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 perl-check)
  23   #:use-module (gnu packages web)
  24   #:use-module (guix packages)
  25   #:use-module (guix build-system perl)
  26   #:use-module ((guix licenses) #:select (gpl2 gpl3 perl-license))
  27   #:use-module (guix download))
  28
  29 (define-public perl-lingua-en-findnumber
  30   (package
  31     (name "perl-lingua-en-findnumber")
  32     (version "1.32")
  33     (source
  34      (origin
  35        (method url-fetch)
  36        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  37                            "Lingua-EN-FindNumber-" version ".tar.gz"))
  38        (sha256
  39         (base32
  40          "015ximzdp42v824llwlg2pd77vd0d172lb4xs55q9f9zhqf6s5qx"))))
  41     (build-system perl-build-system)
  42     (propagated-inputs
  43      `(("perl-lingua-en-words2nums" ,perl-lingua-en-words2nums)))
  44     (home-page "http://search.cpan.org/dist/Lingua-EN-FindNumber")
  45     (synopsis "Locate (written) numbers in English text ")
  46     (description "This module provides a regular expression for finding
  47 numbers in English text.  It also provides functions for extracting and
  48 manipulating such numbers.")
  49     (license perl-license)))
  50
  51 (define-public perl-lingua-en-inflect
  52   (package
  53     (name "perl-lingua-en-inflect")
  54     (version "1.901")
  55     (source
  56      (origin
  57        (method url-fetch)
  58        (uri (string-append "mirror://cpan/authors/id/D/DC/DCONWAY/"
  59                            "Lingua-EN-Inflect-" version ".tar.gz"))
  60        (sha256
  61         (base32
  62          "0mcwlgf6hkh4zm3s1x899f25xj4hyzrc2vssiwfxysqja36yf5ys"))))
  63     (build-system perl-build-system)
  64     (native-inputs `(("perl-module-build" ,perl-module-build)))
  65     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect")
  66     (synopsis "Convert singular to plural")
  67     (description "Lingua::EN::Inflect provides plural inflections,
  68 \"a\"/\"an\" selection for English words, and manipulation of numbers as
  69 words.  Plural forms of all nouns, most verbs, and some adjectives are
  70 provided.  Where appropriate, \"classical\" variants (for example: \"brother\"
  71 -> \"brethren\", \"dogma\" -> \"dogmata\", etc.) are also provided.")
  72     (license perl-license)))
  73
  74 (define-public perl-lingua-en-inflect-number
  75   (package
  76     (name "perl-lingua-en-inflect-number")
  77     (version "1.12")
  78     (source
  79      (origin
  80        (method url-fetch)
  81        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  82                            "Lingua-EN-Inflect-Number-" version ".tar.gz"))
  83        (sha256
  84         (base32
  85          "1gxccynkaqav43ww43jp4rzkyr36x97jd03yb5f6yx0jhn1k7yv6"))))
  86     (build-system perl-build-system)
  87     (propagated-inputs
  88      `(("perl-lingua-en-inflect" ,perl-lingua-en-inflect)))
  89     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect-Number")
  90     (synopsis "Force number of words to singular or plural")
  91     (description "This module extends the functionality of Lingua::EN::Inflect
  92 with three new functions for determining plurality of a word and forcefully
  93 converting a word to singular or plural.")
  94     (license perl-license)))
  95
  96 (define-public perl-lingua-en-inflect-phrase
  97   (package
  98     (name "perl-lingua-en-inflect-phrase")
  99     (version "0.18")
 100     (source
 101      (origin
 102        (method url-fetch)
 103        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 104                            "Lingua-EN-Inflect-Phrase-" version ".tar.gz"))
 105        (sha256
 106         (base32
 107          "19prg6pkgnznrc82ymdkdi3l9schg815axjig7adca5yqa7mn2i9"))))
 108     (build-system perl-build-system)
 109     (native-inputs
 110      `(("perl-test-nowarnings" ,perl-test-nowarnings)))
 111     (propagated-inputs
 112      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)
 113        ("perl-lingua-en-inflect" ,perl-lingua-en-inflect)
 114        ("perl-lingua-en-inflect-number" ,perl-lingua-en-inflect-number)
 115        ("perl-lingua-en-number-isordinal" ,perl-lingua-en-number-isordinal)
 116        ("perl-lingua-en-tagger" ,perl-lingua-en-tagger)))
 117     (home-page "http://search.cpan.org/dist/Lingua-EN-Inflect-Phrase")
 118     (synopsis "Inflect short English phrases")
