gnu/packages/language.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2015, 2016 Eric Bavier <bavier@member.fsf.org>
   3 ;;; Copyright © 2017, 2018 Tobias Geerinckx-Rice <me@tobias.gr>
   4 ;;; Copyright © 2018 Nils Gillmann <ng0@n0.is>
   5 ;;;
   6 ;;; This file is part of GNU Guix.
   7 ;;;
   8 ;;; GNU Guix is free software; you can redistribute it and/or modify it
   9 ;;; under the terms of the GNU General Public License as published by
  10 ;;; the Free Software Foundation; either version 3 of the License, or (at
  11 ;;; your option) any later version.
  12 ;;;
  13 ;;; GNU Guix is distributed in the hope that it will be useful, but
  14 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  15 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 ;;; GNU General Public License for more details.
  17 ;;;
  18 ;;; You should have received a copy of the GNU General Public License
  19 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  20
  21 (define-module (gnu packages language)
  22   #:use-module (gnu packages)
  23   #:use-module (gnu packages perl)
  24   #:use-module (gnu packages perl-check)
  25   #:use-module (gnu packages web)
  26   #:use-module (guix packages)
  27   #:use-module (guix build-system perl)
  28   #:use-module ((guix licenses) #:select (gpl2 gpl3 perl-license))
  29   #:use-module (guix download))
  30
  31 (define-public perl-lingua-en-findnumber
  32   (package
  33     (name "perl-lingua-en-findnumber")
  34     (version "1.32")
  35     (source
  36      (origin
  37        (method url-fetch)
  38        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  39                            "Lingua-EN-FindNumber-" version ".tar.gz"))
  40        (sha256
  41         (base32
  42          "015ximzdp42v824llwlg2pd77vd0d172lb4xs55q9f9zhqf6s5qx"))))
  43     (build-system perl-build-system)
  44     (propagated-inputs
  45      `(("perl-lingua-en-words2nums" ,perl-lingua-en-words2nums)))
  46     (home-page "https://metacpan.org/release/Lingua-EN-FindNumber")
  47     (synopsis "Locate (written) numbers in English text ")
  48     (description "This module provides a regular expression for finding
  49 numbers in English text.  It also provides functions for extracting and
  50 manipulating such numbers.")
  51     (license perl-license)))
  52
  53 (define-public perl-lingua-en-inflect
  54   (package
  55     (name "perl-lingua-en-inflect")
  56     (version "1.903")
  57     (source
  58      (origin
  59        (method url-fetch)
  60        (uri (string-append "mirror://cpan/authors/id/D/DC/DCONWAY/"
  61                            "Lingua-EN-Inflect-" version ".tar.gz"))
  62        (sha256
  63         (base32
  64          "0j8d1f1wvmgc11d71pc8xp8fv5a1nb2yfw1dgd19xhscn1klpvzw"))))
  65     (build-system perl-build-system)
  66     (native-inputs `(("perl-module-build" ,perl-module-build)))
  67     (home-page "https://metacpan.org/release/Lingua-EN-Inflect")
  68     (synopsis "Convert singular to plural")
  69     (description "Lingua::EN::Inflect provides plural inflections,
  70 \"a\"/\"an\" selection for English words, and manipulation of numbers as
  71 words.  Plural forms of all nouns, most verbs, and some adjectives are
  72 provided.  Where appropriate, \"classical\" variants (for example: \"brother\"
  73 -> \"brethren\", \"dogma\" -> \"dogmata\", etc.) are also provided.")
