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 perl-license))
  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 perl)
  32   #:use-module (guix build-system python)
  33   #:use-module (gnu packages)
  34   #:use-module (gnu packages compression)
  35   #:use-module (gnu packages check)
  36   #:use-module (gnu packages databases)
  37   #:use-module (gnu packages freedesktop)
  38   #:use-module (gnu packages linux)
  39   #:use-module (gnu packages perl)
  40   #:use-module (gnu packages pdf)
  41   #:use-module (gnu packages python)
  42   #:use-module (gnu packages python-web)
  43   #:use-module (gnu packages python-xyz)
  44   #:use-module (gnu packages sphinx)
  45   #:use-module (gnu packages web)
  46   #:use-module (gnu packages xdisorg)
  47   #:use-module (gnu packages xml))
  48
  49 (define-public xapian
  50   (package
  51     (name "xapian")
  52     (version "1.4.11")
  53     ;; Note: When updating Xapian, remember to update xapian-bindings below.
  54     (source (origin
  55               (method url-fetch)
  56               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  57                                   "/xapian-core-" version ".tar.xz"))
  58               (sha256
  59                (base32 "01xwqljnp5afjf9097lyfbqc6x5bcqszfdkn9l1j86imwbrv45lz"))))
  60     (build-system gnu-build-system)
  61     (inputs `(("zlib" ,zlib)
  62               ("util-linux" ,util-linux)))
  63     (arguments
  64      `(#:phases
  65        (modify-phases %standard-phases
  66          (replace 'check
  67            ;; As of Xapian 1.3.3, the TCP server implementation uses
  68            ;; getaddrinfo(). This does not work in the build environment,
  69            ;; so exclude those tests. See HACKING for the list of targets.
  70            (lambda _
  71              (invoke "make"
  72                      "check-inmemory"
  73                      "check-remoteprog"
  74                      ;"check-remotetcp"
  75                      "check-multi"
  76                      "check-glass"
  77                      "check-chert"))))))
  78     (synopsis "Search Engine Library")
  79     (description
  80      "Xapian is a highly adaptable toolkit which allows developers to easily
  81 add advanced indexing and search facilities to their own applications.  It
  82 supports the Probabilistic Information Retrieval model and also supports a
  83 rich set of boolean query operators.")
  84     (home-page "https://xapian.org/")
  85     (license (list gpl2+ bsd-3 x11))))
  86
  87 (define-public python-xapian-bindings
  88   (package (inherit xapian)
  89     (name "python-xapian-bindings")
  90     (version (package-version xapian))
  91     (source (origin
  92               (method url-fetch)
  93               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  94                                   "/xapian-bindings-" version ".tar.xz"))
  95               (sha256
  96                (base32
  97                 "13bi2vr5d39ys49nlwmsv64ik5pdwkz28bh08hyylrhanb45d8wx"))))
  98     (build-system gnu-build-system)
  99     (arguments
 100      `(#:configure-flags '("--with-python3")
 101        #:make-flags
 102        (list (string-append "pkgpylibdir="
 103                             (assoc-ref %outputs "out")
 104                             "/lib/python" ,(version-major+minor
 105                                             (package-version python))
 106                             "/site-packages/xapian"))))
 107     (inputs
 108      `(("python" ,python)
 109        ("python-sphinx" ,python-sphinx) ; for documentation
 110        ("xapian" ,xapian)
 111        ("zlib" ,zlib)))
 112     (synopsis "Python bindings for the Xapian search engine library")
 113     (license gpl2+)))
 114
 115 (define-public perl-search-xapian
 116   (package
 117     (name "perl-search-xapian")
 118     (version "1.2.25.2")
 119     (source
 120      (origin
 121        (method url-fetch)
 122        (uri (string-append "mirror://cpan/authors/id/O/OL/OLLY/"
 123                            "Search-Xapian-" version ".tar.gz"))
 124        (sha256
 125         (base32
 126          "0hpa8gi38j0ibq8af6dy69lm1bl5jnq76nsa69dbrzbr88l5m594"))))
 127     (build-system perl-build-system)
 128     (native-inputs
 129      `(("perl-devel-leak" ,perl-devel-leak)))
 130     (inputs
 131      `(("xapian" ,xapian)))
 132     (home-page "https://metacpan.org/release/Search-Xapian")
 133     (synopsis "Perl XS frontend to the Xapian C++ search library")
 134     (description
 135      "Search::Xapian wraps most methods of most Xapian classes.  The missing
 136 classes and methods should be added in the future.  It also provides a
 137 simplified, more 'perlish' interface to some common operations.")
