyms abalieno.derivates # use .get_derivates('n') to restrict by POS abalieno.relatives # use .get_relatives('n') to restrict by POS abalieno.synsets LWN.get('abalien', strict=False) # returns a list of lemmas using wildcard matching LWN.get('abalien', pos='v', strict=False) # restrict the results to verbs synset = LWN.get_synset('n#07462736') # you can find a synset directly, if you know its offset ID synset.lemmas

LWN.get_relations(source=synset) # all semantic relations where 'synset' is the source LWN.get_relations(source=synset, type='@') # restrict to hyponymy relations

Relations are of the following types: Noun relations ————– Semantic: ! antonym @ hypernym ~ hyponym #m member-of #s substance-of #p part-of %m has-member %s has-substance %p has-part = attribute | nearest +r has-role -r is-role-of Lexical: ! antonym @ hypernym ~ hyponym +c composed-of -c composes derived-from / related-to

Verb relations

Semantic: ! antonym @ hypernym ~ hyponym * entailment > causes ^ also-see $ verb-group | nearest

Lexical: ! antonym @ hypernym ~ hyponym +c composed-of -c composes derived-from / related-to

Adjective relations

Semantic: ! antonym & similar-to = is-value-of ^ also-see | nearest

Lexical: ! antonym < participle +c composed-of -c composes derived-from / related-to

Adverb relations

Semantic: ! antonym | nearest

Lexical: ! antonym +c composed-of -c composes derived-from / related-to

LWN.get_semfield_by_code('110') # 'Furniture'

Semfields

160 Acoustics 187 Administration 123 Agriculture 142 Anatomy 125 Animal_Husbandry 145 Animals 188 Anthropology 96 Applied_Science 35 Archaeology 91 Archery 107 Architecture 19 Art 192 Artisanship 49 Astrology 113 Astronautics 139 Astronomy 87 Athletics 161 Atomic_Physic 225 Aviation 66 Badminton 204 Banking 67 Baseball 68 Basketball 62 Betting 141 Biochemistry 140 Biology 194 Body_Care 202 Book_Keeping 94 Bowling 89 Boxing 109 Buildings 63 Card 149 Chemistry 61 Chess 20 Cinema 3 Color 195 Commerce 128 Computer_Science 69 Cricket 77 Cycling 21 Dance 118 Dentistry 215 Diplomacy 85 Diving 29 Drawing 150 Earth 196 Economy 167 Electricity 168 Electronics 114 Electrotechnology 111 Engineering 201 Enterprise 146 Entomology 134 Environment 190 Ethnology 206 Exchange 1 Factotum 207 Fashion 90 Fencing 203 Finance 92 Fishing 191 Folklore 105 Food 70 Football 54 Free_Time 110 Furniture 163 Gas 106 Gastronomy 144 Genetics 165 Geography 151 Geology 157 Geometry 71 Golf 37 Grammar 31 Graphic_Arts 193 Health 34 Heraldry 33 History 80 Hockey 127 Home 9 Humanities 93 Hunting 115 Hydraulics 208 Industry 199 Insurance 42 Jewellery 209 Law 36 Linguistics 38 Literature 156 Mathematics 116 Mechanics 117 Medicine 152 Meteorology 7 Metrology 210 Military 205 Money 81 Mountaineering 25 Music 51 Mythology 227 Nautical 2 Number 43 Numismatics 48 Occultism 153 Oceanography 164 Optics 24 Painting 154 Paleontology 47 Paranormal 211 Pedagogy 5 Person 119 Pharmacy 32 Philately 39 Philology 40 Philosophy 26 Photography 159 Physics 143 Physiology 147 Plants 41 Plastic_Arts 58 Play 214 Politics 220 Post 120 Psychiatry 46 Psychoanalysis 8 Psychological_Features 45 Psychology 216 Publishing 129 Pure_Science 6 Quality 95 Racing 64 Radio_TV 121 Radiology 228 Railway 50 Religion 52 Roman_Catholic 82 Rowing 72 Rugby 212 School 44 Sculpture 217 Sexuality 78 Skating 79 Skiing 73 Soccer 169 Social_Science 218 Sociology 65 Sport 158 Statistics 84 Sub 122 Surgery 83 Swimming 74 Table_Tennis 200 Tax 219 Telecommunication 221 Telegraphy 222 Telephony 75 Tennis 28 Theatre 53 Theology 4 Time_Period 166 Topography 223 Tourism 108 Town_Planning 224 Transport 213 University 226 Vehicles 126 Veterinary 76 Volleyball 88 Wrestling

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

multiwordnet-0.1.4.tar.gz (20.1 MB view hashes)

Uploaded Source

Built Distribution

multiwordnet-0.1.4-py2.py3-none-any.whl (20.5 MB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page