  74     (license perl-license)))
  75
  76 (define-public perl-lingua-en-inflect-number
  77   (package
  78     (name "perl-lingua-en-inflect-number")
  79     (version "1.12")
  80     (source
  81      (origin
  82        (method url-fetch)
  83        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
  84                            "Lingua-EN-Inflect-Number-" version ".tar.gz"))
  85        (sha256
  86         (base32
  87          "1gxccynkaqav43ww43jp4rzkyr36x97jd03yb5f6yx0jhn1k7yv6"))))
  88     (build-system perl-build-system)
  89     (propagated-inputs
  90      `(("perl-lingua-en-inflect" ,perl-lingua-en-inflect)))
  91     (home-page "https://metacpan.org/release/Lingua-EN-Inflect-Number")
  92     (synopsis "Force number of words to singular or plural")
  93     (description "This module extends the functionality of Lingua::EN::Inflect
  94 with three new functions for determining plurality of a word and forcefully
  95 converting a word to singular or plural.")
  96     (license perl-license)))
  97
  98 (define-public perl-lingua-en-inflect-phrase
  99   (package
 100     (name "perl-lingua-en-inflect-phrase")
 101     (version "0.20")
 102     (source
 103      (origin
 104        (method url-fetch)
 105        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 106                            "Lingua-EN-Inflect-Phrase-" version ".tar.gz"))
 107        (sha256
 108         (base32
 109          "1a6y1l2pjim2242wcpgz066di4pbzfgsjjdl7vg5a5wzm48qj1am"))))
 110     (build-system perl-build-system)
 111     (native-inputs
 112      `(("perl-test-nowarnings" ,perl-test-nowarnings)))
 113     (propagated-inputs
 114      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)
 115        ("perl-lingua-en-inflect" ,perl-lingua-en-inflect)
 116        ("perl-lingua-en-inflect-number" ,perl-lingua-en-inflect-number)
 117        ("perl-lingua-en-number-isordinal" ,perl-lingua-en-number-isordinal)
 118        ("perl-lingua-en-tagger" ,perl-lingua-en-tagger)))
 119     (home-page "https://metacpan.org/release/Lingua-EN-Inflect-Phrase")
 120     (synopsis "Inflect short English phrases")
 121     (description "This module attempts to pluralize or singularize short
 122 English phrases.")
 123     (license perl-license)))
 124
 125 (define-public perl-lingua-en-number-isordinal
 126   (package
 127     (name "perl-lingua-en-number-isordinal")
 128     (version "0.05")
 129     (source
 130      (origin
 131        (method url-fetch)
 132        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 133                            "Lingua-EN-Number-IsOrdinal-" version ".tar.gz"))
 134        (sha256
 135         (base32
 136          "1mhqjvh2ad30gjab5b3a6mbr4aysyrscp4wp42yy5x6001a6km98"))))
 137     (build-system perl-build-system)
 138     (native-inputs
 139      `(("perl-try-tiny" ,perl-try-tiny)
 140        ("perl-test-fatal" ,perl-test-fatal)))
 141     (propagated-inputs
 142      `(("perl-lingua-en-findnumber" ,perl-lingua-en-findnumber)))
 143     (home-page "https://metacpan.org/release/Lingua-EN-Number-IsOrdinal")
 144     (synopsis "Detect if English number is ordinal or cardinal")
 145     (description "This module will tell you if a number, either in words or as
 146 digits, is a cardinal or ordinal number.")
 147     (license perl-license)))
 148
 149 (define-public perl-lingua-en-tagger
 150   (package
 151     (name "perl-lingua-en-tagger")
 152     (version "0.29")
 153     (source
 154      (origin
 155        (method url-fetch)
 156        (uri (string-append "mirror://cpan/authors/id/A/AC/ACOBURN/"
 157                            "Lingua-EN-Tagger-" version ".tar.gz"))
 158        (sha256
 159         (base32
 160          "0dssn101kmpkh2ik1430mj2ikk04849vbpgi60382kvh9xn795na"))))
 161     (build-system perl-build-system)
 162     (propagated-inputs
 163      `(("perl-memoize-expirelru" ,perl-memoize-expirelru)
 164        ("perl-lingua-stem" ,perl-lingua-stem)
 165        ("perl-html-parser" ,perl-html-parser)
 166        ("perl-html-tagset" ,perl-html-tagset)))
 167     (home-page "https://metacpan.org/release/Lingua-EN-Tagger")
 168     (synopsis "Part-of-speech tagger for English natural language processing")
 169     (description "This module is a probability based, corpus-trained tagger
 170 that assigns part-of-speech tags to English text based on a lookup dictionary
 171 and a set of probability values.  The tagger assigns appropriate tags based on
 172 conditional probabilities - it examines the preceding tag to determine the
 173 appropriate tag for the current word.  Unknown words are classified according
 174 to word morphology or can be set to be treated as nouns or other parts of
 175 speech.  The tagger also extracts as many nouns and noun phrases as it can,
 176 using a set of regular expressions.")