 138     (license perl-license)))
 139
 140 (define-public libtocc
 141   (package
 142     (name "libtocc")
 143     (version "1.0.1")
 144     (source
 145      (origin
 146        (method url-fetch)
 147        (uri (string-append "https://github.com/aidin36/tocc/releases/download/"
 148                            "v" version "/tocc-" version ".tar.gz"))
 149        (sha256
 150         (base32
 151          "1kd2jd74m8ksc8s7hh0haz0q0c3n0mr39bbky262kk4l58f1g068"))))
 152     (build-system gnu-build-system)
 153     (native-inputs `(("catch" ,catch-framework)))
 154     (inputs `(("unqlite" ,unqlite)))
 155     (arguments
 156      `(#:phases (modify-phases %standard-phases
 157                   (add-before 'configure 'chdir-source
 158                     (lambda _
 159                       (chdir "libtocc/src")
 160                       #t))
 161                   (replace 'check
 162                     (lambda _
 163                       (with-directory-excursion "../tests"
 164                         (invoke "./configure"
 165                                 (string-append "CONFIG_SHELL="
 166                                                (which "sh"))
 167                                 (string-append "SHELL="
 168                                                (which "sh"))
 169                                 "CPPFLAGS=-I../src"
 170                                 (string-append
 171                                  "LDFLAGS=-L../src/.libs "
 172                                  "-Wl,-rpath=../src/.libs"))
 173                         (invoke "make")
 174                         (invoke "./libtocctests")))))))
 175     (home-page "https://t-o-c-c.com/")
 176     (synopsis "Tool for Obsessive Compulsive Classifiers")
 177     (description
 178      "libtocc is the engine of the Tocc project, a tag-based file management
 179 system.  The goal of Tocc is to provide a better system for classifying files
 180 that is more flexible than classic file systems that are based on a tree of
 181 files and directories.")
 182     (license gpl3+)))
 183
 184 (define-public tocc
 185   (package
 186     (name "tocc")
 187     (version (package-version libtocc))
 188     (source (package-source libtocc))
 189     (build-system gnu-build-system)
 190     (inputs
 191      `(("libtocc" ,libtocc)
 192        ("unqlite" ,unqlite)))
 193     (arguments
 194      `(#:tests? #f                      ;No tests
 195        #:phases (modify-phases %standard-phases
 196                   (add-after
 197                    'unpack 'chdir-source
 198                    (lambda _ (chdir "cli/src"))))))
 199     (home-page "https://t-o-c-c.com/")
 200     (synopsis "Command-line interface to libtocc")
 201     (description
 202      "Tocc is a tag-based file management system.  This package contains the
 203 command line tool for interacting with libtocc.")
 204     (license gpl3+)))
 205
 206 (define-public bool
 207   (package
 208     (name "bool")
 209     (version "0.2.2")
 210     (source
 211      (origin
 212        (method url-fetch)
 213        (uri (string-append "mirror://gnu/bool/bool-"
 214                            version ".tar.xz"))
 215        (sha256
 216         (base32
 217          "1frdmgrmb509fxbdpsxxw3lvvwv7xm1pavqrqgm4jg698iix6xfw"))))
 218     (build-system gnu-build-system)
 219     (home-page "https://www.gnu.org/software/bool/")
 220     (synopsis "Finding text and HTML files that match boolean expressions")
 221     (description
 222      "GNU Bool is a utility to perform text searches on files using Boolean
 223 expressions.  For example, a search for \"hello AND world\" would return a
 224 file containing the phrase \"Hello, world!\".  It supports both AND and OR
 225 statements, as well as the NEAR statement to search for the occurrence of
 226 words in close proximity to each other.  It handles context gracefully,
 227 accounting for new lines and paragraph changes.  It also has robust support
 228 for parsing HTML files.")
