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, 2020 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 "1.4.17")
  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 "0bjpaavdckl4viznr8gbq476fvg648sj4rks2vacmc51vrb8bsxm"))))
  60     (build-system gnu-build-system)
  61     (inputs `(("zlib" ,zlib)
  62               ("util-linux" ,util-linux "lib")))
  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                 "1lzb739hjfy9ih2c0ircra50h4jr48wxq1fal7sai963w28mv9j8"))))
  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     (native-inputs
 108      `(("python-sphinx" ,python-sphinx))) ;for documentation
 109     (inputs
 110      `(("python" ,python)
 111        ("xapian" ,xapian)
 112        ("zlib" ,zlib)))
 113     (synopsis "Python bindings for the Xapian search engine library")
 114     (license gpl2+)))
 115
 116 (define-public perl-search-xapian
 117   (package
 118     (name "perl-search-xapian")
 119     (version "1.2.25.2")
 120     (source
 121      (origin
 122        (method url-fetch)
 123        (uri (string-append "mirror://cpan/authors/id/O/OL/OLLY/"
 124                            "Search-Xapian-" version ".tar.gz"))
 125        (sha256
 126         (base32
 127          "0hpa8gi38j0ibq8af6dy69lm1bl5jnq76nsa69dbrzbr88l5m594"))))
 128     (build-system perl-build-system)
 129     (native-inputs
 130      `(("perl-devel-leak" ,perl-devel-leak)))
 131     (inputs
 132      `(("xapian" ,xapian)))
 133     (home-page "https://metacpan.org/release/Search-Xapian")
 134     (synopsis "Perl XS frontend to the Xapian C++ search library")
 135     (description
 136      "Search::Xapian wraps most methods of most Xapian classes.  The missing
 137 classes and methods should be added in the future.  It also provides a
 138 simplified, more 'perlish' interface to some common operations.")
 139     (license perl-license)))
 140
 141 (define-public libtocc
 142   (package
 143     (name "libtocc")
 144     (version "1.0.1")
 145     (source
 146      (origin
 147        (method url-fetch)
 148        (uri (string-append "https://github.com/aidin36/tocc/releases/download/"
 149                            "v" version "/tocc-" version ".tar.gz"))
 150        (sha256
 151         (base32
 152          "1kd2jd74m8ksc8s7hh0haz0q0c3n0mr39bbky262kk4l58f1g068"))))
 153     (build-system gnu-build-system)
 154     (native-inputs `(("catch" ,catch-framework)))
 155     (inputs `(("unqlite" ,unqlite)))
 156     (arguments
 157      `(#:phases (modify-phases %standard-phases
 158                   (add-before 'configure 'chdir-source
 159                     (lambda _
 160                       (chdir "libtocc/src")
 161                       #t))
 162                   (replace 'check
 163                     (lambda _
 164                       (with-directory-excursion "../tests"
 165                         (invoke "./configure"
 166                                 (string-append "CONFIG_SHELL="
 167                                                (which "sh"))
 168                                 (string-append "SHELL="
 169                                                (which "sh"))
 170                                 "CPPFLAGS=-I../src"
 171                                 (string-append
 172                                  "LDFLAGS=-L../src/.libs "
 173                                  "-Wl,-rpath=../src/.libs"))
 174                         (invoke "make")
 175                         (invoke "./libtocctests")))))))
 176     (home-page "https://t-o-c-c.com/")
 177     (synopsis "Tool for Obsessive Compulsive Classifiers")
 178     (description
 179      "libtocc is the engine of the Tocc project, a tag-based file management
 180 system.  The goal of Tocc is to provide a better system for classifying files
 181 that is more flexible than classic file systems that are based on a tree of
 182 files and directories.")
 183     (license gpl3+)))
 184
 185 (define-public tocc
 186   (package
 187     (name "tocc")
 188     (version (package-version libtocc))
 189     (source (package-source libtocc))
 190     (build-system gnu-build-system)
 191     (inputs
 192      `(("libtocc" ,libtocc)
 193        ("unqlite" ,unqlite)))
 194     (arguments
 195      `(#:tests? #f                      ;No tests
 196        #:phases (modify-phases %standard-phases
 197                   (add-after
 198                    'unpack 'chdir-source
 199                    (lambda _ (chdir "cli/src"))))))
 200     (home-page "https://t-o-c-c.com/")
 201     (synopsis "Command-line interface to libtocc")
 202     (description
 203      "Tocc is a tag-based file management system.  This package contains the
 204 command line tool for interacting with libtocc.")
