gnu/packages/language.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2015, 2016 Eric Bavier <bavier@member.fsf.org>
   3 ;;;
   4 ;;; This file is part of GNU Guix.
   5 ;;;
   6 ;;; GNU Guix is free software; you can redistribute it and/or modify it
   7 ;;; under the terms of the GNU General Public License as published by
   8 ;;; the Free Software Foundation; either version 3 of the License, or (at
   9 ;;; your option) any later version.
  10 ;;;
  11 ;;; GNU Guix is distributed in the hope that it will be useful, but
  12 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  13 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 ;;; GNU General Public License for more details.
  15 ;;;
  16 ;;; You should have received a copy of the GNU General Public License
  17 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  18
  19 (define-module (gnu packages language)
  20   #:use-module (gnu packages)
  21   #:use-module (gnu packages perl)
  22   #:use-module (gnu packages web)
  23   #:use-module (guix packages)
  24   #:use-module (guix build-system perl)
  25   #:use-module ((guix licenses) #:select (gpl2 gpl3))
  26   #:use-module (guix download))
  27
  28 (define-public perl-lingua-en-findnumber
  29   (package
  30     (name "perl-lingua-en-findnumber")
  31     (version "1.30")
  32     (source
  33      (origin
  34        (method url-fetch)
  35        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  36                            "Lingua-EN-FindNumber-" version ".tar.gz"))
  37        (sha256
  38         (base32
  39          "0g5h6bwwggizwh3dd3xyx92537s4qy8nspshp1lrm9qlxh3prc28"))))
  40     (build-system perl-build-system)
  41     (propagated-inputs
  42      `(("perl-lingua-en-words2nums" ,perl-lingua-en-words2nums)))
  43     (home-page "http://search.cpan.org/dist/Lingua-EN-FindNumber")
  44     (synopsis "Locate (written) numbers in English text ")
  45     (description "This module provides a regular expression for finding
  46 numbers in English text.  It also provides functions for extracting and
  47 manipulating such numbers.")
  48     (license (package-license perl))))
  49
  50 (define-public perl-lingua-en-inflect
  51   (package
  52     (name "perl-lingua-en-inflect")
  53     (version "1.895")
  54     (source
  55      (origin
  56        (method url-fetch)
  57        (uri (string-append "mirror://cpan/authors/id/D/DC/DCONWAY/"
  58                            "Lingua-EN-Inflect-" version ".tar.gz"))
  59        (sha256
  60         (base32
  61          "0drzg9a2dkjxgf00n6jg0jzhd8972bh3j4wdnmdxpqi3zmfqhwcy"))))
  62     (build-system perl-build-system)
  63     (native-inputs `(("perl-module-build" ,perl-module-build)))
  64     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect")
  65     (synopsis "Convert singular to plural")
  66     (description "Lingua::EN::Inflect provides plural inflections,
  67 \"a\"/\"an\" selection for English words, and manipulation of numbers as
  68 words.  Plural forms of all nouns, most verbs, and some adjectives are
  69 provided.  Where appropriate, \"classical\" variants (for example: \"brother\"
  70 -> \"brethren\", \"dogma\" -> \"dogmata\", etc.) are also provided.")
  71     (license (package-license perl))))
  72
  73 (define-public perl-lingua-en-inflect-number
  74   (package
  75     (name "perl-lingua-en-inflect-number")
  76     (version "1.11")
  77     (source
  78      (origin
  79        (method url-fetch)
  80        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  81                            "Lingua-EN-Inflect-Number-" version ".tar.gz"))
  82        (sha256
  83         (base32
  84          "0rvgrff96ja7fqr79dszmiyv2wz4izw82znah0mx9szkir657gfz"))))
  85     (build-system perl-build-system)
  86     (propagated-inputs
  87      `(("perl-lingua-en-inflect" ,perl-lingua-en-inflect)))
  88     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect-Number")
  89     (synopsis "Force number of words to singular or plural")
  90     (description "This module extends the functionality of Lingua::EN::Inflect
  91 with three new functions for determining plurality of a word and forcefully
  92 converting a word to singular or plural.")