 229     (license gpl3+)))
 230
 231 (define-public hyperestraier
 232   (package
 233     (name "hyperestraier")
 234     (version "1.4.13")
 235     (source
 236       (origin
 237         (method url-fetch)
 238         (uri (string-append "http://fallabs.com/" name "/"
 239                             name "-" version ".tar.gz"))
 240         (sha256
 241          (base32
 242           "1qk3pxgzyrpcz5qfyd5xs2hw9q1cbb7j5zd4kp1diq501wcj2vs9"))))
 243     (inputs
 244      `(("qdbm" ,qdbm)
 245        ("zlib" ,zlib)))
 246     (build-system gnu-build-system)
 247     (arguments
 248      `(#:configure-flags (list (string-append "LDFLAGS=-Wl,-rpath="
 249                                               (assoc-ref %outputs "out")
 250                                               "/lib"))))
 251     (home-page "http://fallabs.com/hyperestraier")
 252     (synopsis "Full-text search system")
 253     (description "Hyper Estraier can be used to integrate full-text
 254 search into applications, using either the provided command line and CGI
 255 interfaces, or a C API.")
 256     (license lgpl2.1+)))
 257
 258 (define-public mlocate
 259   (package
 260     (name "mlocate")
 261     (version "0.26")
 262     (source (origin
 263               (method url-fetch)
 264               (uri (string-append "http://releases.pagure.org/mlocate/"
 265                                   "mlocate-" version ".tar.xz"))
 266               (sha256
 267                (base32
 268                 "0gi6y52gkakhhlnzy0p6izc36nqhyfx5830qirhvk3qrzrwxyqrh"))))
 269     (build-system gnu-build-system)
 270     (home-page "https://pagure.io/mlocate")
 271     (synopsis "Locate files on the file system")
 272     (description
 273      "mlocate is a locate/updatedb implementation.  The @code{m} stands for
 274 \"merging\": @code{updatedb} reuses the existing database to avoid rereading
 275 most of the file system, which makes it faster and does not trash the system
 276 caches as much.  The locate(1) utility is intended to be completely compatible
 277 with slocate, and attempts to be compatible to GNU locate when it does not
 278 conflict with slocate compatibility.")
 279     (license gpl2)))
 280
 281 (define-public swish-e
 282   (package
 283     (name "swish-e")
 284     (version "2.4.7")
 285     (source (origin
 286               (method url-fetch)
 287               (uri (list (string-append
 288                           "https://web.archive.org/web/20160730145202/"
 289                           "http://swish-e.org/distribution/"
 290                           "swish-e-" version ".tar.gz")
 291                          (string-append "http://http.debian.net/debian/pool/"
 292                                         "main/s/swish-e/swish-e_" version
 293                                         ".orig.tar.gz")))
 294               (file-name (string-append name "-" version ".tar.gz"))
 295               (sha256
 296                (base32
 297                 "0qkrk7z25yp9hynj21vxkyn7yi8gcagcfxnass5cgczcz0gm9pax"))
 298               (patches (search-patches "swish-e-search.patch"
 299                                        "swish-e-format-security.patch"))))
 300     (build-system gnu-build-system)
 301     ;; Several other packages and perl modules may be installed alongside
 302     ;; swish-e to extend its features at runtime, but are not required for
 303     ;; building: xpdf, catdoc, MP3::Tag, Spreadsheet::ParseExcel,
 304     ;; HTML::Entities.
 305     (inputs
 306      `(("perl" ,perl)
 307        ("perl-uri" ,perl-uri)
 308        ("perl-html-parser" ,perl-html-parser)
 309        ("perl-html-tagset" ,perl-html-tagset)
 310        ("perl-mime-types" ,perl-mime-types)))
 311     (arguments
 312      `(;; XXX: This fails to build with zlib (API mismatch) and tests fail
 313        ;; with libxml2, so disable both.
 314        #:configure-flags (list (string-append "--without-zlib")
 315                                (string-append "--without-libxml2"))
 316        #:phases (modify-phases %standard-phases
 317                   (add-after 'install 'wrap-programs
 318                     (lambda* (#:key inputs outputs #:allow-other-keys)
 319                       (let* ((out (assoc-ref outputs "out")))
 320                         (for-each
 321                          (lambda (program)
 322                            (wrap-program program
 323                              `("PERL5LIB" ":" prefix
 324                                ,(map (lambda (i)
 325                                        (string-append (assoc-ref inputs i)
 326                                                       "/lib/perl5/site_perl"))
 327                                      ;; These perl modules have no propagated
 328                                      ;; inputs, so no further analysis needed.