 177     (license gpl3)))
 178
 179 (define-public perl-lingua-en-words2nums
 180   (package
 181     (name "perl-lingua-en-words2nums")
 182     (version "0.18")
 183     (source
 184      (origin
 185        (method url-fetch)
 186        (uri (string-append "mirror://cpan/authors/id/J/JO/JOEY/"
 187                            "Lingua-EN-Words2Nums-" version ".tar.gz"))
 188        (sha256
 189         (base32
 190          "118xx8qr1zbx30psv7ic55w65h15mc1vz6zicshfm96jgiwmcrb8"))))
 191     (build-system perl-build-system)
 192     (home-page "https://metacpan.org/release/Lingua-EN-Words2Nums")
 193     (synopsis "Convert English text to numbers")
 194     (description "This module converts English text into numbers.  It supports
 195 both ordinal and cardinal numbers, negative numbers, and very large numbers.")
 196     (license perl-license)))
 197
 198 (define-public perl-lingua-pt-stemmer
 199   (package
 200     (name "perl-lingua-pt-stemmer")
 201     (version "0.02")
 202     (source
 203      (origin
 204        (method url-fetch)
 205        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
 206                            "Lingua-PT-Stemmer-" version ".tar.gz"))
 207        (sha256
 208         (base32
 209          "17c48sfbgwd2ivlgf59sr6jdhwa3aim8750f8pyzz7xpi8gz0var"))))
 210     (build-system perl-build-system)
 211     (home-page "https://metacpan.org/release/Lingua-PT-Stemmer")
 212     (synopsis "Portuguese language stemming")
 213     (description "This module implements a Portuguese stemming algorithm
 214 proposed in the paper A Stemming Algorithm for the Portuguese Language by
 215 Moreira, V. and Huyck, C.")
 216     (license perl-license)))
 217
 218 (define-public perl-lingua-stem
 219   (package
 220     (name "perl-lingua-stem")
 221     (version "0.84")
 222     (source
 223      (origin
 224        (method url-fetch)
 225        (uri (string-append "mirror://cpan/authors/id/S/SN/SNOWHARE/"
 226                            "Lingua-Stem-" version ".tar.gz"))
 227        (sha256
 228         (base32
 229          "12avh2mnnc7llmmshrr5bgb473fvydxnlqrqbl2815mf2dp4pxcg"))))
 230     (build-system perl-build-system)
 231     (native-inputs
 232      `(("perl-module-build" ,perl-module-build)))
 233     (propagated-inputs
 234      `(("perl-lingua-pt-stemmer" ,perl-lingua-pt-stemmer)
 235        ("perl-lingua-stem-fr" ,perl-lingua-stem-fr)
 236        ("perl-lingua-stem-it" ,perl-lingua-stem-it)
 237        ("perl-lingua-stem-ru" ,perl-lingua-stem-ru)
 238        ("perl-lingua-stem-snowball-da" ,perl-lingua-stem-snowball-da)
 239        ("perl-snowball-norwegian" ,perl-snowball-norwegian)
 240        ("perl-snowball-swedish" ,perl-snowball-swedish)
 241        ("perl-text-german" ,perl-text-german)))
 242     (home-page "https://metacpan.org/release/Lingua-Stem")
 243     (synopsis "Stemming of words in various languages")
 244     (description "This routine applies stemming algorithms to its parameters,
 245 returning the stemmed words as appropriate to the selected locale.")
