gnu/packages/search.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2014 Mark H Weaver <mhw@netris.org>
   3 ;;; Copyright © 2015, 2016 Eric Bavier <bavier@member.fsf.org>
   4 ;;; Copyright © 2017 Thomas Danckaert <post@thomasdanckaert.be>
   5 ;;; Copyright © 2017 Ricardo Wurmus <rekado@elephly.net>
   6 ;;; Copyright © 2018 Tobias Geerinckx-Rice <me@tobias.gr>
   7 ;;; Copyright © 2018 Adam Massmann <massmannak@gmail.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 search)
  25   #:use-module ((guix licenses)
  26                 #:select (gpl2 gpl2+ gpl3+ lgpl2.1+ bsd-3 x11))
  27   #:use-module (guix packages)
  28   #:use-module (guix download)
  29   #:use-module (guix utils)
  30   #:use-module (guix build-system gnu)
  31   #:use-module (guix build-system python)
  32   #:use-module (gnu packages)
  33   #:use-module (gnu packages compression)
  34   #:use-module (gnu packages check)
  35   #:use-module (gnu packages databases)
  36   #:use-module (gnu packages freedesktop)
  37   #:use-module (gnu packages linux)
  38   #:use-module (gnu packages perl)
  39   #:use-module (gnu packages pdf)
  40   #:use-module (gnu packages python)
  41   #:use-module (gnu packages python-web)
  42   #:use-module (gnu packages web)
  43   #:use-module (gnu packages xdisorg)
  44   #:use-module (gnu packages xml))
  45
  46 (define-public xapian
  47   (package
  48     (name "xapian")
  49     (version "1.4.5")
  50     ;; Note: When updating Xapian, remember to update xapian-bindings below.
  51     (source (origin
  52               (method url-fetch)
  53               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  54                                   "/xapian-core-" version ".tar.xz"))
  55               (sha256
  56                (base32 "0axhqrj202hbll9mcx1qdm8gsqj19216w3z02gyjbycxvr9gkdc5"))))
  57     (build-system gnu-build-system)
  58     (inputs `(("zlib" ,zlib)
  59               ("util-linux" ,util-linux)))
  60     (arguments
  61      `(#:phases
  62        (modify-phases %standard-phases
  63          (replace 'check
  64            ;; As of Xapian 1.3.3, the TCP server implementation uses
  65            ;; getaddrinfo(). This does not work in the build environment,
  66            ;; so exclude those tests. See HACKING for the list of targets.
  67            (lambda _
  68              (invoke "make"
  69                      "check-inmemory"
  70                      "check-remoteprog"
  71                      ;"check-remotetcp"
  72                      "check-multi"
  73                      "check-glass"
  74                      "check-chert"))))))
  75     (synopsis "Search Engine Library")
  76     (description
  77      "Xapian is a highly adaptable toolkit which allows developers to easily
  78 add advanced indexing and search facilities to their own applications.  It
  79 supports the Probabilistic Information Retrieval model and also supports a
  80 rich set of boolean query operators.")