 329                                      '("perl-uri"
 330                                        "perl-html-parser"
 331                                        "perl-html-tagset"
 332                                        "perl-mime-types")))))
 333                          (list (string-append out "/lib/swish-e/swishspider")
 334                                (string-append out "/bin/swish-filter-test")))
 335                         #t))))))
 336     (home-page (string-append "https://web.archive.org/web/20160730145202/"
 337                               "http://swish-e.org"))
 338     (synopsis "Web indexing system")
 339     (description
 340      "Swish-e is Simple Web Indexing System for Humans - Enhanced.  Swish-e
 341 can quickly and easily index directories of files or remote web sites and
 342 search the generated indexes.")
 343     (license gpl2+)))                   ;with exception
 344
 345 (define-public xapers
 346   (package
 347     (name "xapers")
 348     (version "0.8.2")
 349     (source
 350      (origin
 351        (method url-fetch)
 352        (uri (string-append
 353              "https://finestructure.net/xapers/releases/xapers-"
 354              version ".tar.gz"))
 355        (sha256
 356         (base32
 357          "0ykz6hn3qj46w3c99d6q0pi5ncq2894simcl7vapv047zm3cylmd"))))
 358     (build-system python-build-system)
 359     (propagated-inputs
 360      `(("poppler" ,poppler)
 361        ("python-urwid" ,python-urwid)
 362        ("xclip" ,xclip)
 363        ("xdg-utils" ,xdg-utils)))
 364     (inputs
 365      `(("python-latexcodec" ,python-latexcodec)
 366        ("python-pybtex" ,python-pybtex)
 367        ("python-pycurl" ,python-pycurl)
 368        ("python-pyyaml" ,python-pyyaml)
 369        ("python-six" ,python-six)
 370        ("python-xapian-bindings" ,python-xapian-bindings)))
 371     (arguments
 372      `(#:modules ((ice-9 rdelim)
 373                   (guix build python-build-system)
 374                   (guix build utils))
 375        #:phases
 376        (modify-phases %standard-phases
 377          (add-after 'install 'install-doc
 378            (lambda* (#:key inputs outputs #:allow-other-keys)
 379              (define (purge-term-support input output)
 380                (let loop ((line (read-line input)))
 381                  (if (string-prefix? "if [[ \"$term\"" line)
 382                      (begin (display "eval \"$cmd\"\n" output)
 383                             #t)
 384                      (begin (display (string-append line "\n") output)
 385                             (loop (read-line input))))))
 386              (let* ((out (assoc-ref outputs "out"))
 387                     (bin (string-append out "/bin"))
 388                     (adder-out (string-append bin "/xapers-adder"))
 389                     (man1 (string-append out "/share/man/man1")))
 390                (install-file "man/man1/xapers.1"  man1)
 391                (install-file "man/man1/xapers-adder.1" man1)
 392                ;; below is equivalent to setting --no-term option
 393                ;; permanently on; this is desirable to avoid imposing
 394                ;; an x-terminal installation on the user but breaks
 395                ;; some potential xapers-adder uses like auto browser
 396                ;; pdf handler, but user could instead still use
 397                ;; e.g. "xterm -e xapers-adder %F" for same use.
 398                ;; alternatively we could propagate xterm as an input
 399                ;; and replace 'x-terminal-emulator' with 'xterm'
 400                (call-with-input-file "bin/xapers-adder"
 401                  (lambda (input)
 402                    (call-with-output-file adder-out
 403                      (lambda (output)
 404                        (purge-term-support input output)))))
 405                (chmod adder-out #o555)))))))
 406     (home-page "https://finestructure.net/xapers/")
 407     (synopsis "Personal document indexing system")
 408     (description
 409      "Xapers is a personal document indexing system,
 410 geared towards academic journal articles build on the Xapian search engine.
 411 Think of it as your own personal document search engine, or a local cache of
 412 online libraries.  It provides fast search of document text and
 413 bibliographic data and simple document and bibtex retrieval.")
 414     (license gpl3+)))
 415
 416 ;;; search.scm ends here