 119     (description "This module attempts to pluralize or singularize short
 120 English phrases.")
 121     (license perl-license)))
 122
 123 (define-public perl-lingua-en-number-isordinal
 124   (package
 125     (name "perl-lingua-en-number-isordinal")
 126     (version "0.05")
 127     (source
 128      (origin
 129        (method url-fetch)
 130        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 131                            "Lingua-EN-Number-IsOrdinal-" version ".tar.gz"))
 132        (sha256
 133         (base32
 134          "1mhqjvh2ad30gjab5b3a6mbr4aysyrscp4wp42yy5x6001a6km98"))))
 135     (build-system perl-build-system)
 136     (native-inputs
 137      `(("perl-try-tiny" ,perl-try-tiny)
 138        ("perl-test-fatal" ,perl-test-fatal)))
 139     (propagated-inputs
 140      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)))
 141     (home-page "http://search.cpan.org/dist/Lingua-EN-Number-IsOrdinal")
 142     (synopsis "Detect if English number is ordinal or cardinal")
 143     (description "This module will tell you if a number, either in words or as
 144 digits, is a cardinal or ordinal number.")
 145     (license perl-license)))
 146
 147 (define-public perl-lingua-en-tagger
 148   (package
 149     (name "perl-lingua-en-tagger")
 150     (version "0.28")
 151     (source
 152      (origin
 153        (method url-fetch)
 154        (uri (string-append "mirror://cpan/authors/id/A/AC/ACOBURN/"
 155                            "Lingua-EN-Tagger-" version ".tar.gz"))
 156        (sha256
 157         (base32
 158          "1rm223d871s150dgnhbar71w56bim950cp45f0904w6px5fifcqa"))))
 159     (build-system perl-build-system)
 160     (propagated-inputs
 161      `(("perl-memoize-expirelru" ,perl-memoize-expirelru)
 162        ("perl-lingua-stem" ,perl-lingua-stem)
 163        ("perl-html-parser" ,perl-html-parser)
 164        ("perl-html-tagset" ,perl-html-tagset)))
 165     (home-page "http://search.cpan.org/dist/Lingua-EN-Tagger")
 166     (synopsis "Part-of-speech tagger for English natural language processing")
 167     (description "This module is a probability based, corpus-trained tagger
 168 that assigns part-of-speech tags to English text based on a lookup dictionary
 169 and a set of probability values.  The tagger assigns appropriate tags based on
 170 conditional probabilities - it examines the preceding tag to determine the
 171 appropriate tag for the current word.  Unknown words are classified according
 172 to word morphology or can be set to be treated as nouns or other parts of
 173 speech.  The tagger also extracts as many nouns and noun phrases as it can,
 174 using a set of regular expressions.")
 175     (license gpl3)))
 176
 177 (define-public perl-lingua-en-words2nums
 178   (package
 179     (name "perl-lingua-en-words2nums")
 180     (version "0.18")
 181     (source
 182      (origin
 183        (method url-fetch)
 184        (uri (string-append "mirror://cpan/authors/id/J/JO/JOEY/"
 185                            "Lingua-EN-Words2Nums-" version ".tar.gz"))
 186        (sha256
 187         (base32
 188          "118xx8qr1zbx30psv7ic55w65h15mc1vz6zicshfm96jgiwmcrb8"))))
 189     (build-system perl-build-system)
 190     (home-page "http://search.cpan.org/dist/Lingua-EN-Words2Nums")
 191     (synopsis "Convert English text to numbers")
 192     (description "This module converts English text into numbers.  It supports
 193 both ordinal and cardinal numbers, negative numbers, and very large numbers.")
 194     (license perl-license)))
 195
 196 (define-public perl-lingua-pt-stemmer
 197   (package
 198     (name "perl-lingua-pt-stemmer")
 199     (version "0.01")
 200     (source
 201      (origin
 202        (method url-fetch)
 203        (uri (string-append "mirror://cpan/authors/id/X/XE/XERN/"
 204                            "Lingua-PT-Stemmer-" version ".tar.gz"))
 205        (sha256
 206         (base32
 207          "11rqc5pqnkl9c13vy7sihiyas14ci0pj3k6chrgrgjv5sjv2m4a5"))))
 208     (build-system perl-build-system)
 209     (home-page "http://search.cpan.org/dist/Lingua-PT-Stemmer")
 210     (synopsis "Portuguese language stemming")
