gnu/packages/textutils.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2015 Taylan Ulrich Bayırlı/Kammer <taylanbayirli@gmail.com>
   3 ;;; Copyright © 2015, 2016 Ricardo Wurmus <rekado@elephly.net>
   4 ;;; Copyright © 2015, 2016 Ben Woodcroft <donttrustben@gmail.com>
   5 ;;; Copyright © 2015 Roel Janssen <roel@gnu.org>
   6 ;;; Copyright © 2016 Jelle Licht <jlicht@fsfe.org>
   7 ;;; Copyright © 2016 Alex Griffin <a@ajgrf.com>
   8 ;;;
   9 ;;; This file is part of GNU Guix.
  10 ;;;
  11 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  12 ;;; under the terms of the GNU General Public License as published by
  13 ;;; the Free Software Foundation; either version 3 of the License, or (at
  14 ;;; your option) any later version.
  15 ;;;
  16 ;;; GNU Guix is distributed in the hope that it will be useful, but
  17 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  18 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 ;;; GNU General Public License for more details.
  20 ;;;
  21 ;;; You should have received a copy of the GNU General Public License
  22 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  23
  24 (define-module (gnu packages textutils)
  25   #:use-module ((guix licenses) #:prefix license:)
  26   #:use-module (guix packages)
  27   #:use-module (guix download)
  28   #:use-module (guix git-download)
  29   #:use-module (guix build-system gnu)
  30   #:use-module (guix build-system cmake)
  31   #:use-module (guix build-system trivial)
  32   #:use-module (gnu packages autotools)
  33   #:use-module (gnu packages python)
  34   #:use-module (gnu packages zip))
  35
  36 (define-public recode
  37   (package
  38     (name "recode")
  39     ;; Last beta release (3.7-beta2) is from 2008; last commit from Feb 2014.
  40     ;; So we use that commit instead.
  41     (version "3.7.0.201402")
  42     (source
  43      (origin
  44        (method git-fetch)
  45        (uri (git-reference
  46              (url "https://github.com/pinard/Recode.git")
  47              (commit "2d7092a9999194fc0e9449717a8048c8d8e26c18")))
  48        (sha256
  49         (base32 "1wssv8z6g3ryrw33sksz4rjhlnhgvvdqszw1ggl4rcwks34n86zm"))
  50        (file-name (string-append name "-" version "-checkout"))))
  51     (build-system gnu-build-system)
  52     (native-inputs `(("python" ,python-2)))
  53     (arguments
  54      '(#:phases
  55        (alist-cons-before
  56         'check 'pre-check
  57         (lambda _
  58           (substitute* "tests/setup.py"
  59             (("([[:space:]]*)include_dirs=.*" all space)
  60              (string-append all space "library_dirs=['../src/.libs'],\n")))
  61           ;; The test extension 'Recode.so' lacks RUNPATH for 'librecode.so'.
  62           (setenv "LD_LIBRARY_PATH" (string-append (getcwd) "/src/.libs")))
  63         %standard-phases)))
  64     (home-page "https://github.com/pinard/Recode")
  65     (synopsis "Text encoding converter")
  66     (description "The Recode library converts files between character sets and
  67 usages.  It recognises or produces over 200 different character sets (or about
  68 300 if combined with an iconv library) and transliterates files between almost
  69 any pair.  When exact transliteration are not possible, it gets rid of
  70 offending characters or falls back on approximations.  The recode program is a
  71 handy front-end to the library.")
  72     (license license:gpl2+)))
  73
  74 (define-public enca
  75   (package
  76     (name "enca")
  77     (version "1.16")
  78     (source
  79      (origin
  80        (method url-fetch)
  81        (uri (string-append
  82              "https://github.com/nijel/enca/archive/" version ".tar.gz"))
  83        (sha256
  84         (base32 "1xik00x0yvhswsw2isnclabhv536xk1s42cf5z54gfbpbhc7ni8l"))
  85        (file-name (string-append name "-" version ".tar.gz"))))
  86     (build-system gnu-build-system)
  87     (inputs `(("recode" ,recode)))
  88
  89     ;; Both 'test-convert-64.sh' and 'test-convert-filter.sh' manipulate a
  90     ;; 'test.tmp' file, so they have to run in sequence.