  81     (home-page "https://xapian.org/")
  82     (license (list gpl2+ bsd-3 x11))))
  83
  84 (define-public python-xapian-bindings
  85   (package (inherit xapian)
  86     (name "python-xapian-bindings")
  87     (version (package-version xapian))
  88     (source (origin
  89               (method url-fetch)
  90               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  91                                   "/xapian-bindings-" version ".tar.xz"))
  92               (sha256
  93                (base32
  94                 "0cwx39764w24xd25w271had4w78lnw1dgz36yvlw1g3i19rqcy34"))))
  95     (build-system gnu-build-system)
  96     (arguments
  97      `(#:configure-flags '("--with-python3")
  98        #:make-flags
  99        (list (string-append "pkgpylibdir="
 100                             (assoc-ref %outputs "out")
 101                             "/lib/python" ,(version-major+minor
 102                                             (package-version python))
 103                             "/site-packages/xapian"))))
 104     (inputs
 105      `(("python" ,python)
 106        ("python-sphinx" ,python-sphinx) ; for documentation
 107        ("xapian" ,xapian)
 108        ("zlib" ,zlib)))
 109     (synopsis "Python bindings for the Xapian search engine library")
 110     (license gpl2+)))
 111
 112 (define-public libtocc
 113   (package
 114     (name "libtocc")
 115     (version "1.0.1")
 116     (source
 117      (origin
 118        (method url-fetch)
 119        (uri (string-append "https://github.com/aidin36/tocc/releases/download/"
 120                            "v" version "/tocc-" version ".tar.gz"))
 121        (sha256
 122         (base32
 123          "1kd2jd74m8ksc8s7hh0haz0q0c3n0mr39bbky262kk4l58f1g068"))))
 124     (build-system gnu-build-system)
 125     (native-inputs `(("catch" ,catch-framework)))
 126     (inputs `(("unqlite" ,unqlite)))
 127     (arguments
 128      `(#:phases (modify-phases %standard-phases
 129                   (add-before
 130                    'configure 'chdir-source
 131                    (lambda _ (chdir "libtocc/src")))
 132                   (replace
 133                    'check
 134                    (lambda _
 135                      (with-directory-excursion "../tests"
 136                        (and (zero? (system* "./configure"
 137                                             (string-append "CONFIG_SHELL="
 138                                                            (which "sh"))
 139                                             (string-append "SHELL="
 140                                                            (which "sh"))
 141                                             "CPPFLAGS=-I../src"
 142                                             (string-append
 143                                              "LDFLAGS=-L../src/.libs "
 144                                              "-Wl,-rpath=../src/.libs")))
 145                             (zero? (system* "make"))
 146                             (zero? (system* "./libtocctests")))))))))
 147     (home-page "https://t-o-c-c.com/")
 148     (synopsis "Tool for Obsessive Compulsive Classifiers")
 149     (description
 150      "libtocc is the engine of the Tocc project, a tag-based file management
 151 system.  The goal of Tocc is to provide a better system for classifying files
 152 that is more flexible than classic file systems that are based on a tree of
 153 files and directories.")
 154     (license gpl3+)))
 155
 156 (define-public tocc
 157   (package
 158     (name "tocc")
 159     (version (package-version libtocc))
 160     (source (package-source libtocc))
 161     (build-system gnu-build-system)
 162     (inputs
 163      `(("libtocc" ,libtocc)
 164        ("unqlite" ,unqlite)))
 165     (arguments
 166      `(#:tests? #f                      ;No tests
 167        #:phases (modify-phases %standard-phases
 168                   (add-after
 169                    'unpack 'chdir-source
 170                    (lambda _ (chdir "cli/src"))))))
 171     (home-page "https://t-o-c-c.com/")
 172     (synopsis "Command-line interface to libtocc")
 173     (description
 174      "Tocc is a tag-based file management system.  This package contains the
 175 command line tool for interacting with libtocc.")
 176     (license gpl3+)))
 177
 178 (define-public bool
 179   (package
 180     (name "bool")
 181     (version "0.2.2")
 182     (source
 183      (origin
 184        (method url-fetch)
 185        (uri (string-append "mirror://gnu/bool/bool-"
 186                            version ".tar.xz"))
 187        (sha256
 188         (base32
 189          "1frdmgrmb509fxbdpsxxw3lvvwv7xm1pavqrqgm4jg698iix6xfw"))))
 190     (build-system gnu-build-system)
 191     (home-page "https://www.gnu.org/software/bool/")
 192     (synopsis "Finding text and HTML files that match boolean expressions")
 193     (description
 194      "GNU Bool is a utility to perform text searches on files using Boolean
 195 expressions.  For example, a search for \"hello AND world\" would return a
 196 file containing the phrase \"Hello, world!\".  It supports both AND and OR
 197 statements, as well as the NEAR statement to search for the occurrence of
 198 words in close proximity to each other.  It handles context gracefully,
 199 accounting for new lines and paragraph changes.  It also has robust support
 200 for parsing HTML files.")
