gnu/packages/machine-learning.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2015, 2016, 2017, 2018, 2019 Ricardo Wurmus <rekado@elephly.net>
   3 ;;; Copyright © 2016 Efraim Flashner <efraim@flashner.co.il>
   4 ;;; Copyright © 2016, 2017 Marius Bakke <mbakke@fastmail.com>
   5 ;;; Copyright © 2016 Hartmut Goebel <h.goebel@crazy-compilers.com>
   6 ;;; Copyright © 2018 Tobias Geerinckx-Rice <me@tobias.gr>
   7 ;;; Copyright © 2018 Kei Kebreau <kkebreau@posteo.net>
   8 ;;; Copyright © 2018 Mark Meyer <mark@ofosos.org>
   9 ;;; Copyright © 2018 Ben Woodcroft <donttrustben@gmail.com>
  10 ;;; Copyright © 2018 Fis Trivial <ybbs.daans@hotmail.com>
  11 ;;; Copyright © 2018 Julien Lepiller <julien@lepiller.eu>
  12 ;;; Copyright © 2018 Björn Höfling <bjoern.hoefling@bjoernhoefling.de>
  13 ;;;
  14 ;;; This file is part of GNU Guix.
  15 ;;;
  16 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  17 ;;; under the terms of the GNU General Public License as published by
  18 ;;; the Free Software Foundation; either version 3 of the License, or (at
  19 ;;; your option) any later version.
  20 ;;;
  21 ;;; GNU Guix is distributed in the hope that it will be useful, but
  22 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  23 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  24 ;;; GNU General Public License for more details.
  25 ;;;
  26 ;;; You should have received a copy of the GNU General Public License
  27 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  28
  29 (define-module (gnu packages machine-learning)
  30   #:use-module ((guix licenses) #:prefix license:)
  31   #:use-module (guix packages)
  32   #:use-module (guix utils)
  33   #:use-module (guix download)
  34   #:use-module (guix svn-download)
  35   #:use-module (guix build-system cmake)
  36   #:use-module (guix build-system gnu)
  37   #:use-module (guix build-system ocaml)
  38   #:use-module (guix build-system python)
  39   #:use-module (guix build-system r)
  40   #:use-module (guix git-download)
  41   #:use-module (gnu packages)
  42   #:use-module (gnu packages algebra)
  43   #:use-module (gnu packages audio)
  44   #:use-module (gnu packages autotools)
  45   #:use-module (gnu packages base)
  46   #:use-module (gnu packages bash)
  47   #:use-module (gnu packages boost)
  48   #:use-module (gnu packages check)
  49   #:use-module (gnu packages compression)
  50   #:use-module (gnu packages cran)
  51   #:use-module (gnu packages dejagnu)
  52   #:use-module (gnu packages gcc)
  53   #:use-module (gnu packages glib)
  54   #:use-module (gnu packages gstreamer)
  55   #:use-module (gnu packages image)
  56   #:use-module (gnu packages linux)
  57   #:use-module (gnu packages maths)
  58   #:use-module (gnu packages mpi)
  59   #:use-module (gnu packages ocaml)
  60   #:use-module (gnu packages onc-rpc)
  61   #:use-module (gnu packages perl)
  62   #:use-module (gnu packages pkg-config)
  63   #:use-module (gnu packages python)
  64   #:use-module (gnu packages python-web)
  65   #:use-module (gnu packages python-xyz)
  66   #:use-module (gnu packages statistics)
  67   #:use-module (gnu packages swig)
  68   #:use-module (gnu packages web)
  69   #:use-module (gnu packages xml)
  70   #:use-module (gnu packages xorg))
  71
  72 (define-public fann
  73   ;; The last release is >100 commits behind, so we package from git.
