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 "1.4.7")
  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               (patches (search-patches "xapian-revert-5489fb2f8.patch"))
  56               (sha256
  57                (base32 "1lxmlds3v5s1gng9nk1rvmln1zcksrw5ds509y0glylwch5qmw0k"))))
  58     (build-system gnu-build-system)
  59     (inputs `(("zlib" ,zlib)
  60               ("util-linux" ,util-linux)))
  61     (arguments
  62      `(#:phases
  63        (modify-phases %standard-phases
  64          (replace 'check
  65            ;; As of Xapian 1.3.3, the TCP server implementation uses
  66            ;; getaddrinfo(). This does not work in the build environment,
  67            ;; so exclude those tests. See HACKING for the list of targets.
  68            (lambda _
  69              (invoke "make"
  70                      "check-inmemory"
  71                      "check-remoteprog"
  72                      ;"check-remotetcp"
  73                      "check-multi"
  74                      "check-glass"
  75                      "check-chert"))))))
  76     (synopsis "Search Engine Library")
  77     (description
  78      "Xapian is a highly adaptable toolkit which allows developers to easily
  79 add advanced indexing and search facilities to their own applications.  It
  80 supports the Probabilistic Information Retrieval model and also supports a
  81 rich set of boolean query operators.")
  82     (home-page "https://xapian.org/")
  83     (license (list gpl2+ bsd-3 x11))))
  84
  85 (define-public python-xapian-bindings
  86   (package (inherit xapian)
  87     (name "python-xapian-bindings")
  88     (version (package-version xapian))
  89     (source (origin
  90               (method url-fetch)
  91               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  92                                   "/xapian-bindings-" version ".tar.xz"))
  93               (sha256
  94                (base32
  95                 "0sjf9ck3a6p7xnd84w09l6s0xn2g03k9a9417f4mjnywfq9pa6a5"))))
  96     (build-system gnu-build-system)
  97     (arguments
  98      `(#:configure-flags '("--with-python3")
  99        #:make-flags
 100        (list (string-append "pkgpylibdir="
 101                             (assoc-ref %outputs "out")
 102                             "/lib/python" ,(version-major+minor
 103                                             (package-version python))
 104                             "/site-packages/xapian"))))
 105     (inputs
 106      `(("python" ,python)
 107        ("python-sphinx" ,python-sphinx) ; for documentation
 108        ("xapian" ,xapian)
 109        ("zlib" ,zlib)))
 110     (synopsis "Python bindings for the Xapian search engine library")
 111     (license gpl2+)))
 112
 113 (define-public libtocc
 114   (package
 115     (name "libtocc")
 116     (version "1.0.1")
 117     (source
 118      (origin
 119        (method url-fetch)
 120        (uri (string-append "https://github.com/aidin36/tocc/releases/download/"
 121                            "v" version "/tocc-" version ".tar.gz"))
 122        (sha256
 123         (base32
 124          "1kd2jd74m8ksc8s7hh0haz0q0c3n0mr39bbky262kk4l58f1g068"))))
 125     (build-system gnu-build-system)
 126     (native-inputs `(("catch" ,catch-framework)))
 127     (inputs `(("unqlite" ,unqlite)))
 128     (arguments
 129      `(#:phases (modify-phases %standard-phases
 130                   (add-before 'configure 'chdir-source
 131                     (lambda _
 132                       (chdir "libtocc/src")
 133                       #t))
 134                   (replace 'check
 135                     (lambda _
 136                       (with-directory-excursion "../tests"
 137                         (invoke "./configure"
 138                                 (string-append "CONFIG_SHELL="
 139                                                (which "sh"))
 140                                 (string-append "SHELL="
 141                                                (which "sh"))
 142                                 "CPPFLAGS=-I../src"
 143                                 (string-append
 144                                  "LDFLAGS=-L../src/.libs "
 145                                  "-Wl,-rpath=../src/.libs"))
 146                         (invoke "make")
 147                         (invoke "./libtocctests")))))))
 148     (home-page "https://t-o-c-c.com/")
 149     (synopsis "Tool for Obsessive Compulsive Classifiers")
 150     (description
 151      "libtocc is the engine of the Tocc project, a tag-based file management
 152 system.  The goal of Tocc is to provide a better system for classifying files
 153 that is more flexible than classic file systems that are based on a tree of
 154 files and directories.")
 155     (license gpl3+)))
 156
 157 (define-public tocc
 158   (package
 159     (name "tocc")
 160     (version (package-version libtocc))
 161     (source (package-source libtocc))
 162     (build-system gnu-build-system)
 163     (inputs
 164      `(("libtocc" ,libtocc)
 165        ("unqlite" ,unqlite)))
 166     (arguments
 167      `(#:tests? #f                      ;No tests
 168        #:phases (modify-phases %standard-phases
 169                   (add-after
 170                    'unpack 'chdir-source
 171                    (lambda _ (chdir "cli/src"))))))
 172     (home-page "https://t-o-c-c.com/")
 173     (synopsis "Command-line interface to libtocc")
 174     (description
 175      "Tocc is a tag-based file management system.  This package contains the
 176 command line tool for interacting with libtocc.")