 205     (license gpl3+)))
 206
 207 (define-public bool
 208   (package
 209     (name "bool")
 210     (version "0.2.2")
 211     (source
 212      (origin
 213        (method url-fetch)
 214        (uri (string-append "mirror://gnu/bool/bool-"
 215                            version ".tar.xz"))
 216        (sha256
 217         (base32
 218          "1frdmgrmb509fxbdpsxxw3lvvwv7xm1pavqrqgm4jg698iix6xfw"))))
 219     (build-system gnu-build-system)
 220     (home-page "https://www.gnu.org/software/bool/")
 221     (synopsis "Finding text and HTML files that match boolean expressions")
 222     (description
 223      "GNU Bool is a utility to perform text searches on files using Boolean
 224 expressions.  For example, a search for \"hello AND world\" would return a
 225 file containing the phrase \"Hello, world!\".  It supports both AND and OR
 226 statements, as well as the NEAR statement to search for the occurrence of
 227 words in close proximity to each other.  It handles context gracefully,
 228 accounting for new lines and paragraph changes.  It also has robust support
 229 for parsing HTML files.")
 230     (license gpl3+)))
 231
 232 (define-public hyperestraier
 233   (package
 234     (name "hyperestraier")
 235     (version "1.4.13")
 236     (source
 237       (origin
 238         (method url-fetch)
 239         (uri (string-append "http://fallabs.com/" name "/"
 240                             name "-" version ".tar.gz"))
 241         (sha256
 242          (base32
 243           "1qk3pxgzyrpcz5qfyd5xs2hw9q1cbb7j5zd4kp1diq501wcj2vs9"))))
 244     (inputs
 245      `(("qdbm" ,qdbm)
 246        ("zlib" ,zlib)))
 247     (build-system gnu-build-system)
 248     (arguments
 249      `(#:configure-flags (list (string-append "LDFLAGS=-Wl,-rpath="
 250                                               (assoc-ref %outputs "out")
 251                                               "/lib"))))
 252     (home-page "https://fallabs.com/hyperestraier")
 253     (synopsis "Full-text search system")
 254     (description "Hyper Estraier can be used to integrate full-text
 255 search into applications, using either the provided command line and CGI
 256 interfaces, or a C API.")
 257     (license lgpl2.1+)))
 258
 259 (define-public mlocate
 260   (package
 261     (name "mlocate")
 262     (version "0.26")
 263     (source (origin
 264               (method url-fetch)
 265               (uri (string-append "http://releases.pagure.org/mlocate/"
 266                                   "mlocate-" version ".tar.xz"))
 267               (sha256
 268                (base32
 269                 "0gi6y52gkakhhlnzy0p6izc36nqhyfx5830qirhvk3qrzrwxyqrh"))))
 270     (build-system gnu-build-system)
 271     (home-page "https://pagure.io/mlocate")
 272     (synopsis "Locate files on the file system")
 273     (description
 274      "mlocate is a locate/updatedb implementation.  The @code{m} stands for
 275 \"merging\": @code{updatedb} reuses the existing database to avoid rereading
 276 most of the file system, which makes it faster and does not trash the system
 277 caches as much.  The locate(1) utility is intended to be completely compatible
 278 with slocate, and attempts to be compatible to GNU locate when it does not
 279 conflict with slocate compatibility.")
 280     (license gpl2)))
 281
 282 (define-public swish-e
 283   (package
 284     (name "swish-e")
 285     (version "2.4.7")
 286     (source (origin
 287               (method url-fetch)
 288               (uri (list (string-append
 289                           "https://web.archive.org/web/20160730145202/"
 290                           "http://swish-e.org/distribution/"
 291                           "swish-e-" version ".tar.gz")
 292                          (string-append "http://http.debian.net/debian/pool/"
 293                                         "main/s/swish-e/swish-e_" version
 294                                         ".orig.tar.gz")))
 295               (file-name (string-append name "-" version ".tar.gz"))
 296               (sha256
 297                (base32
 298                 "0qkrk7z25yp9hynj21vxkyn7yi8gcagcfxnass5cgczcz0gm9pax"))
 299               (patches (search-patches "swish-e-search.patch"
 300                                        "swish-e-format-security.patch"))))
 301     (build-system gnu-build-system)
 302     ;; Several other packages and perl modules may be installed alongside
 303     ;; swish-e to extend its features at runtime, but are not required for
 304     ;; building: xpdf, catdoc, MP3::Tag, Spreadsheet::ParseExcel,
 305     ;; HTML::Entities.
 306     (inputs
 307      `(("perl" ,perl)
 308        ("perl-uri" ,perl-uri)
 309        ("perl-html-parser" ,perl-html-parser)
 310        ("perl-html-tagset" ,perl-html-tagset)
 311        ("perl-mime-types" ,perl-mime-types)))
 312     (arguments
 313      `(;; XXX: This fails to build with zlib (API mismatch) and tests fail
 314        ;; with libxml2, so disable both.