  93     (license (package-license perl))))
  94
  95 (define-public perl-lingua-en-inflect-phrase
  96   (package
  97     (name "perl-lingua-en-inflect-phrase")
  98     (version "0.18")
  99     (source
 100      (origin
 101        (method url-fetch)
 102        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 103                            "Lingua-EN-Inflect-Phrase-" version ".tar.gz"))
 104        (sha256
 105         (base32
 106          "19prg6pkgnznrc82ymdkdi3l9schg815axjig7adca5yqa7mn2i9"))))
 107     (build-system perl-build-system)
 108     (native-inputs
 109      `(("perl-test-nowarnings" ,perl-test-nowarnings)))
 110     (propagated-inputs
 111      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)
 112        ("perl-lingua-en-inflect" ,perl-lingua-en-inflect)
 113        ("perl-lingua-en-inflect-number" ,perl-lingua-en-inflect-number)
 114        ("perl-lingua-en-number-isordinal" ,perl-lingua-en-number-isordinal)
 115        ("perl-lingua-en-tagger" ,perl-lingua-en-tagger)))
 116     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect-Phrase")
 117     (synopsis "Inflect short English phrases")
 118     (description "This module attempts to pluralize or singularize short
 119 English phrases.")
 120     (license (package-license perl))))
 121
 122 (define-public perl-lingua-en-number-isordinal
 123   (package
 124     (name "perl-lingua-en-number-isordinal")
 125     (version "0.04")
 126     (source
 127      (origin
 128        (method url-fetch)
 129        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 130                            "Lingua-EN-Number-IsOrdinal-" version ".tar.gz"))
 131        (sha256
 132         (base32
 133          "1321fm4pyqcamffd0qsjm1gb07ijqndlb29qkcqf22kaibngfm4i"))))
 134     (build-system perl-build-system)
 135     (native-inputs
 136      `(("perl-try-tiny" ,perl-try-tiny)
 137        ("perl-test-fatal" ,perl-test-fatal)))
 138     (propagated-inputs
 139      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)))
 140     (home-page "http://search.cpan.org/dist/Lingua-EN-Number-IsOrdinal")
 141     (synopsis "Detect if English number is ordinal or cardinal")
 142     (description "This module will tell you if a number, either in words or as
 143 digits, is a cardinal or ordinal number.")
 144     (license (package-license perl))))
 145
 146 (define-public perl-lingua-en-tagger
 147   (package
 148     (name "perl-lingua-en-tagger")
 149     (version "0.24")
 150     (source
 151      (origin
 152        (method url-fetch)
 153        (uri (string-append "mirror://cpan/authors/id/A/AC/ACOBURN/"
 154                            "Lingua-EN-Tagger-" version ".tar.gz"))
 155        (sha256
 156         (base32
 157          "0qksqh1zi8fz76a29s2ll4g6yr8y6agmzgq7ngccvgj3gza5q241"))))
 158     (build-system perl-build-system)
 159     (propagated-inputs
 160      `(("perl-memoize-expirelru" ,perl-memoize-expirelru)
 161        ("perl-lingua-stem" ,perl-lingua-stem)
 162        ("perl-html-parser" ,perl-html-parser)
 163        ("perl-html-tagset" ,perl-html-tagset)))
 164     (home-page "http://search.cpan.org/dist/Lingua-EN-Tagger")
 165     (synopsis "Part-of-speech tagger for English natural language processing")
 166     (description "This module is a probability based, corpus-trained tagger
 167 that assigns part-of-speech tags to English text based on a lookup dictionary
 168 and a set of probability values.  The tagger assigns appropriate tags based on
 169 conditional probabilities - it examines the preceding tag to determine the
 170 appropriate tag for the current word.  Unknown words are classified according
 171 to word morphology or can be set to be treated as nouns or other parts of
 172 speech.  The tagger also extracts as many nouns and noun phrases as it can,
 173 using a set of regular expressions.")
 174     (license gpl3)))
 175
 176 (define-public perl-lingua-en-words2nums
 177   (package
 178     (name "perl-lingua-en-words2nums")
 179     (version "0.18")
 180     (source
 181      (origin
 182        (method url-fetch)
 183        (uri (string-append "mirror://cpan/authors/id/J/JO/JOEY/"
 184                            "Lingua-EN-Words2Nums-" version ".tar.gz"))
 185        (sha256
 186         (base32
 187          "118xx8qr1zbx30psv7ic55w65h15mc1vz6zicshfm96jgiwmcrb8"))))
 188     (build-system perl-build-system)
 189     (home-page "http://search.cpan.org/dist/Lingua-EN-Words2Nums")
 190     (synopsis "Convert English text to numbers")
 191     (description "This module converts English text into numbers.  It supports
 192 both ordinal and cardinal numbers, negative numbers, and very large numbers.")
