gnu/packages/language.scm

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