 201     (license gpl3+)))
 202
 203 (define-public hyperestraier
 204   (package
 205     (name "hyperestraier")
 206     (version "1.4.13")
 207     (source
 208       (origin
 209         (method url-fetch)
 210         (uri (string-append "http://fallabs.com/" name "/"
 211                             name "-" version ".tar.gz"))
 212         (sha256
 213          (base32
 214           "1qk3pxgzyrpcz5qfyd5xs2hw9q1cbb7j5zd4kp1diq501wcj2vs9"))))
 215     (inputs
 216      `(("qdbm" ,qdbm)
 217        ("zlib" ,zlib)))
 218     (build-system gnu-build-system)
 219     (arguments
 220      `(#:configure-flags (list (string-append "LDFLAGS=-Wl,-rpath="
 221                                               (assoc-ref %outputs "out")
 222                                               "/lib"))))
 223     (home-page "http://fallabs.com/hyperestraier")
 224     (synopsis "Full-text search system")
 225     (description "Hyper Estraier can be used to integrate full-text
 226 search into applications, using either the provided command line and CGI
 227 interfaces, or a C API.")
 228     (license lgpl2.1+)))
 229
 230 (define-public mlocate
 231   (package
 232     (name "mlocate")
 233     (version "0.26")
 234     (source (origin
 235               (method url-fetch)
 236               (uri (string-append "http://releases.pagure.org/mlocate/"
 237                                   "mlocate-" version ".tar.xz"))
 238               (sha256
 239                (base32
 240                 "0gi6y52gkakhhlnzy0p6izc36nqhyfx5830qirhvk3qrzrwxyqrh"))))
 241     (build-system gnu-build-system)
 242     (home-page "https://pagure.io/mlocate")
 243     (synopsis "Locate files on the file system")
 244     (description
 245      "mlocate is a locate/updatedb implementation.  The 'm' stands for
 246 \"merging\": @code{updatedb} reuses the existing database to avoid rereading
 247 most of the file system, which makes it faster and does not trash the system
 248 caches as much.  The locate(1) utility is intended to be completely compatible
 249 with slocate, and attempts to be compatible to GNU locate when it does not
 250 conflict with slocate compatibility.")
 251     (license gpl2)))
 252
 253 (define-public swish-e
 254   (package
 255     (name "swish-e")
 256     (version "2.4.7")
 257     (source (origin
 258               (method url-fetch)
 259               (uri (list (string-append
 260                           "https://web.archive.org/web/20160730145202/"
 261                           "http://swish-e.org/distribution/"
 262                           "swish-e-" version ".tar.gz")
 263                          (string-append "http://http.debian.net/debian/pool/"
 264                                         "main/s/swish-e/swish-e_" version
 265                                         ".orig.tar.gz")))
 266               (file-name (string-append name "-" version ".tar.gz"))
 267               (sha256
 268                (base32
 269                 "0qkrk7z25yp9hynj21vxkyn7yi8gcagcfxnass5cgczcz0gm9pax"))
 270               (patches (search-patches "swish-e-search.patch"
 271                                        "swish-e-format-security.patch"))))
 272     (build-system gnu-build-system)
 273     ;; Several other packages and perl modules may be installed alongside
 274     ;; swish-e to extend its features at runtime, but are not required for
 275     ;; building: xpdf, catdoc, MP3::Tag, Spreadsheet::ParseExcel,
 276     ;; HTML::Entities.
 277     (inputs
 278      `(("perl" ,perl)
 279        ("perl-uri" ,perl-uri)
 280        ("perl-html-parser" ,perl-html-parser)
 281        ("perl-html-tagset" ,perl-html-tagset)
 282        ("perl-mime-types" ,perl-mime-types)))
 283     (arguments
 284      `(;; XXX: This fails to build with zlib (API mismatch) and tests fail
 285        ;; with libxml2, so disable both.
 286        #:configure-flags (list (string-append "--without-zlib")
 287                                (string-append "--without-libxml2"))
 288        #:phases (modify-phases %standard-phases
 289                   (add-after 'install 'wrap-programs
 290                     (lambda* (#:key inputs outputs #:allow-other-keys)
 291                       (let* ((out (assoc-ref outputs "out")))
 292                         (for-each
 293                          (lambda (program)
 294                            (wrap-program program
 295                              `("PERL5LIB" ":" prefix
 296                                ,(map (lambda (i)
 297                                        (string-append (assoc-ref inputs i)
 298                                                       "/lib/perl5/site_perl"))
 299                                      ;; These perl modules have no propagated
 300                                      ;; inputs, so no further analysis needed.