 193     (license (package-license perl))))
 194
 195 (define-public perl-lingua-pt-stemmer
 196   (package
 197     (name "perl-lingua-pt-stemmer")
 198     (version "0.01")
 199     (source
 200      (origin
 201        (method url-fetch)
 202        (uri (string-append "mirror://cpan/authors/id/X/XE/XERN/"
 203                            "Lingua-PT-Stemmer-" version ".tar.gz"))
 204        (sha256
 205         (base32
 206          "11rqc5pqnkl9c13vy7sihiyas14ci0pj3k6chrgrgjv5sjv2m4a5"))))
 207     (build-system perl-build-system)
 208     (home-page "http://search.cpan.org/dist/Lingua-PT-Stemmer")
 209     (synopsis "Portuguese language stemming")
 210     (description "This module implements a Portuguese stemming algorithm
 211 proposed in the paper A Stemming Algorithm for the Portuguese Language by
 212 Moreira, V. and Huyck, C.")
 213     (license (package-license perl))))
 214
 215 (define-public perl-lingua-stem
 216   (package
 217     (name "perl-lingua-stem")
 218     (version "0.84")
 219     (source
 220      (origin
 221        (method url-fetch)
 222        (uri (string-append "mirror://cpan/authors/id/S/SN/SNOWHARE/"
 223                            "Lingua-Stem-" version ".tar.gz"))
 224        (sha256
 225         (base32
 226          "12avh2mnnc7llmmshrr5bgb473fvydxnlqrqbl2815mf2dp4pxcg"))))
 227     (build-system perl-build-system)
 228     (native-inputs
 229      `(("perl-module-build" ,perl-module-build)))
 230     (propagated-inputs
 231      `(("perl-lingua-pt-stemmer" ,perl-lingua-pt-stemmer)
 232        ("perl-lingua-stem-fr" ,perl-lingua-stem-fr)
 233        ("perl-lingua-stem-it" ,perl-lingua-stem-it)
 234        ("perl-lingua-stem-ru" ,perl-lingua-stem-ru)
 235        ("perl-lingua-stem-snowball-da" ,perl-lingua-stem-snowball-da)
 236        ("perl-snowball-norwegian" ,perl-snowball-norwegian)
 237        ("perl-snowball-swedish" ,perl-snowball-swedish)
 238        ("perl-text-german" ,perl-text-german)))
 239     (home-page "http://search.cpan.org/dist/Lingua-Stem")
 240     (synopsis "Stemming of words in various languages")
 241     (description "This routine applies stemming algorithms to its parameters,
 242 returning the stemmed words as appropriate to the selected locale.")
 243     (license (package-license perl))))
 244
 245 (define-public perl-lingua-stem-fr
 246   (package
 247     (name "perl-lingua-stem-fr")
 248     (version "0.02")
 249     (source
 250      (origin
 251        (method url-fetch)
 252        (uri (string-append "mirror://cpan/authors/id/S/SD/SDP/"
 253                            "Lingua-Stem-Fr-" version ".tar.gz"))
 254        (sha256
 255         (base32
 256          "0vyrspwzaqjxm5mqshf4wvwa3938mkajd1918d9ii2l9m2rn8kwx"))))
 257     (build-system perl-build-system)
 258     (home-page "http://search.cpan.org/dist/Lingua-Stem-Fr")
 259     (synopsis "Porter's stemming algorithm for French")
 260     (description "This module uses a modified version of the Porter Stemming
 261 Algorithm to return a stemmed French word.")
 262     (license (package-license perl))))
 263
 264 (define-public perl-lingua-stem-it
 265   (package
 266     (name "perl-lingua-stem-it")
 267     (version "0.02")
 268     (source
 269      (origin
 270        (method url-fetch)
 271        (uri (string-append "mirror://cpan/authors/id/A/AC/ACALPINI/"
 272                            "Lingua-Stem-It-" version ".tar.gz"))
 273        (sha256
 274         (base32
 275          "1207r183s5hlh4mfwa6p46vzm0dhvrs2dnss5s41a0gyfkxp7riq"))))
 276     (build-system perl-build-system)
 277     (home-page "http://search.cpan.org/dist/Lingua-Stem-It")
 278     (synopsis "Porter's stemming algorithm for Italian")
 279     (description "This module applies the Porter Stemming Algorithm to its
 280 parameters, returning the stemmed Italian word.")
 281     (license (package-license perl))))
 282
 283 (define-public perl-lingua-stem-ru
 284   (package
 285     (name "perl-lingua-stem-ru")
 286     (version "0.01")
 287     (source
 288      (origin
 289        (method url-fetch)
 290        (uri (string-append "mirror://cpan/authors/id/A/AL/ALGDR/"
 291                            "Lingua-Stem-Ru-" version ".tar.gz"))
 292        (sha256
 293         (base32
 294          "0pqgg442fkf12ayh9fgmpa8q9x0iqx6s96181r52yn7s7pcs61h6"))))
 295     (build-system perl-build-system)
 296     (home-page "http://search.cpan.org/dist/Lingua-Stem-Ru")
 297     (synopsis "Porter's stemming algorithm for Russian")
 298     (description "This module applies the Porter Stemming Algorithm to its
 299 parameters, returning the stemmed Russian (KOI8-R only) word.")