 246     (license perl-license)))
 247
 248 (define-public perl-lingua-stem-fr
 249   (package
 250     (name "perl-lingua-stem-fr")
 251     (version "0.02")
 252     (source
 253      (origin
 254        (method url-fetch)
 255        (uri (string-append "mirror://cpan/authors/id/S/SD/SDP/"
 256                            "Lingua-Stem-Fr-" version ".tar.gz"))
 257        (sha256
 258         (base32
 259          "0vyrspwzaqjxm5mqshf4wvwa3938mkajd1918d9ii2l9m2rn8kwx"))))
 260     (build-system perl-build-system)
 261     (home-page "https://metacpan.org/release/Lingua-Stem-Fr")
 262     (synopsis "Porter's stemming algorithm for French")
 263     (description "This module uses a modified version of the Porter Stemming
 264 Algorithm to return a stemmed French word.")
 265     (license perl-license)))
 266
 267 (define-public perl-lingua-stem-it
 268   (package
 269     (name "perl-lingua-stem-it")
 270     (version "0.02")
 271     (source
 272      (origin
 273        (method url-fetch)
 274        (uri (string-append "mirror://cpan/authors/id/A/AC/ACALPINI/"
 275                            "Lingua-Stem-It-" version ".tar.gz"))
 276        (sha256
 277         (base32
 278          "1207r183s5hlh4mfwa6p46vzm0dhvrs2dnss5s41a0gyfkxp7riq"))))
 279     (build-system perl-build-system)
 280     (home-page "https://metacpan.org/release/Lingua-Stem-It")
 281     (synopsis "Porter's stemming algorithm for Italian")
 282     (description "This module applies the Porter Stemming Algorithm to its
 283 parameters, returning the stemmed Italian word.")
 284     (license perl-license)))
 285
 286 (define-public perl-lingua-stem-ru
 287   (package
 288     (name "perl-lingua-stem-ru")
 289     (version "0.04")
 290     (source
 291      (origin
 292        (method url-fetch)
 293        (uri (string-append "mirror://cpan/authors/id/N/NE/NEILB/"
 294                            "Lingua-Stem-Ru-" version ".tar.gz"))
 295        (sha256
 296         (base32
 297          "0a2jmdz7jn32qj5hyiw5kbv8fvlpmws8i00a6xcbkzb48yvwww0j"))))
 298     (build-system perl-build-system)
 299     (home-page "https://metacpan.org/release/Lingua-Stem-Ru")
 300     (synopsis "Porter's stemming algorithm for Russian")
 301     (description "This module applies the Porter Stemming Algorithm to its
 302 parameters, returning the stemmed Russian (KOI8-R only) word.")
 303     (license perl-license)))
 304
 305 (define-public perl-lingua-stem-snowball-da
 306   (package
 307     (name "perl-lingua-stem-snowball-da")
 308     (version "1.01")
 309     (source
 310      (origin
 311        (method url-fetch)
 312        (uri (string-append "mirror://cpan/authors/id/C/CI/CINE/"
 313                            "Lingua-Stem-Snowball-Da-" version ".tar.gz"))
 314        (sha256
 315         (base32
 316          "0mm0m7glm1s6i9f6a78jslw6wh573208arxhq93yriqmw17bwf9f"))))
 317     (build-system perl-build-system)
 318     (home-page "https://metacpan.org/release/Lingua-Stem-Snowball-Da")
 319     (synopsis "Porters stemming algorithm for Danish")
 320     (description "Lingua::Stem::Snowball::Da is a perl port of the danish
 321 stemmer at http://snowball.sourceforge.net, it was originally altered from the