  74   (let ((commit "d71d54788bee56ba4cf7522801270152da5209d7"))
  75     (package
  76       (name "fann")
  77       (version (string-append "2.2.0-1." (string-take commit 8)))
  78       (source (origin
  79                 (method git-fetch)
  80                 (uri (git-reference
  81                       (url "https://github.com/libfann/fann.git")
  82                       (commit commit)))
  83                 (file-name (string-append name "-" version "-checkout"))
  84                 (sha256
  85                  (base32
  86                   "0ibwpfrjs6q2lijs8slxjgzb2llcl6rk3v2ski4r6215g5jjhg3x"))))
  87       (build-system cmake-build-system)
  88       (arguments
  89        `(#:phases
  90          (modify-phases %standard-phases
  91            (replace 'check
  92              (lambda* (#:key outputs #:allow-other-keys)
  93                (let* ((out (assoc-ref outputs "out")))
  94                  (with-directory-excursion (string-append (getcwd) "/tests")
  95                    (invoke "./fann_tests"))))))))
  96       (home-page "http://leenissen.dk/fann/wp/")
  97       (synopsis "Fast Artificial Neural Network")
  98       (description
  99        "FANN is a free open source neural network library, which implements
 100 multilayer artificial neural networks in C with support for both fully
 101 connected and sparsely connected networks.")
 102       (license license:lgpl2.1))))
 103
 104 (define-public libsvm
 105   (package
 106     (name "libsvm")
 107     (version "3.22")
 108     (source
 109      (origin
 110        (method url-fetch)
 111        (uri (string-append "https://www.csie.ntu.edu.tw/~cjlin/libsvm/"
 112                            name "-" version ".tar.gz"))
 113        (sha256
 114         (base32
 115          "0zd7s19y5vb7agczl6456bn45cj1y64739sslaskw1qk7dywd0bd"))))
 116     (build-system gnu-build-system)
 117     (arguments
 118      `(#:tests? #f ;no "check" target
 119        #:phases (modify-phases %standard-phases
 120                   (delete 'configure)
 121                   (replace
 122                    'install             ; no ‘install’ target
 123                    (lambda* (#:key outputs #:allow-other-keys)
 124                      (let* ((out (assoc-ref outputs "out"))
 125                             (bin (string-append out "/bin/")))
 126                        (mkdir-p bin)
 127                        (for-each (lambda (file)
 128                                    (copy-file file (string-append bin file)))
 129                                  '("svm-train"
 130                                    "svm-predict"
 131                                    "svm-scale")))
 132                      #t)))))
 133     (home-page "http://www.csie.ntu.edu.tw/~cjlin/libsvm/")
 134     (synopsis "Library for Support Vector Machines")
 135     (description
 136      "LIBSVM is a machine learning library for support vector
 137 classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and
 138 distribution estimation (one-class SVM).  It supports multi-class
 139 classification.")
 140     (license license:bsd-3)))
 141
 142 (define-public python-libsvm
 143   (package (inherit libsvm)
 144     (name "python-libsvm")
 145     (build-system gnu-build-system)
 146     (arguments
 147      `(#:tests? #f ;no "check" target
 148        #:make-flags '("-C" "python")
 149        #:phases
 150        (modify-phases %standard-phases
 151          (delete 'configure)
 152          (replace
 153           'install                      ; no ‘install’ target
 154           (lambda* (#:key inputs outputs #:allow-other-keys)
 155             (let ((site (string-append (assoc-ref outputs "out")
 156                                        "/lib/python"
 157                                        (string-take
 158                                         (string-take-right
 159                                          (assoc-ref inputs "python") 5) 3)
 160                                        "/site-packages/")))
 161               (substitute* "python/svm.py"
 162                 (("../libsvm.so.2") "libsvm.so.2"))
 163               (mkdir-p site)
 164               (for-each (lambda (file)
 165                           (copy-file file (string-append site (basename file))))
 166                         (find-files "python" "\\.py"))
 167               (copy-file "libsvm.so.2"
 168                          (string-append site "libsvm.so.2")))
 169             #t)))))
 170     (inputs
 171      `(("python" ,python)))
 172     (synopsis "Python bindings of libSVM")))
 173
 174 (define-public ghmm
 175   ;; The latest release candidate is several years and a couple of fixes have
 176   ;; been published since.  This is why we download the sources from the SVN
 177   ;; repository.