 177     (license gpl3+)))
 178
 179 (define-public bool
 180   (package
 181     (name "bool")
 182     (version "0.2.2")
 183     (source
 184      (origin
 185        (method url-fetch)
 186        (uri (string-append "mirror://gnu/bool/bool-"
 187                            version ".tar.xz"))
 188        (sha256
 189         (base32
 190          "1frdmgrmb509fxbdpsxxw3lvvwv7xm1pavqrqgm4jg698iix6xfw"))))
 191     (build-system gnu-build-system)
 192     (home-page "https://www.gnu.org/software/bool/")
 193     (synopsis "Finding text and HTML files that match boolean expressions")
 194     (description
 195      "GNU Bool is a utility to perform text searches on files using Boolean
 196 expressions.  For example, a search for \"hello AND world\" would return a
 197 file containing the phrase \"Hello, world!\".  It supports both AND and OR
 198 statements, as well as the NEAR statement to search for the occurrence of
 199 words in close proximity to each other.  It handles context gracefully,
 200 accounting for new lines and paragraph changes.  It also has robust support
 201 for parsing HTML files.")
 202     (license gpl3+)))
 203
 204 (define-public hyperestraier
 205   (package
 206     (name "hyperestraier")
 207     (version "1.4.13")
 208     (source
 209       (origin
 210         (method url-fetch)
 211         (uri (string-append "http://fallabs.com/" name "/"
 212                             name "-" version ".tar.gz"))
 213         (sha256
 214          (base32
 215           "1qk3pxgzyrpcz5qfyd5xs2hw9q1cbb7j5zd4kp1diq501wcj2vs9"))))
 216     (inputs
 217      `(("qdbm" ,qdbm)
 218        ("zlib" ,zlib)))
 219     (build-system gnu-build-system)
 220     (arguments
 221      `(#:configure-flags (list (string-append "LDFLAGS=-Wl,-rpath="
 222                                               (assoc-ref %outputs "out")
 223                                               "/lib"))))
 224     (home-page "http://fallabs.com/hyperestraier")
 225     (synopsis "Full-text search system")
 226     (description "Hyper Estraier can be used to integrate full-text
 227 search into applications, using either the provided command line and CGI
 228 interfaces, or a C API.")
 229     (license lgpl2.1+)))
 230
 231 (define-public mlocate
 232   (package
 233     (name "mlocate")
 234     (version "0.26")
 235     (source (origin
 236               (method url-fetch)
 237               (uri (string-append "http://releases.pagure.org/mlocate/"
 238                                   "mlocate-" version ".tar.xz"))
 239               (sha256
 240                (base32
 241                 "0gi6y52gkakhhlnzy0p6izc36nqhyfx5830qirhvk3qrzrwxyqrh"))))
 242     (build-system gnu-build-system)
 243     (home-page "https://pagure.io/mlocate")
 244     (synopsis "Locate files on the file system")
 245     (description
 246      "mlocate is a locate/updatedb implementation.  The 'm' stands for
 247 \"merging\": @code{updatedb} reuses the existing database to avoid rereading
 248 most of the file system, which makes it faster and does not trash the system
 249 caches as much.  The locate(1) utility is intended to be completely compatible
 250 with slocate, and attempts to be compatible to GNU locate when it does not
 251 conflict with slocate compatibility.")
 252     (license gpl2)))
 253
 254 (define-public swish-e
 255   (package
 256     (name "swish-e")
 257     (version "2.4.7")
 258     (source (origin
 259               (method url-fetch)
 260               (uri (list (string-append
 261                           "https://web.archive.org/web/20160730145202/"
 262                           "http://swish-e.org/distribution/"
 263                           "swish-e-" version ".tar.gz")
 264                          (string-append "http://http.debian.net/debian/pool/"
 265                                         "main/s/swish-e/swish-e_" version
 266                                         ".orig.tar.gz")))
 267               (file-name (string-append name "-" version ".tar.gz"))
 268               (sha256
 269                (base32
 270                 "0qkrk7z25yp9hynj21vxkyn7yi8gcagcfxnass5cgczcz0gm9pax"))
 271               (patches (search-patches "swish-e-search.patch"
 272                                        "swish-e-format-security.patch"))))
 273     (build-system gnu-build-system)
 274     ;; Several other packages and perl modules may be installed alongside
 275     ;; swish-e to extend its features at runtime, but are not required for
 276     ;; building: xpdf, catdoc, MP3::Tag, Spreadsheet::ParseExcel,
 277     ;; HTML::Entities.
 278     (inputs
 279      `(("perl" ,perl)
 280        ("perl-uri" ,perl-uri)
 281        ("perl-html-parser" ,perl-html-parser)
 282        ("perl-html-tagset" ,perl-html-tagset)
 283        ("perl-mime-types" ,perl-mime-types)))
 284     (arguments
 285      `(;; XXX: This fails to build with zlib (API mismatch) and tests fail
 286        ;; with libxml2, so disable both.
