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, 2021 Tobias Geerinckx-Rice <me@tobias.gr>
   7 ;;; Copyright © 2018 Adam Massmann <massmannak@gmail.com>
   8 ;;; Copyright © 2020 Hartmut Goebel <h.goebel@crazy-compilers.com>
   9 ;;;
  10 ;;; This file is part of GNU Guix.
  11 ;;;
  12 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  13 ;;; under the terms of the GNU General Public License as published by
  14 ;;; the Free Software Foundation; either version 3 of the License, or (at
  15 ;;; your option) any later version.
  16 ;;;
  17 ;;; GNU Guix is distributed in the hope that it will be useful, but
  18 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  19 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  20 ;;; GNU General Public License for more details.
  21 ;;;
  22 ;;; You should have received a copy of the GNU General Public License
  23 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  24
  25 (define-module (gnu packages search)
  26   #:use-module ((guix licenses)
  27                 #:select (gpl2 gpl2+ gpl3+ lgpl2.1+ bsd-3 x11 perl-license))
  28   #:use-module (guix packages)
  29   #:use-module (guix download)
  30   #:use-module (guix git-download)
  31   #:use-module (guix utils)
  32   #:use-module (guix build-system gnu)
  33   #:use-module (guix build-system perl)
  34   #:use-module (guix build-system python)
  35   #:use-module (gnu packages)
  36   #:use-module (gnu packages compression)
  37   #:use-module (gnu packages check)
  38   #:use-module (gnu packages databases)
  39   #:use-module (gnu packages freedesktop)
  40   #:use-module (gnu packages less)
  41   #:use-module (gnu packages linux)
  42   #:use-module (gnu packages pcre)
  43   #:use-module (gnu packages perl)
  44   #:use-module (gnu packages pdf)
  45   #:use-module (gnu packages python)
  46   #:use-module (gnu packages python-web)
  47   #:use-module (gnu packages python-xyz)
  48   #:use-module (gnu packages sphinx)
  49   #:use-module (gnu packages web)
  50   #:use-module (gnu packages xdisorg)
  51   #:use-module (gnu packages xml))
  52
  53 (define-public xapian
  54   (package
  55     (name "xapian")
  56     (version "1.4.18")
  57     ;; Note: When updating Xapian, remember to update xapian-bindings below.
  58     (source (origin
  59               (method url-fetch)
  60               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  61                                   "/xapian-core-" version ".tar.xz"))
  62               (sha256
  63                (base32 "0xsb4ihf3p767f0zx9p4janwni6r9sg5j6lry0002i8hmnsdnv8r"))))
  64     (build-system gnu-build-system)
  65     (inputs `(("zlib" ,zlib)
  66               ("util-linux" ,util-linux "lib")))
  67     (arguments
  68      `(#:phases
  69        (modify-phases %standard-phases
  70          (replace 'check
  71            ;; As of Xapian 1.3.3, the TCP server implementation uses
  72            ;; getaddrinfo(). This does not work in the build environment,
  73            ;; so exclude those tests. See HACKING for the list of targets.
  74            (lambda _
  75              (invoke "make"
  76                      "check-inmemory"
  77                      "check-remoteprog"
  78                      ;"check-remotetcp"
  79                      "check-multi"
  80                      "check-glass"
  81                      "check-chert"))))))
  82     (synopsis "Search Engine Library")
  83     (description
  84      "Xapian is a highly adaptable toolkit which allows developers to easily
  85 add advanced indexing and search facilities to their own applications.  It
  86 supports the Probabilistic Information Retrieval model and also supports a
  87 rich set of boolean query operators.")