 178   (let ((svn-revision 2341))
 179     (package
 180       (name "ghmm")
 181       (version (string-append "0.9-rc3-0." (number->string svn-revision)))
 182       (source (origin
 183                 (method svn-fetch)
 184                 (uri (svn-reference
 185                       (url "http://svn.code.sf.net/p/ghmm/code/trunk")
 186                       (revision svn-revision)))
 187                 (file-name (string-append name "-" version))
 188                 (sha256
 189                  (base32
 190                   "0qbq1rqp94l530f043qzp8aw5lj7dng9wq0miffd7spd1ff638wq"))))
 191       (build-system gnu-build-system)
 192       (arguments
 193        `(#:imported-modules (,@%gnu-build-system-modules
 194                              (guix build python-build-system))
 195          #:phases
 196          (modify-phases %standard-phases
 197            (add-after 'unpack 'enter-dir
 198              (lambda _ (chdir "ghmm") #t))
 199            (delete 'check)
 200            (add-after 'install 'check
 201              (assoc-ref %standard-phases 'check))
 202            (add-before 'check 'fix-PYTHONPATH
 203              (lambda* (#:key inputs outputs #:allow-other-keys)
 204                (let ((python-version ((@@ (guix build python-build-system)
 205                                            get-python-version)
 206                                       (assoc-ref inputs "python"))))
 207                  (setenv "PYTHONPATH"
 208                          (string-append (getenv "PYTHONPATH")
 209                                         ":" (assoc-ref outputs "out")
 210                                         "/lib/python" python-version
 211                                         "/site-packages")))
 212                #t))
 213            (add-after 'enter-dir 'fix-runpath
 214              (lambda* (#:key outputs #:allow-other-keys)
 215                (substitute* "ghmmwrapper/setup.py"
 216                  (("^(.*)extra_compile_args = \\[" line indent)
 217                   (string-append indent
 218                                  "extra_link_args = [\"-Wl,-rpath="
 219                                  (assoc-ref outputs "out") "/lib\"],\n"
 220                                  line
 221                                  "\"-Wl,-rpath="
 222                                  (assoc-ref outputs "out")
 223                                  "/lib\", ")))
 224                #t))
 225            (add-after 'enter-dir 'disable-broken-tests
 226              (lambda _
 227                (substitute* "tests/Makefile.am"
 228                  ;; GHMM_SILENT_TESTS is assumed to be a command.
 229                  (("TESTS_ENVIRONMENT.*") "")
 230                  ;; Do not build broken tests.
 231                  (("chmm .*") "")
 232                  (("read_fa .*") "")
 233                  (("mcmc .*") "")
 234                  (("label_higher_order_test.*$")
 235                   "label_higher_order_test\n"))
 236
 237                ;; These Python unittests are broken as there is no gato.
 238                ;; See https://sourceforge.net/p/ghmm/support-requests/3/
 239                (substitute* "ghmmwrapper/ghmmunittests.py"
 240                  (("^(.*)def (testNewXML|testMultipleTransitionClasses|testNewXML)"
 241                    line indent)
 242                   (string-append indent
 243                                  "@unittest.skip(\"Disabled by Guix\")\n"
 244                                  line)))
 245                #t))
 246            (add-after 'disable-broken-tests 'autogen
 247              (lambda _
 248                (invoke "bash" "autogen.sh"))))))
 249       (inputs
 250        `(("python" ,python-2) ; only Python 2 is supported
 251          ("libxml2" ,libxml2)))
 252       (native-inputs
 253        `(("pkg-config" ,pkg-config)
 254          ("dejagnu" ,dejagnu)
 255          ("swig" ,swig)
 256          ("autoconf" ,autoconf)
 257          ("automake" ,automake)
 258          ("libtool" ,libtool)))
 259       (home-page "http://ghmm.org")
 260       (synopsis "Hidden Markov Model library")
 261       (description
 262        "The General Hidden Markov Model library (GHMM) is a C library with
 263 additional Python bindings implementing a wide range of types of @dfn{Hidden
 264 Markov Models} (HMM) and algorithms: discrete, continuous emissions, basic
 265 training, HMM clustering, HMM mixtures.")