 322 Lingua::Stem::Snowball::Se.")
 323     (license gpl2)))
 324
 325 (define-public perl-snowball-norwegian
 326   (package
 327     (name "perl-snowball-norwegian")
 328     (version "1.2")
 329     (source
 330      (origin
 331        (method url-fetch)
 332        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 333                            "Snowball-Norwegian-" version ".tar.gz"))
 334        (sha256
 335         (base32
 336          "0675v45bbsh7vr7kpf36xs2q79g02iq1kmfw22h20xdk4rzqvkqx"))))
 337     (build-system perl-build-system)
 338     (native-inputs `(("perl-module-build" ,perl-module-build)))
 339     (home-page "https://metacpan.org/release/Snowball-Norwegian")
 340     (synopsis "Porters stemming algorithm for Norwegian")
 341     (description "Lingua::Stem::Snowball::No is a perl port of the norwegian
 342 stemmer at http://snowball.tartarus.org.")
 343     (license perl-license)))
 344
 345 (define-public perl-snowball-swedish
 346   (package
 347     (name "perl-snowball-swedish")
 348     (version "1.2")
 349     (source
 350      (origin
 351        (method url-fetch)
 352        (uri (string-append "mirror://cpan/authors/id/A/AS/ASKSH/"
 353                            "Snowball-Swedish-" version ".tar.gz"))
 354        (sha256
 355         (base32
 356          "0agwc12jk5kmabnpsplw3wf4ii5w1zb159cpin44x3srb0sr5apg"))))
 357     (build-system perl-build-system)
 358     (native-inputs `(("perl-module-build" ,perl-module-build)))
 359     (home-page "https://metacpan.org/release/Snowball-Swedish")
 360     (synopsis "Porters stemming algorithm for Swedish")
 361     (description "Lingua::Stem::Snowball::Se is a perl port of the swedish
 362 stemmer at http://snowball.sourceforge.net.")
 363     (license perl-license)))
 364
 365 (define-public perl-string-toidentifier-en
 366   (package
 367     (name "perl-string-toidentifier-en")
 368     (version "0.12")
 369     (source
 370      (origin
 371        (method url-fetch)
 372        (uri (string-append "mirror://cpan/authors/id/R/RK/RKITOVER/"
 373                            "String-ToIdentifier-EN-" version ".tar.gz"))
 374        (sha256
 375         (base32
 376          "12nw7h2yiybhdw0vnnpc7bif8ylhsn6kqf6s39dsrf9h54iq9yrs"))))
 377     (build-system perl-build-system)
 378     (propagated-inputs
 379      `(("perl-lingua-en-inflect-phrase" ,perl-lingua-en-inflect-phrase)
 380        ("perl-text-unidecode" ,perl-text-unidecode)
 381        ("perl-namespace-clean" ,perl-namespace-clean)))
 382     (home-page "https://metacpan.org/release/String-ToIdentifier-EN")
 383     (synopsis "Convert strings to English program identifiers")
 384     (description "This module provides a utility method, \"to_identifier\" for
 385 converting an arbitrary string into a readable representation using the ASCII
 386 subset of \"\\w\" for use as an identifier in a computer program.  The intent
 387 is to make unique identifier names from which the content of the original
 388 string can be easily inferred by a human just by reading the identifier.")
 389     (license perl-license)))
 390
 391 (define-public perl-text-german
 392   (package
 393     (name "perl-text-german")
 394     (version "0.06")
 395     (source
 396      (origin
 397        (method url-fetch)
 398        (uri (string-append "mirror://cpan/authors/id/U/UL/ULPFR/"
 399                            "Text-German-" version ".tar.gz"))
 400        (sha256
 401         (base32
 402          "1p87pgap99lw0nv62i3ghvsi7yg90lhn8vsa3yqp75rd04clybcj"))))
 403     (build-system perl-build-system)
 404     (home-page "https://metacpan.org/release/Text-German")
 405     (synopsis "German grundform reduction")
 406     (description "This module is a rather incomplete implementation of work
 407 done by Gudrun Putze-Meier.")
 408     (license perl-license)))