 300     (license (package-license perl))))
 301
 302 (define-public perl-lingua-stem-snowball-da
 303   (package
 304     (name "perl-lingua-stem-snowball-da")
 305     (version "1.01")
 306     (source
 307      (origin
 308        (method url-fetch)
 309        (uri (string-append "mirror://cpan/authors/id/C/CI/CINE/"
 310                            "Lingua-Stem-Snowball-Da-" version ".tar.gz"))
 311        (sha256
 312         (base32
 313          "0mm0m7glm1s6i9f6a78jslw6wh573208arxhq93yriqmw17bwf9f"))))
 314     (build-system perl-build-system)
 315     (home-page "http://search.cpan.org/dist/Lingua-Stem-Snowball-Da")
 316     (synopsis "Porters stemming algorithm for Danish")
 317     (description "Lingua::Stem::Snowball::Da is a perl port of the danish
 318 stemmer at http://snowball.sourceforge.net, it was originally altered from the
 319 Lingua::Stem::Snowball::Se.")
 320     (license gpl2)))
 321
 322 (define-public perl-snowball-norwegian
 323   (package
 324     (name "perl-snowball-norwegian")
 325     (version "1.2")
 326     (source
 327      (origin
 328        (method url-fetch)
 329        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 330                            "Snowball-Norwegian-" version ".tar.gz"))
 331        (sha256
 332         (base32
 333          "0675v45bbsh7vr7kpf36xs2q79g02iq1kmfw22h20xdk4rzqvkqx"))))
 334     (build-system perl-build-system)
 335     (native-inputs `(("perl-module-build" ,perl-module-build)))
 336     (home-page "http://search.cpan.org/dist/Snowball-Norwegian")
 337     (synopsis "Porters stemming algorithm for Norwegian")
 338     (description "Lingua::Stem::Snowball::No is a perl port of the norwegian
 339 stemmer at http://snowball.tartarus.org.")
 340     (license (package-license perl))))
 341
 342 (define-public perl-snowball-swedish
 343   (package
 344     (name "perl-snowball-swedish")
 345     (version "1.2")
 346     (source
 347      (origin
 348        (method url-fetch)
 349        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 350                            "Snowball-Swedish-" version ".tar.gz"))
 351        (sha256
 352         (base32
 353          "0agwc12jk5kmabnpsplw3wf4ii5w1zb159cpin44x3srb0sr5apg"))))
 354     (build-system perl-build-system)
 355     (native-inputs `(("perl-module-build" ,perl-module-build)))
 356     (home-page "http://search.cpan.org/dist/Snowball-Swedish")
 357     (synopsis "Porters stemming algorithm for Swedish")
 358     (description "Lingua::Stem::Snowball::Se is a perl port of the swedish
 359 stemmer at http://snowball.sourceforge.net.")
 360     (license (package-license perl))))
 361
 362 (define-public perl-string-toidentifier-en
 363   (package
 364     (name "perl-string-toidentifier-en")
 365     (version "0.11")
 366     (source
 367      (origin
 368        (method url-fetch)
 369        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 370                            "String-ToIdentifier-EN-" version ".tar.gz"))
 371        (sha256
 372         (base32
 373          "1bawghkgkkx7j3avnrj5sg3vix1z5564ks6wf9az3jc2knh8s5nh"))))
 374     (build-system perl-build-system)
 375     (propagated-inputs
 376      `(("perl-lingua-en-inflect-phrase" ,perl-lingua-en-inflect-phrase)
 377        ("perl-text-unidecode" ,perl-text-unidecode)
 378        ("perl-namespace-clean" ,perl-namespace-clean)))
 379     (home-page "http://search.cpan.org/dist/String-ToIdentifier-EN")
 380     (synopsis "Convert strings to English program identifiers")
 381     (description "This module provides a utility method, \"to_identifier\" for
 382 converting an arbitrary string into a readable representation using the ASCII
 383 subset of \"\\w\" for use as an identifier in a computer program.  The intent
 384 is to make unique identifier names from which the content of the original
 385 string can be easily inferred by a human just by reading the identifier.")
 386     (license (package-license perl))))
 387
 388 (define-public perl-text-german
 389   (package
 390     (name "perl-text-german")
 391     (version "0.06")
 392     (source
 393      (origin
 394        (method url-fetch)
 395        (uri (string-append "mirror://cpan/authors/id/U/UL/ULPFR/"
 396                            "Text-German-" version ".tar.gz"))
 397        (sha256
 398         (base32
 399          "1p87pgap99lw0nv62i3ghvsi7yg90lhn8vsa3yqp75rd04clybcj"))))
 400     (build-system perl-build-system)
 401     (home-page "http://search.cpan.org/dist/Text-German")
 402     (synopsis "German grundform reduction")
 403     (description "This module is a rather incomplete implementation of work
 404 done by Gudrun Putze-Meier.")
 405     (license (package-license perl))))