  88     (home-page "https://xapian.org/")
  89     (license (list gpl2+ bsd-3 x11))))
  90
  91 (define-public python-xapian-bindings
  92   (package (inherit xapian)
  93     (name "python-xapian-bindings")
  94     (version (package-version xapian))
  95     (source (origin
  96               (method url-fetch)
  97               (uri (string-append "https://oligarchy.co.uk/xapian/" version
  98                                   "/xapian-bindings-" version ".tar.xz"))
  99               (sha256
 100                (base32
 101                 "13ziql8027glgihgvnbsa75vkcn82g83mbihj60zf0njj170clpy"))))
 102     (build-system gnu-build-system)
 103     (arguments
 104      `(#:configure-flags '("--with-python3")
 105        #:make-flags
 106        (list (string-append "pkgpylibdir="
 107                             (assoc-ref %outputs "out")
 108                             "/lib/python" ,(version-major+minor
 109                                             (package-version python))
 110                             "/site-packages/xapian"))))
 111     (native-inputs
 112      `(("python-sphinx" ,python-sphinx))) ;for documentation
 113     (inputs
 114      `(("python" ,python)
 115        ("xapian" ,xapian)
 116        ("zlib" ,zlib)))
 117     (synopsis "Python bindings for the Xapian search engine library")
 118     (license gpl2+)))
 119
 120 (define-public perl-search-xapian
 121   (package
 122     (name "perl-search-xapian")
 123     (version "1.2.25.4")
 124     (source
 125      (origin
 126        (method url-fetch)
 127        (uri (string-append "mirror://cpan/authors/id/O/OL/OLLY/"
 128                            "Search-Xapian-" version ".tar.gz"))
 129        (sha256
 130         (base32 "1pbl8pbgmbs3i8yik4p63g4pd9bhn0dp3d7l667dkvw0kccl66c7"))))
 131     (build-system perl-build-system)
 132     (native-inputs
 133      `(("perl-devel-leak" ,perl-devel-leak)))
 134     (inputs
 135      `(("xapian" ,xapian)))
 136     (home-page "https://metacpan.org/release/Search-Xapian")
 137     (synopsis "Perl XS frontend to the Xapian C++ search library")
 138     (description
 139      "Search::Xapian wraps most methods of most Xapian classes.  The missing
 140 classes and methods should be added in the future.  It also provides a
 141 simplified, more 'perlish' interface to some common operations.")
 142     (license perl-license)))
 143
 144 (define-public libtocc
 145   (package
 146     (name "libtocc")
 147     (version "1.0.1")
 148     (source
 149      (origin
 150        (method url-fetch)
 151        (uri (string-append "https://github.com/aidin36/tocc/releases/download/"
 152                            "v" version "/tocc-" version ".tar.gz"))
 153        (sha256
 154         (base32
 155          "1kd2jd74m8ksc8s7hh0haz0q0c3n0mr39bbky262kk4l58f1g068"))))
 156     (build-system gnu-build-system)
 157     (native-inputs `(("catch" ,catch-framework)))
 158     (inputs `(("unqlite" ,unqlite)))
 159     (arguments
 160      `(#:phases (modify-phases %standard-phases
 161                   (add-before 'configure 'chdir-source
 162                     (lambda _
 163                       (chdir "libtocc/src")
 164                       #t))
 165                   (replace 'check
 166                     (lambda _
 167                       (with-directory-excursion "../tests"
 168                         (invoke "./configure"
 169                                 (string-append "CONFIG_SHELL="
 170                                                (which "sh"))
 171                                 (string-append "SHELL="
 172                                                (which "sh"))
 173                                 "CPPFLAGS=-I../src"
 174                                 (string-append
 175                                  "LDFLAGS=-L../src/.libs "
 176                                  "-Wl,-rpath=../src/.libs"))
 177                         (invoke "make")
 178                         (invoke "./libtocctests")))))))
 179     (home-page "https://t-o-c-c.com/")
 180     (synopsis "Tool for Obsessive Compulsive Classifiers")
 181     (description
 182      "libtocc is the engine of the Tocc project, a tag-based file management
 183 system.  The goal of Tocc is to provide a better system for classifying files
 184 that is more flexible than classic file systems that are based on a tree of
 185 files and directories.")
 186     (license gpl3+)))
 187
 188 (define-public tocc
 189   (package
 190     (name "tocc")
 191     (version (package-version libtocc))
 192     (source (package-source libtocc))
 193     (build-system gnu-build-system)
 194     (inputs
 195      `(("libtocc" ,libtocc)
 196        ("unqlite" ,unqlite)))
 197     (arguments
 198      `(#:tests? #f                      ;No tests
 199        #:phases (modify-phases %standard-phases
 200                   (add-after
 201                    'unpack 'chdir-source
 202                    (lambda _ (chdir "cli/src"))))))
 203     (home-page "https://t-o-c-c.com/")
 204     (synopsis "Command-line interface to libtocc")
 205     (description
 206      "Tocc is a tag-based file management system.  This package contains the
 207 command line tool for interacting with libtocc.")