 315        #:configure-flags (list (string-append "--without-zlib")
 316                                (string-append "--without-libxml2"))
 317        #:phases (modify-phases %standard-phases
 318                   (add-after 'install 'wrap-programs
 319                     (lambda* (#:key inputs outputs #:allow-other-keys)
 320                       (let* ((out (assoc-ref outputs "out")))
 321                         (for-each
 322                          (lambda (program)
 323                            (wrap-program program
 324                              `("PERL5LIB" ":" prefix
 325                                ,(map (lambda (i)
 326                                        (string-append (assoc-ref inputs i)
 327                                                       "/lib/perl5/site_perl"))
 328                                      ;; These perl modules have no propagated
 329                                      ;; inputs, so no further analysis needed.
 330                                      '("perl-uri"
 331                                        "perl-html-parser"
 332                                        "perl-html-tagset"
 333                                        "perl-mime-types")))))
 334                          (list (string-append out "/lib/swish-e/swishspider")
 335                                (string-append out "/bin/swish-filter-test")))
 336                         #t))))))
 337     (home-page (string-append "https://web.archive.org/web/20160730145202/"
 338                               "http://swish-e.org"))
 339     (synopsis "Web indexing system")
 340     (description
 341      "Swish-e is Simple Web Indexing System for Humans - Enhanced.  Swish-e
 342 can quickly and easily index directories of files or remote web sites and
 343 search the generated indexes.")
 344     (license gpl2+)))                   ;with exception
 345
 346 (define-public xapers
 347   (package
 348     (name "xapers")
 349     (version "0.8.2")
 350     (source
 351      (origin
 352        (method url-fetch)
 353        (uri (string-append
 354              "https://finestructure.net/xapers/releases/xapers-"
 355              version ".tar.gz"))
 356        (sha256
 357         (base32
 358          "0ykz6hn3qj46w3c99d6q0pi5ncq2894simcl7vapv047zm3cylmd"))))
 359     (build-system python-build-system)
 360     (propagated-inputs
 361      `(("poppler" ,poppler)
 362        ("python-urwid" ,python-urwid)
 363        ("xclip" ,xclip)
 364        ("xdg-utils" ,xdg-utils)))
 365     (inputs
 366      `(("python-latexcodec" ,python-latexcodec)
 367        ("python-pybtex" ,python-pybtex)
 368        ("python-pycurl" ,python-pycurl)
 369        ("python-pyyaml" ,python-pyyaml)
 370        ("python-six" ,python-six)
 371        ("python-xapian-bindings" ,python-xapian-bindings)))
 372     (arguments
 373      `(#:modules ((ice-9 rdelim)
 374                   (guix build python-build-system)
 375                   (guix build utils))
 376        #:phases
 377        (modify-phases %standard-phases
 378          (add-after 'install 'install-doc
 379            (lambda* (#:key inputs outputs #:allow-other-keys)
 380              (define (purge-term-support input output)
 381                (let loop ((line (read-line input)))
 382                  (if (string-prefix? "if [[ \"$term\"" line)
 383                      (begin (display "eval \"$cmd\"\n" output)
 384                             #t)
 385                      (begin (display (string-append line "\n") output)
 386                             (loop (read-line input))))))
 387              (let* ((out (assoc-ref outputs "out"))
 388                     (bin (string-append out "/bin"))
 389                     (adder-out (string-append bin "/xapers-adder"))
 390                     (man1 (string-append out "/share/man/man1")))
 391                (install-file "man/man1/xapers.1"  man1)
 392                (install-file "man/man1/xapers-adder.1" man1)
 393                ;; below is equivalent to setting --no-term option
 394                ;; permanently on; this is desirable to avoid imposing
 395                ;; an x-terminal installation on the user but breaks
 396                ;; some potential xapers-adder uses like auto browser
 397                ;; pdf handler, but user could instead still use
 398                ;; e.g. "xterm -e xapers-adder %F" for same use.
 399                ;; alternatively we could propagate xterm as an input
 400                ;; and replace 'x-terminal-emulator' with 'xterm'
 401                (call-with-input-file "bin/xapers-adder"
 402                  (lambda (input)
 403                    (call-with-output-file adder-out
 404                      (lambda (output)
 405                        (purge-term-support input output)))))
 406                (chmod adder-out #o555)))))))
 407     (home-page "https://finestructure.net/xapers/")
 408     (synopsis "Personal document indexing system")
 409     (description
 410      "Xapers is a personal document indexing system,
 411 geared towards academic journal articles build on the Xapian search engine.
 412 Think of it as your own personal document search engine, or a local cache of
 413 online libraries.  It provides fast search of document text and
 414 bibliographic data and simple document and bibtex retrieval.")
 415     (license gpl3+)))
 416
 417 ;;; search.scm ends here