 211     (description "This module implements a Portuguese stemming algorithm
 212 proposed in the paper A Stemming Algorithm for the Portuguese Language by
 213 Moreira, V. and Huyck, C.")
 214     (license perl-license)))
 215
 216 (define-public perl-lingua-stem
 217   (package
 218     (name "perl-lingua-stem")
 219     (version "0.84")
 220     (source
 221      (origin
 222        (method url-fetch)
 223        (uri (string-append "mirror://cpan/authors/id/S/SN/SNOWHARE/"
 224                            "Lingua-Stem-" version ".tar.gz"))
 225        (sha256
 226         (base32
 227          "12avh2mnnc7llmmshrr5bgb473fvydxnlqrqbl2815mf2dp4pxcg"))))
 228     (build-system perl-build-system)
 229     (native-inputs
 230      `(("perl-module-build" ,perl-module-build)))
 231     (propagated-inputs
 232      `(("perl-lingua-pt-stemmer" ,perl-lingua-pt-stemmer)
 233        ("perl-lingua-stem-fr" ,perl-lingua-stem-fr)
 234        ("perl-lingua-stem-it" ,perl-lingua-stem-it)
 235        ("perl-lingua-stem-ru" ,perl-lingua-stem-ru)
 236        ("perl-lingua-stem-snowball-da" ,perl-lingua-stem-snowball-da)
 237        ("perl-snowball-norwegian" ,perl-snowball-norwegian)
 238        ("perl-snowball-swedish" ,perl-snowball-swedish)
 239        ("perl-text-german" ,perl-text-german)))
 240     (home-page "http://search.cpan.org/dist/Lingua-Stem")
 241     (synopsis "Stemming of words in various languages")
 242     (description "This routine applies stemming algorithms to its parameters,
 243 returning the stemmed words as appropriate to the selected locale.")
 244     (license perl-license)))
 245
 246 (define-public perl-lingua-stem-fr
 247   (package
 248     (name "perl-lingua-stem-fr")
 249     (version "0.02")
 250     (source
 251      (origin
 252        (method url-fetch)
 253        (uri (string-append "mirror://cpan/authors/id/S/SD/SDP/"
 254                            "Lingua-Stem-Fr-" version ".tar.gz"))
 255        (sha256
 256         (base32
 257          "0vyrspwzaqjxm5mqshf4wvwa3938mkajd1918d9ii2l9m2rn8kwx"))))
 258     (build-system perl-build-system)
 259     (home-page "http://search.cpan.org/dist/Lingua-Stem-Fr")
 260     (synopsis "Porter's stemming algorithm for French")
 261     (description "This module uses a modified version of the Porter Stemming
 262 Algorithm to return a stemmed French word.")
 263     (license perl-license)))
 264
 265 (define-public perl-lingua-stem-it
 266   (package
 267     (name "perl-lingua-stem-it")
 268     (version "0.02")
 269     (source
 270      (origin
 271        (method url-fetch)
 272        (uri (string-append "mirror://cpan/authors/id/A/AC/ACALPINI/"
 273                            "Lingua-Stem-It-" version ".tar.gz"))
 274        (sha256
 275         (base32
 276          "1207r183s5hlh4mfwa6p46vzm0dhvrs2dnss5s41a0gyfkxp7riq"))))
 277     (build-system perl-build-system)
 278     (home-page "http://search.cpan.org/dist/Lingua-Stem-It")
 279     (synopsis "Porter's stemming algorithm for Italian")
 280     (description "This module applies the Porter Stemming Algorithm to its
 281 parameters, returning the stemmed Italian word.")
 282     (license perl-license)))
 283
 284 (define-public perl-lingua-stem-ru
 285   (package
 286     (name "perl-lingua-stem-ru")
 287     (version "0.01")
 288     (source
 289      (origin
 290        (method url-fetch)
 291        (uri (string-append "mirror://cpan/authors/id/A/AL/ALGDR/"
 292                            "Lingua-Stem-Ru-" version ".tar.gz"))
 293        (sha256
 294         (base32
 295          "0pqgg442fkf12ayh9fgmpa8q9x0iqx6s96181r52yn7s7pcs61h6"))))
 296     (build-system perl-build-system)
 297     (home-page "http://search.cpan.org/dist/Lingua-Stem-Ru")
 298     (synopsis "Porter's stemming algorithm for Russian")
 299     (description "This module applies the Porter Stemming Algorithm to its
 300 parameters, returning the stemmed Russian (KOI8-R only) word.")