 208     (license gpl3+)))
 209
 210 (define-public bool
 211   (package
 212     (name "bool")
 213     (version "0.2.2")
 214     (source
 215      (origin
 216        (method url-fetch)
 217        (uri (string-append "mirror://gnu/bool/bool-"
 218                            version ".tar.xz"))
 219        (sha256
 220         (base32
 221          "1frdmgrmb509fxbdpsxxw3lvvwv7xm1pavqrqgm4jg698iix6xfw"))))
 222     (build-system gnu-build-system)
 223     (home-page "https://www.gnu.org/software/bool/")
 224     (synopsis "Finding text and HTML files that match boolean expressions")
 225     (description
 226      "GNU Bool is a utility to perform text searches on files using Boolean
 227 expressions.  For example, a search for \"hello AND world\" would return a
 228 file containing the phrase \"Hello, world!\".  It supports both AND and OR
 229 statements, as well as the NEAR statement to search for the occurrence of
 230 words in close proximity to each other.  It handles context gracefully,
 231 accounting for new lines and paragraph changes.  It also has robust support
 232 for parsing HTML files.")
 233     (license gpl3+)))
 234
 235 (define-public hyperestraier
 236   (package
 237     (name "hyperestraier")
 238     (version "1.4.13")
 239     (source
 240       (origin
 241         (method url-fetch)
 242         (uri (string-append "http://fallabs.com/" name "/"
 243                             name "-" version ".tar.gz"))
 244         (sha256
 245          (base32
 246           "1qk3pxgzyrpcz5qfyd5xs2hw9q1cbb7j5zd4kp1diq501wcj2vs9"))))
 247     (inputs
 248      `(("qdbm" ,qdbm)
 249        ("zlib" ,zlib)))
 250     (build-system gnu-build-system)
 251     (arguments
 252      `(#:configure-flags (list (string-append "LDFLAGS=-Wl,-rpath="
 253                                               (assoc-ref %outputs "out")
 254                                               "/lib"))))
 255     (home-page "https://fallabs.com/hyperestraier")
 256     (synopsis "Full-text search system")
 257     (description "Hyper Estraier can be used to integrate full-text
 258 search into applications, using either the provided command line and CGI
 259 interfaces, or a C API.")
 260     (license lgpl2.1+)))
 261
 262 (define-public mlocate
 263   (package
 264     (name "mlocate")
 265     (version "0.26")
 266     (source (origin
 267               (method url-fetch)
 268               (uri (string-append "http://releases.pagure.org/mlocate/"
 269                                   "mlocate-" version ".tar.xz"))
 270               (sha256
 271                (base32
 272                 "0gi6y52gkakhhlnzy0p6izc36nqhyfx5830qirhvk3qrzrwxyqrh"))))
 273     (build-system gnu-build-system)
 274     (home-page "https://pagure.io/mlocate")
 275     (synopsis "Locate files on the file system")
 276     (description
 277      "mlocate is a locate/updatedb implementation.  The @code{m} stands for
 278 \"merging\": @code{updatedb} reuses the existing database to avoid rereading
 279 most of the file system, which makes it faster and does not trash the system
 280 caches as much.  The locate(1) utility is intended to be completely compatible
 281 with slocate, and attempts to be compatible to GNU locate when it does not
 282 conflict with slocate compatibility.")
 283     (license gpl2)))
 284
 285 (define-public swish-e
 286   (package
 287     (name "swish-e")
 288     (version "2.4.7")
 289     (source (origin
 290               (method url-fetch)
 291               (uri (list (string-append
 292                           "https://web.archive.org/web/20160730145202/"
 293                           "http://swish-e.org/distribution/"
 294                           "swish-e-" version ".tar.gz")
 295                          (string-append "http://http.debian.net/debian/pool/"
 296                                         "main/s/swish-e/swish-e_" version
 297                                         ".orig.tar.gz")))
 298               (file-name (string-append name "-" version ".tar.gz"))
 299               (sha256
 300                (base32
 301                 "0qkrk7z25yp9hynj21vxkyn7yi8gcagcfxnass5cgczcz0gm9pax"))
 302               (patches (search-patches "swish-e-search.patch"
 303                                        "swish-e-format-security.patch"))))
 304     (build-system gnu-build-system)
 305     ;; Several other packages and perl modules may be installed alongside
 306     ;; swish-e to extend its features at runtime, but are not required for
 307     ;; building: xpdf, catdoc, MP3::Tag, Spreadsheet::ParseExcel,
 308     ;; HTML::Entities.