  91     (arguments '(#:parallel-tests? #f))
  92
  93     (home-page "https://github.com/nijel/enca")
  94     (synopsis "Text encoding detection tool")
  95     (description "Enca (Extremely Naive Charset Analyser) consists of libenca,
  96 an encoding detection library, and enca, a command line frontend, integrating
  97 libenca and several charset conversion libraries and tools.")
  98     (license license:gpl2)))
  99
 100 (define-public utf8proc
 101   (package
 102     (name "utf8proc")
 103     (version "1.3.1")
 104     (source
 105      (origin
 106        (method url-fetch)
 107        (uri (string-append
 108              "https://github.com/JuliaLang/utf8proc/archive/v"
 109              version ".tar.gz"))
 110        (file-name (string-append name "-" version ".tar.gz"))
 111        (sha256
 112         (base32 "1k48as5kjkar4yj3dwxyll8ykj4k723ib5a6mnw1g86q3zi0zdl3"))))
 113     (build-system gnu-build-system)
 114     (arguments
 115      '(#:tests? #f ;no "check" target
 116        #:make-flags (list "CC=gcc"
 117                           (string-append "prefix=" (assoc-ref %outputs "out")))
 118        #:phases
 119        (modify-phases %standard-phases
 120          (delete 'configure))))
 121     (home-page "http://julialang.org/utf8proc/")
 122     (synopsis "C library for processing UTF-8 Unicode data")
 123     (description "utf8proc is a small C library that provides Unicode
 124 normalization, case-folding, and other operations for data in the UTF-8
 125 encoding, supporting Unicode version 7.0.")
 126     (license license:expat)))
 127
 128 (define-public libgtextutils
 129   (package
 130     (name "libgtextutils")
 131     (version "0.7")
 132     (source
 133      (origin
 134        (method url-fetch)
 135        (uri (string-append
 136              "https://github.com/agordon/libgtextutils/releases/download/"
 137              version "/libgtextutils-" version ".tar.gz"))
 138        (sha256
 139         (base32 "0jiybkb2z58wa2msvllnphr4js2hvjvh988pavb3mzkgr6ihwbkr"))))
 140     (build-system gnu-build-system)
 141     (arguments
 142      '(#:phases
 143        (alist-cons-after
 144         'unpack 'autoreconf
 145         (lambda _ (zero? (system* "autoreconf" "-vif")))
 146         %standard-phases)))
 147     (native-inputs
 148      `(("autoconf" ,autoconf)
 149        ("automake" ,automake)
 150        ("libtool" ,libtool)))
 151     (home-page "https://github.com/agordon/libgtextutils")
 152     (synopsis "Gordon's text utils library")
 153     (description
 154      "libgtextutils is a text utilities library used by the fastx toolkit from
 155 the Hannon Lab.")
 156     (license license:agpl3+)))
 157
 158 (define-public cityhash
 159   (let ((commit "8af9b8c"))
 160     (package
 161       (name "cityhash")
 162       (version (string-append "1.1-2." commit))
 163       (source (origin
 164                 (method git-fetch)
 165                 (uri (git-reference
 166                       (url "https://github.com/google/cityhash.git")
 167                       (commit commit)))
 168                 (file-name (string-append name "-" version ".tar.gz"))
 169                 (sha256
 170                  (base32
 171                   "0n6skf5dv8yfl1ckax8dqhvsbslkwc9158zf2ims0xqdvzsahbi6"))))
 172       (build-system gnu-build-system)
 173       (arguments
 174        '(#:make-flags (list "CXXFLAGS=-g -O3")
 175          #:phases
 176          (modify-phases %standard-phases
 177            ;; citycrc is not installed by default but is used by some
 178            ;; programs.
 179            (add-after 'install 'install-citycrc
 180              (lambda* (#:key outputs #:allow-other-keys)
 181                (let* ((out (assoc-ref outputs "out"))
 182                       (include (string-append out "/include")))
 183                  (install-file "src/citycrc.h" include))
 184                #t)))))
 185       (home-page "https://github.com/google/cityhash")
 186       (synopsis "C++ hash functions for strings")
 187       (description
 188        "CityHash provides hash functions for strings.  The functions mix the
 189 input bits thoroughly but are not suitable for cryptography.")