 266       (license license:lgpl2.0+))))
 267
 268 (define-public mcl
 269   (package
 270     (name "mcl")
 271     (version "14.137")
 272     (source (origin
 273               (method url-fetch)
 274               (uri (string-append
 275                     "http://micans.org/mcl/src/mcl-"
 276                     (string-replace-substring version "." "-")
 277                     ".tar.gz"))
 278               (sha256
 279                (base32
 280                 "15xlax3z31lsn62vlg94hkm75nm40q4679amnfg13jm8m2bnhy5m"))))
 281     (build-system gnu-build-system)
 282     (arguments
 283      `(#:configure-flags (list "--enable-blast")))
 284     (inputs
 285      `(("perl" ,perl)))
 286     (home-page "http://micans.org/mcl/")
 287     (synopsis "Clustering algorithm for graphs")
 288     (description
 289      "The MCL algorithm is short for the @dfn{Markov Cluster Algorithm}, a
 290 fast and scalable unsupervised cluster algorithm for graphs (also known as
 291 networks) based on simulation of (stochastic) flow in graphs.")
 292     ;; In the LICENCE file and web page it says "The software is licensed
 293     ;; under the GNU General Public License, version 3.", but in several of
 294     ;; the source code files it suggests GPL3 or later.
 295     ;; http://listserver.ebi.ac.uk/pipermail/mcl-users/2016/000376.html
 296     (license license:gpl3)))
 297
 298 (define-public ocaml-mcl
 299   (package
 300     (name "ocaml-mcl")
 301     (version "12-068oasis4")
 302     (source
 303      (origin
 304        (method url-fetch)
 305        (uri (string-append
 306              "https://github.com/fhcrc/mcl/archive/"
 307              version ".tar.gz"))
 308        (file-name (string-append name "-" version ".tar.gz"))
 309        (sha256
 310         (base32
 311          "1l5jbhwjpsj38x8b9698hfpkv75h8hn3kj0gihjhn8ym2cwwv110"))))
 312     (build-system ocaml-build-system)
 313     (arguments
 314      `(#:ocaml ,ocaml-4.02
 315        #:findlib ,ocaml4.02-findlib
 316        #:phases
 317        (modify-phases %standard-phases
 318          (add-before 'configure 'patch-paths
 319            (lambda _
 320              (substitute* "configure"
 321                (("SHELL = /bin/sh") (string-append "SHELL = "(which "sh"))))
 322              (substitute* "setup.ml"
 323                (("LDFLAGS=-fPIC")
 324                 (string-append "LDFLAGS=-fPIC\"; \"SHELL=" (which "sh"))))
 325              #t)))))
 326     (home-page "https://github.com/fhcrc/mcl")
 327     (synopsis "OCaml wrappers around MCL")
 328     (description
 329      "This package provides OCaml bindings for the MCL graph clustering
 330 algorithm.")