 309     (inputs
 310      `(("perl" ,perl)
 311        ("perl-uri" ,perl-uri)
 312        ("perl-html-parser" ,perl-html-parser)
 313        ("perl-html-tagset" ,perl-html-tagset)
 314        ("perl-mime-types" ,perl-mime-types)))
 315     (arguments
 316      `(;; XXX: This fails to build with zlib (API mismatch) and tests fail
 317        ;; with libxml2, so disable both.
 318        #:configure-flags (list (string-append "--without-zlib")
 319                                (string-append "--without-libxml2"))
 320        #:phases (modify-phases %standard-phases
 321                   (add-after 'install 'wrap-programs
 322                     (lambda* (#:key inputs outputs #:allow-other-keys)
 323                       (let* ((out (assoc-ref outputs "out")))
 324                         (for-each
 325                          (lambda (program)
 326                            (wrap-program program
 327                              `("PERL5LIB" ":" prefix
 328                                ,(map (lambda (i)
 329                                        (string-append (assoc-ref inputs i)
 330                                                       "/lib/perl5/site_perl"))
 331                                      ;; These perl modules have no propagated
 332                                      ;; inputs, so no further analysis needed.
 333                                      '("perl-uri"
 334                                        "perl-html-parser"
 335                                        "perl-html-tagset"
 336                                        "perl-mime-types")))))
 337                          (list (string-append out "/lib/swish-e/swishspider")
 338                                (string-append out "/bin/swish-filter-test")))
 339                         #t))))))
 340     (home-page (string-append "https://web.archive.org/web/20160730145202/"
 341                               "http://swish-e.org"))
 342     (synopsis "Web indexing system")
 343     (description
 344      "Swish-e is Simple Web Indexing System for Humans - Enhanced.  Swish-e
 345 can quickly and easily index directories of files or remote web sites and
 346 search the generated indexes.")
 347     (license gpl2+)))                   ;with exception
 348
 349 (define-public xapers
 350   (package
 351     (name "xapers")
 352     (version "0.8.2")
 353     (source
 354      (origin
 355        (method url-fetch)
 356        (uri (string-append
 357              "https://finestructure.net/xapers/releases/xapers-"
 358              version ".tar.gz"))
 359        (sha256
 360         (base32
 361          "0ykz6hn3qj46w3c99d6q0pi5ncq2894simcl7vapv047zm3cylmd"))))
 362     (build-system python-build-system)
 363     (propagated-inputs
 364      `(("poppler" ,poppler)
 365        ("python-urwid" ,python-urwid)
 366        ("xclip" ,xclip)
 367        ("xdg-utils" ,xdg-utils)))
 368     (inputs
 369      `(("python-latexcodec" ,python-latexcodec)
 370        ("python-pybtex" ,python-pybtex)
 371        ("python-pycurl" ,python-pycurl)
 372        ("python-pyyaml" ,python-pyyaml)
 373        ("python-six" ,python-six)
 374        ("python-xapian-bindings" ,python-xapian-bindings)))
 375     (arguments
 376      `(#:modules ((ice-9 rdelim)
 377                   (guix build python-build-system)
 378                   (guix build utils))
 379        #:phases
 380        (modify-phases %standard-phases
 381          (add-after 'install 'install-doc
 382            (lambda* (#:key inputs outputs #:allow-other-keys)
 383              (define (purge-term-support input output)
 384                (let loop ((line (read-line input)))
 385                  (if (string-prefix? "if [[ \"$term\"" line)
 386                      (begin (display "eval \"$cmd\"\n" output)
 387                             #t)
 388                      (begin (display (string-append line "\n") output)
 389                             (loop (read-line input))))))
 390              (let* ((out (assoc-ref outputs "out"))
 391                     (bin (string-append out "/bin"))
 392                     (adder-out (string-append bin "/xapers-adder"))
 393                     (man1 (string-append out "/share/man/man1")))
 394                (install-file "man/man1/xapers.1"  man1)
 395                (install-file "man/man1/xapers-adder.1" man1)
 396                ;; below is equivalent to setting --no-term option
 397                ;; permanently on; this is desirable to avoid imposing
 398                ;; an x-terminal installation on the user but breaks
 399                ;; some potential xapers-adder uses like auto browser
 400                ;; pdf handler, but user could instead still use
 401                ;; e.g. "xterm -e xapers-adder %F" for same use.