 190       (license license:expat))))
 191
 192 (define-public ustr
 193   (package
 194     (name "ustr")
 195     (version "1.0.4")
 196     (source (origin
 197               (method url-fetch)
 198               (uri (string-append "http://www.and.org/ustr/" version
 199                                   "/ustr-" version ".tar.bz2"))
 200               (sha256
 201                (base32
 202                 "1i623ygdj7rkizj7985q9d6vj5amwg686aqb5j3ixpkqkyp6xbrx"))))
 203     (build-system gnu-build-system)
 204     (arguments
 205      `(#:make-flags
 206        (list "CC=gcc"
 207              "HIDE="
 208              ;; Override "/sbin/ldconfig" with "echo" because we don't need
 209              ;; "ldconfig".
 210              "LDCONFIG=echo"
 211              (string-append "prefix=" (assoc-ref %outputs "out"))
 212              "all-shared")
 213        #:phases
 214        (modify-phases %standard-phases
 215          (add-after 'unpack 'disable-check-for-stdint
 216            (lambda _
 217              ;; Of course we have stdint.h, just not in /usr/include
 218              (substitute* '("Makefile"
 219                             "ustr-import.in")
 220                (("-f \"/usr/include/stdint.h\"") "-z \"\""))
 221              #t))
 222          ;; No configure script
 223          (delete 'configure))))
 224     (home-page "http://www.and.org/ustr/")
 225     (synopsis "String library with very low memory overhead")
 226     (description
 227      "Ustr is a string library for C with very low memory overhead.")
 228     ;; Quoted from the home page: "The License for the code is MIT, new-BSD,
 229     ;; LGPL, etc. ... if you need another license to help compatibility, just
 230     ;; ask for it.  It's basically public domain, without all the legal
 231     ;; problems for everyone that trying to make something public domain
 232     ;; entails."
 233     (license license:public-domain)))
 234
 235 (define-public libconfig
 236   (package
 237     (name "libconfig")
 238     (version "1.5")
 239     (source (origin
 240               (method url-fetch)
 241               (uri (string-append "http://www.hyperrealm.com/libconfig/"
 242                                   "libconfig-" version ".tar.gz"))
 243               (sha256
 244                (base32
 245                 "1xh3hzk63v4y8815lc5209m3s6ms2cpgw4h5hg462i4f1lwsl7g3"))))
 246     (build-system gnu-build-system)
 247     (home-page "http://www.hyperrealm.com/libconfig/")
 248     (synopsis "C/C++ configuration file library")
 249     (description
 250      "Libconfig is a simple library for manipulating structured configuration
 251 files.  This file format is more compact and more readable than XML.  And
 252 unlike XML, it is type-aware, so it is not necessary to do string parsing in
 253 application code.")
 254     (license license:lgpl2.1+)))
 255
 256 (define-public pfff
 257   (package
 258     (name "pfff")
 259     (version "1.0")
 260     (source (origin
 261               (method url-fetch)
 262               (uri (string-append "https://github.com/pfff/pfff/archive/v"
 263                                   version ".tar.gz"))
 264               (file-name (string-append name "-" version ".tar.gz"))
 265               (sha256
 266                (base32
 267                 "00m553aa277iarxj6dalmklyb64r7ias49bfwzbacsfg8h3kar8m"))))
 268     (build-system cmake-build-system)
 269     (home-page "http://biit.cs.ut.ee/pfff/")
 270     (synopsis "Probabilistic fast file fingerprinting tool")
 271     (description
 272      "pfff is a tool for calculating a compact digital fingerprint of a file
 273 by sampling randomly from the file instead of reading it in full.
 274 Consequently, the computation has a flat performance characteristic,
 275 correlated with data variation rather than file size.  pfff can be as reliable
 276 as existing hashing techniques, with provably negligible risk of collisions.")