 331     (license license:gpl3)))
 332
 333 (define-public randomjungle
 334   (package
 335     (name "randomjungle")
 336     (version "2.1.0")
 337     (source
 338      (origin
 339        (method url-fetch)
 340        (uri (string-append
 341              "https://www.imbs.uni-luebeck.de/fileadmin/files/Software"
 342              "/randomjungle/randomjungle-" version ".tar_.gz"))
 343        (patches (search-patches "randomjungle-disable-static-build.patch"))
 344        (sha256
 345         (base32
 346          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 347     (build-system gnu-build-system)
 348     (arguments
 349      `(#:configure-flags
 350        (list "--disable-static"
 351              (string-append "--with-boost="
 352                             (assoc-ref %build-inputs "boost")))
 353        #:phases
 354        (modify-phases %standard-phases
 355          (add-before
 356           'configure 'set-CXXFLAGS
 357           (lambda _
 358             (setenv "CXXFLAGS" "-fpermissive ")
 359             #t)))))
 360     (inputs
 361      `(("boost" ,boost)
 362        ("gsl" ,gsl)
 363        ("libxml2" ,libxml2)
 364        ("zlib" ,zlib)))
 365     (native-inputs
 366      `(("gfortran" ,gfortran)
 367        ("gfortran:lib" ,gfortran "lib")))
 368     ;; Non-portable assembly instructions are used so building fails on
 369     ;; platforms other than x86_64 or i686.
 370     (supported-systems '("x86_64-linux" "i686-linux"))
 371     (home-page "https://www.imbs.uni-luebeck.de/forschung/software/details.html#c224")
 372     (synopsis "Implementation of the Random Forests machine learning method")
 373     (description
 374      "Random Jungle is an implementation of Random Forests.  It is supposed to
 375 analyse high dimensional data.  In genetics, it can be used for analysing big
 376 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 377 learning method.  Most interesting features are variable selection, missing
 378 value imputation, classifier creation, generalization error estimation and
 379 sample proximities between pairs of cases.")
 380     (license license:gpl3+)))
 381
 382 (define-public openfst
 383   (package
 384     (name "openfst")
 385     (version "1.7.1")
 386     (source (origin
 387               (method url-fetch)
 388               (uri (string-append "http://www.openfst.org/twiki/pub/FST/"
 389                                   "FstDownload/openfst-" version ".tar.gz"))
 390               (sha256
 391                (base32
 392                 "0x9wfcqd8hq4h349s7j77sr60h8xjdfshqw1m3a2n6z5bdr9qkm1"))))
 393     (build-system gnu-build-system)
 394     (home-page "http://www.openfst.org")
 395     (synopsis "Library for weighted finite-state transducers")
 396     (description "OpenFst is a library for constructing, combining,
 397 optimizing, and searching weighted finite-state transducers (FSTs).")
 398     (license license:asl2.0)))
 399
 400 (define-public shogun
 401   (package
 402     (name "shogun")
 403     (version "6.1.3")
 404     (source
 405      (origin
 406        (method url-fetch)
 407        (uri (string-append
 408              "ftp://shogun-toolbox.org/shogun/releases/"
 409              (version-major+minor version)
 410              "/sources/shogun-" version ".tar.bz2"))
 411        (sha256
 412         (base32
 413          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 414        (modules '((guix build utils)
 415                   (ice-9 rdelim)))
 416        (snippet
 417         '(begin
 418            ;; Remove non-free sources and files referencing them
 419            (for-each delete-file
 420                      (find-files "src/shogun/classifier/svm/"
 421                                  "SVMLight\\.(cpp|h)"))
 422            (for-each delete-file
 423                      (find-files "examples/undocumented/libshogun/"
 424                                  (string-append
 425                                   "(classifier_.*svmlight.*|"
 426                                   "evaluation_cross_validation_locked_comparison).cpp")))
 427            ;; Remove non-free functions.
 428            (define (delete-ifdefs file)
 429              (with-atomic-file-replacement file
 430                (lambda (in out)
 431                  (let loop ((line (read-line in 'concat))
 432                             (skipping? #f))
 433                    (if (eof-object? line)
 434                        #t
 435                        (let ((skip-next?
 436                               (or (and skipping?
 437                                        (not (string-prefix?
 438                                              "#endif //USE_SVMLIGHT" line)))
 439                                   (string-prefix?