 402                ;; alternatively we could propagate xterm as an input
 403                ;; and replace 'x-terminal-emulator' with 'xterm'
 404                (call-with-input-file "bin/xapers-adder"
 405                  (lambda (input)
 406                    (call-with-output-file adder-out
 407                      (lambda (output)
 408                        (purge-term-support input output)))))
 409                (chmod adder-out #o555)))))))
 410     (home-page "https://finestructure.net/xapers/")
 411     (synopsis "Personal document indexing system")
 412     (description
 413      "Xapers is a personal document indexing system,
 414 geared towards academic journal articles build on the Xapian search engine.
 415 Think of it as your own personal document search engine, or a local cache of
 416 online libraries.  It provides fast search of document text and
 417 bibliographic data and simple document and bibtex retrieval.")
 418     (license gpl3+)))
 419
 420 (define-public ugrep
 421   (package
 422     (name "ugrep")
 423     (version "3.1.11")
 424     (source (origin
 425               (method git-fetch)
 426               (uri (git-reference
 427                     (url "https://github.com/Genivia/ugrep")
 428                     (commit (string-append "v" version))))
 429               (sha256
 430                (base32 "1g3sxnrcaz1jxwa8nwrxpr63g4y0ha5zcf10053ciy9wjh6wqs1w"))
 431               (file-name (string-append name "-" version "-checkout"))
 432               (modules '((guix build utils)))
 433               (snippet
 434                '(begin
 435                   (delete-file-recursively "bin")  ;; pre-build executables
 436                   (for-each delete-file (find-files "tests" "^archive\\..*"))
 437                   (for-each delete-file (find-files "tests" "^.*\\.pdf$"))
 438                   (for-each delete-file (find-files "tests" "^.*\\.class$"))
 439                   #t))))
 440     (build-system gnu-build-system)
 441     (inputs
 442      `(("bzip2" ,bzip2)
 443        ("less" ,less)
 444        ("lz4" ,lz4)
 445        ("lzip" ,lzip)  ;; lzma
 446        ("pcre2" ,pcre2)
 447        ("zlib" ,zlib)))
 448     (arguments
 449      `(#:tests? #f                  ; no way to rebuild the binary input files
 450        #:test-target "test"
 451        #:phases
 452        (modify-phases %standard-phases
 453          (add-before 'check 'check-setup
 454            (lambda _
 455              ;; Unpatch shebangs in tests.
 456              (substitute* '("tests/Hello.bat"
 457                             "tests/Hello.sh")
 458                (("#!/gnu/store/.*/bin/sh") "#!/bin/sh")))))))
 459     (home-page "https://github.com/Genivia/ugrep/")
 460     (synopsis "Faster grep with an interactive query UI")
 461     (description "Ugrep is a ultra fast searcher of file systems, text
 462 and binary files, source code, archives, compressed files, documents, and
 463 more.
 464
 465 While still being compatible with the standard GNU/BSD grep command-line
 466 options, ugrep supports fuzzy search as well as structured and (adjustable)
 467 colored output, piped through \"less\" for pagination.  An interactive query
 468 UI allows refinement and has a built-in help (press F1).  Ugrep implements
 469 multi-threaded and other techniques to speed up search, pattern-matching and
 470 decompression.  Many pre-defined regexps ease searching e.g. C typdefs or XML
 471 attributes.  Results can be output in several structured or self-defined
 472 formats.")
 473     (license bsd-3)))
 474
 475 ;;; search.scm ends here