 287        #:configure-flags (list (string-append "--without-zlib")
 288                                (string-append "--without-libxml2"))
 289        #:phases (modify-phases %standard-phases
 290                   (add-after 'install 'wrap-programs
 291                     (lambda* (#:key inputs outputs #:allow-other-keys)
 292                       (let* ((out (assoc-ref outputs "out")))
 293                         (for-each
 294                          (lambda (program)
 295                            (wrap-program program
 296                              `("PERL5LIB" ":" prefix
 297                                ,(map (lambda (i)
 298                                        (string-append (assoc-ref inputs i)
 299                                                       "/lib/perl5/site_perl"))
 300                                      ;; These perl modules have no propagated
 301                                      ;; inputs, so no further analysis needed.
 302                                      '("perl-uri"
 303                                        "perl-html-parser"
 304                                        "perl-html-tagset"
 305                                        "perl-mime-types")))))
 306                          (list (string-append out "/lib/swish-e/swishspider")
 307                                (string-append out "/bin/swish-filter-test")))
 308                         #t))))))
 309     (home-page (string-append "https://web.archive.org/web/20160730145202/"
 310                               "http://swish-e.org"))
 311     (synopsis "Web indexing system")
 312     (description
 313      "Swish-e is Simple Web Indexing System for Humans - Enhanced.  Swish-e
 314 can quickly and easily index directories of files or remote web sites and
 315 search the generated indexes.")
 316     (license gpl2+)))                   ;with exception
 317
 318 (define-public xapers
 319   (package
 320     (name "xapers")
 321     (version "0.8.2")
 322     (source
 323      (origin
 324        (method url-fetch)
 325        (uri (string-append
 326              "https://finestructure.net/xapers/releases/xapers-"
 327              version ".tar.gz"))
 328        (sha256
 329         (base32
 330          "0ykz6hn3qj46w3c99d6q0pi5ncq2894simcl7vapv047zm3cylmd"))))
 331     (build-system python-build-system)
 332     (propagated-inputs
 333      `(("poppler" ,poppler)
 334        ("python-urwid" ,python-urwid)
 335        ("xclip" ,xclip)
 336        ("xdg-utils" ,xdg-utils)))
 337     (inputs
 338      `(("python-latexcodec" ,python-latexcodec)
 339        ("python-pybtex" ,python-pybtex)
 340        ("python-pycurl" ,python-pycurl)
 341        ("python-pyyaml" ,python-pyyaml)
 342        ("python-six" ,python-six)
 343        ("python-xapian-bindings" ,python-xapian-bindings)))
 344     (arguments
 345      `(#:modules ((ice-9 rdelim)
 346                   (guix build python-build-system)
 347                   (guix build utils))
 348        #:phases
 349        (modify-phases %standard-phases
 350          (add-after 'install 'install-doc
 351            (lambda* (#:key inputs outputs #:allow-other-keys)
 352              (define (purge-term-support input output)
 353                (let loop ((line (read-line input)))
 354                  (if (string-prefix? "if [[ \"$term\"" line)
 355                      (begin (display "eval \"$cmd\"\n" output)
 356                             #t)
 357                      (begin (display (string-append line "\n") output)
 358                             (loop (read-line input))))))
 359              (let* ((out (assoc-ref outputs "out"))
 360                     (bin (string-append out "/bin"))
 361                     (adder-out (string-append bin "/xapers-adder"))
 362                     (man1 (string-append out "/share/man/man1")))
 363                (install-file "man/man1/xapers.1"  man1)
 364                (install-file "man/man1/xapers-adder.1" man1)
 365                ;; below is equivalent to setting --no-term option
 366                ;; permanently on; this is desirable to avoid imposing
 367                ;; an x-terminal installation on the user but breaks
 368                ;; some potential xapers-adder uses like auto browser
 369                ;; pdf handler, but user could instead still use
 370                ;; e.g. "xterm -e xapers-adder %F" for same use.
 371                ;; alternatively we could propagate xterm as an input
 372                ;; and replace 'x-terminal-emulator' with 'xterm'
 373                (call-with-input-file "bin/xapers-adder"
 374                  (lambda (input)
 375                    (call-with-output-file adder-out
 376                      (lambda (output)
 377                        (purge-term-support input output)))))
 378                (chmod adder-out #o555)))))))
 379     (home-page "https://finestructure.net/xapers/")
 380     (synopsis "Personal document indexing system")
 381     (description
 382      "Xapers is a personal document indexing system,
 383 geared towards academic journal articles build on the Xapian search engine.
 384 Think of it as your own personal document search engine, or a local cache of
 385 online libraries.  It provides fast search of document text and
 386 bibliographic data and simple document and bibtex retrieval.")
 387     (license gpl3+)))
 388
 389 ;;; search.scm ends here