 301     (license perl-license)))
 302
 303 (define-public perl-lingua-stem-snowball-da
 304   (package
 305     (name "perl-lingua-stem-snowball-da")
 306     (version "1.01")
 307     (source
 308      (origin
 309        (method url-fetch)
 310        (uri (string-append "mirror://cpan/authors/id/C/CI/CINE/"
 311                            "Lingua-Stem-Snowball-Da-" version ".tar.gz"))
 312        (sha256
 313         (base32
 314          "0mm0m7glm1s6i9f6a78jslw6wh573208arxhq93yriqmw17bwf9f"))))
 315     (build-system perl-build-system)
 316     (home-page "http://search.cpan.org/dist/Lingua-Stem-Snowball-Da")
 317     (synopsis "Porters stemming algorithm for Danish")
 318     (description "Lingua::Stem::Snowball::Da is a perl port of the danish
 319 stemmer at http://snowball.sourceforge.net, it was originally altered from the
 320 Lingua::Stem::Snowball::Se.")
 321     (license gpl2)))
 322
 323 (define-public perl-snowball-norwegian
 324   (package
 325     (name "perl-snowball-norwegian")
 326     (version "1.2")
 327     (source
 328      (origin
 329        (method url-fetch)
 330        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 331                            "Snowball-Norwegian-" version ".tar.gz"))
 332        (sha256
 333         (base32
 334          "0675v45bbsh7vr7kpf36xs2q79g02iq1kmfw22h20xdk4rzqvkqx"))))
 335     (build-system perl-build-system)
 336     (native-inputs `(("perl-module-build" ,perl-module-build)))
 337     (home-page "http://search.cpan.org/dist/Snowball-Norwegian")
 338     (synopsis "Porters stemming algorithm for Norwegian")
 339     (description "Lingua::Stem::Snowball::No is a perl port of the norwegian
 340 stemmer at http://snowball.tartarus.org.")
 341     (license perl-license)))
 342
 343 (define-public perl-snowball-swedish
 344   (package
 345     (name "perl-snowball-swedish")
 346     (version "1.2")
 347     (source
 348      (origin
 349        (method url-fetch)
 350        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 351                            "Snowball-Swedish-" version ".tar.gz"))
 352        (sha256
 353         (base32
 354          "0agwc12jk5kmabnpsplw3wf4ii5w1zb159cpin44x3srb0sr5apg"))))
 355     (build-system perl-build-system)
 356     (native-inputs `(("perl-module-build" ,perl-module-build)))
 357     (home-page "http://search.cpan.org/dist/Snowball-Swedish")
 358     (synopsis "Porters stemming algorithm for Swedish")
 359     (description "Lingua::Stem::Snowball::Se is a perl port of the swedish
 360 stemmer at http://snowball.sourceforge.net.")
 361     (license perl-license)))
 362
 363 (define-public perl-string-toidentifier-en
 364   (package
 365     (name "perl-string-toidentifier-en")
 366     (version "0.11")
 367     (source
 368      (origin
 369        (method url-fetch)
 370        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 371                            "String-ToIdentifier-EN-" version ".tar.gz"))
 372        (sha256
 373         (base32
 374          "1bawghkgkkx7j3avnrj5sg3vix1z5564ks6wf9az3jc2knh8s5nh"))))
 375     (build-system perl-build-system)
 376     (propagated-inputs
 377      `(("perl-lingua-en-inflect-phrase" ,perl-lingua-en-inflect-phrase)
 378        ("perl-text-unidecode" ,perl-text-unidecode)
 379        ("perl-namespace-clean" ,perl-namespace-clean)))
 380     (home-page "http://search.cpan.org/dist/String-ToIdentifier-EN")
 381     (synopsis "Convert strings to English program identifiers")
 382     (description "This module provides a utility method, \"to_identifier\" for
 383 converting an arbitrary string into a readable representation using the ASCII
 384 subset of \"\\w\" for use as an identifier in a computer program.  The intent
 385 is to make unique identifier names from which the content of the original
 386 string can be easily inferred by a human just by reading the identifier.")
 387     (license perl-license)))
 388
 389 (define-public perl-text-german
 390   (package
 391     (name "perl-text-german")
 392     (version "0.06")
 393     (source
 394      (origin
 395        (method url-fetch)
 396        (uri (string-append "mirror://cpan/authors/id/U/UL/ULPFR/"
 397                            "Text-German-" version ".tar.gz"))
 398        (sha256
 399         (base32
 400          "1p87pgap99lw0nv62i3ghvsi7yg90lhn8vsa3yqp75rd04clybcj"))))
 401     (build-system perl-build-system)
 402     (home-page "http://search.cpan.org/dist/Text-German")
 403     (synopsis "German grundform reduction")
 404     (description "This module is a rather incomplete implementation of work
 405 done by Gudrun Putze-Meier.")
 406     (license perl-license)))