 301                                      '("perl-uri"
 302                                        "perl-html-parser"
 303                                        "perl-html-tagset"
 304                                        "perl-mime-types")))))
 305                          (list (string-append out "/lib/swish-e/swishspider")
 306                                (string-append out "/bin/swish-filter-test")))
 307                         #t))))))
 308     (home-page (string-append "https://web.archive.org/web/20160730145202/"
 309                               "http://swish-e.org"))
 310     (synopsis "Web indexing system")
 311     (description
 312      "Swish-e is Simple Web Indexing System for Humans - Enhanced.  Swish-e
 313 can quickly and easily index directories of files or remote web sites and
 314 search the generated indexes.")
 315     (license gpl2+)))                   ;with exception
 316
 317 (define-public xapers
 318   (package
 319     (name "xapers")
 320     (version "0.8.2")
 321     (source
 322      (origin
 323        (method url-fetch)
 324        (uri (string-append
 325              "https://finestructure.net/xapers/releases/xapers-"
 326              version ".tar.gz"))
 327        (sha256
 328         (base32
 329          "0ykz6hn3qj46w3c99d6q0pi5ncq2894simcl7vapv047zm3cylmd"))))
 330     (build-system python-build-system)
 331     (propagated-inputs
 332      `(("poppler" ,poppler)
 333        ("python-urwid" ,python-urwid)
 334        ("xclip" ,xclip)
 335        ("xdg-utils" ,xdg-utils)))
 336     (inputs
 337      `(("python-latexcodec" ,python-latexcodec)
 338        ("python-pybtex" ,python-pybtex)
 339        ("python-pycurl" ,python-pycurl)
 340        ("python-pyyaml" ,python-pyyaml)
 341        ("python-six" ,python-six)
 342        ("python-xapian-bindings" ,python-xapian-bindings)))
 343     (arguments
 344      `(#:modules ((ice-9 rdelim)
 345                   (guix build python-build-system)
 346                   (guix build utils))
 347        #:phases
 348        (modify-phases %standard-phases
 349          (add-after 'install 'install-doc
 350            (lambda* (#:key inputs outputs #:allow-other-keys)
 351              (define (purge-term-support input output)
 352                (let loop ((line (read-line input)))
 353                  (if (string-prefix? "if [[ \"$term\"" line)
 354                      (begin (display "eval \"$cmd\"\n" output)
 355                             #t)
 356                      (begin (display (string-append line "\n") output)
 357                             (loop (read-line input))))))
 358              (let* ((out (assoc-ref outputs "out"))
 359                     (bin (string-append out "/bin"))
 360                     (adder-out (string-append bin "/xapers-adder"))
 361                     (man1 (string-append out "/share/man/man1")))
 362                (install-file "man/man1/xapers.1"  man1)
 363                (install-file "man/man1/xapers-adder.1" man1)
 364                ;; below is equivalent to setting --no-term option
 365                ;; permanently on; this is desirable to avoid imposing
 366                ;; an x-terminal installation on the user but breaks
 367                ;; some potential xapers-adder uses like auto browser
 368                ;; pdf handler, but user could instead still use
 369                ;; e.g. "xterm -e xapers-adder %F" for same use.
 370                ;; alternatively we could propagate xterm as an input
 371                ;; and replace 'x-terminal-emulator' with 'xterm'
 372                (call-with-input-file "bin/xapers-adder"
 373                  (lambda (input)
 374                    (call-with-output-file adder-out
 375                      (lambda (output)
 376                        (purge-term-support input output)))))
 377                (chmod adder-out #o555)))))))
 378     (home-page "https://finestructure.net/xapers/")
 379     (synopsis "Personal document indexing system")
 380     (description
 381      "Xapers is a personal document indexing system,
 382 geared towards academic journal articles build on the Xapian search engine.
 383 Think of it as your own personal document search engine, or a local cache of
 384 online libraries.  It provides fast search of document text and
 385 bibliographic data and simple document and bibtex retrieval.")
 386     (license gpl3+)))
 387
 388 ;;; search.scm ends here