 277     (license license:bsd-3)))
 278
 279 (define-public oniguruma
 280   (package
 281     (name "oniguruma")
 282     (version "5.9.6")
 283     (source (origin
 284               (method url-fetch)
 285               (uri (string-append "https://github.com/kkos/"
 286                                   "oniguruma/releases/download/v" version
 287                                   "/onig-" version ".tar.gz"))
 288               (sha256
 289                (base32
 290                 "19s79vsclqn170mw0ajwv7j37qsbn4f1yjz3yavnhvva6c820r6m"))))
 291     (build-system gnu-build-system)
 292     (home-page "https://github.com/kkos/oniguruma")
 293     (synopsis "Regular expression library")
 294     (description "Oniguruma is a regular expressions library.  The special
 295 characteristic of this library is that different character encoding for every
 296 regular expression object can be specified.")
 297     (license license:bsd-2)))
 298
 299 (define-public antiword
 300   (package
 301     (name "antiword")
 302     (version "0.37")
 303     (source (origin
 304               (method url-fetch)
 305               (uri (string-append "http://www.winfield.demon.nl/linux"
 306                                   "/antiword-" version ".tar.gz"))
 307               (sha256
 308                (base32
 309                 "1b7mi1l20jhj09kyh0bq14qzz8vdhhyf35gzwsq43mn6rc7h0b4f"))))
 310     (build-system gnu-build-system)
 311     (arguments
 312      `(#:tests? #f ; There are no tests
 313        #:make-flags
 314        (list "-f" "Makefile.Linux"
 315              (string-append "GLOBAL_INSTALL_DIR="
 316                             (assoc-ref %outputs "out") "/bin")
 317              (string-append "GLOBAL_RESOURCES_DIR="
 318                             (assoc-ref %outputs "out") "/share/antiword"))
 319        #:phases
 320        (modify-phases %standard-phases
 321          (delete 'configure)
 322          (replace 'install
 323            (lambda* (#:key make-flags #:allow-other-keys)
 324              (zero? (apply system* "make" `("global_install" ,@make-flags))))))))
 325     (home-page "http://www.winfield.demon.nl/")
 326     (synopsis "Microsoft Word document reader")
 327     (description "Antiword is an application for displaying Microsoft Word
 328 documents.  It can also convert the document to PostScript or XML.  Only
 329 documents made by MS Word version 2 and version 6 or later are supported.  The
 330 name comes from: \"The antidote against people who send Microsoft Word files
 331 to everybody, because they believe that everybody runs Windows and therefore
 332 runs Word\".")
 333     (license license:gpl2+)))
 334
 335 (define-public utfcpp
 336   (package
 337     (name "utfcpp")
 338     (version "2.3.4")
 339     (source (origin
 340               (method url-fetch)
 341               (uri
 342                (string-append
 343                 "mirror://sourceforge/project/utfcpp/utf8cpp_2x/Release%20"
 344                 version "/utf8_v"
 345                 (string-map (lambda (x) (if (eq? x #\.) #\_ x)) version)
 346                 ".zip"))
 347               (file-name (string-append name "-" version ".zip"))
 348               (sha256
 349                (base32
 350                 "1vqhs0aipcvvdrwcs7h3jsryg6mgbmc4s34n5cm6d36q4nxwwwrk"))))
 351     (build-system trivial-build-system)
 352     (arguments
 353      `(#:modules ((guix build utils))
 354        #:builder
 355        (begin
 356          (use-modules (guix build utils))
 357          (let ((source (assoc-ref %build-inputs "source"))
 358                (out    (assoc-ref %outputs "out"))
 359                (unzip  (string-append (assoc-ref %build-inputs "unzip")
 360                                       "/bin/unzip")))
 361            (mkdir-p out)
 362            (with-directory-excursion out
 363              (system* unzip source)
 364              (mkdir-p "share/doc")
 365              (rename-file "doc" "share/doc/utfcpp")
 366              (rename-file "source" "include"))))))
 367     (native-inputs `(("unzip" ,unzip)))
 368     (home-page "https://github.com/nemtrif/utfcpp")
 369     (synopsis "Portable C++ library for handling UTF-8")
 370     (description "UTF8-CPP is a C++ library for handling UTF-8 encoded text
 371 in a portable way.")
 372     (license license:boost1.0)))