 440                                    "#ifdef USE_SVMLIGHT" line))))
 441                          (when (or (not skipping?)
 442                                    (and skipping? (not skip-next?)))
 443                            (display line out))
 444                          (loop (read-line in 'concat) skip-next?)))))))
 445            (for-each delete-ifdefs
 446                      (append
 447                       (find-files "src/shogun/classifier/mkl"
 448                                   "^MKLClassification\\.cpp")
 449                       (find-files "src/shogun/classifier/svm"
 450                                   "^SVMLightOneClass\\.(cpp|h)")
 451                       (find-files "src/shogun/multiclass"
 452                                   "^ScatterSVM\\.(cpp|h)")
 453                       (find-files "src/shogun/kernel/"
 454                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 455                       (find-files "src/shogun/regression/svr"
 456                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 457                       (find-files "src/shogun/transfer/domain_adaptation"
 458                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 459            #t))))
 460     (build-system cmake-build-system)
 461     (arguments
 462      '(#:tests? #f ;no check target
 463        #:phases
 464        (modify-phases %standard-phases
 465          (add-after 'unpack 'delete-broken-symlinks
 466            (lambda _
 467              (for-each delete-file '("applications/arts/data"
 468                                      "applications/asp/data"
 469                                      "applications/easysvm/data"
 470                                      "applications/msplicer/data"
 471                                      "applications/ocr/data"
 472                                      "examples/meta/data"
 473                                      "examples/undocumented/data"))
 474              #t))
 475          (add-after 'unpack 'change-R-target-path
 476            (lambda* (#:key outputs #:allow-other-keys)
 477              (substitute* '("src/interfaces/r/CMakeLists.txt"
 478                             "examples/meta/r/CMakeLists.txt")
 479                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 480                 (string-append (assoc-ref outputs "out")
 481                                "/lib/R/library/")))
 482              #t))
 483          (add-after 'unpack 'fix-octave-modules
 484            (lambda* (#:key outputs #:allow-other-keys)
 485              (substitute* "src/interfaces/octave/CMakeLists.txt"
 486                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 487                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 488                ;; change target directory
 489                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 490                 (string-append (assoc-ref outputs "out")
 491                                "/share/octave/packages")))
 492              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 493                             "src/interfaces/octave/sg_print_functions.cpp")
 494                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 495                (("octave/config\\.h") "octave/octave-config.h")
 496                (("octave/oct-obj.h") "octave/ovl.h"))
 497              #t))
 498          (add-after 'unpack 'move-rxcpp
 499            (lambda* (#:key inputs #:allow-other-keys)
 500              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 501                (mkdir-p rxcpp-dir)
 502                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 503                #t)))
 504          (add-before 'build 'set-HOME
 505            ;; $HOME needs to be set at some point during the build phase
 506            (lambda _ (setenv "HOME" "/tmp") #t)))
 507        #:configure-flags
 508        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 509              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 510              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 511              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 512              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 513              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 514              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 515              "-DINTERFACE_OCTAVE=ON"
 516              "-DINTERFACE_PYTHON=ON"
 517              "-DINTERFACE_R=ON")))
 518     (inputs
 519      `(("python" ,python)
 520        ("numpy" ,python-numpy)
 521        ("r-minimal" ,r-minimal)
 522        ("octave" ,octave-cli)
 523        ("swig" ,swig)
 524        ("eigen" ,eigen)
 525        ("hdf5" ,hdf5)
 526        ("atlas" ,atlas)
 527        ("arpack" ,arpack-ng)
 528        ("lapack" ,lapack)
 529        ("glpk" ,glpk)
 530        ("libxml2" ,libxml2)
 531        ("lzo" ,lzo)
 532        ("zlib" ,zlib)))
 533     (native-inputs
 534      `(("pkg-config" ,pkg-config)
 535        ("rxcpp" ,rxcpp)))
 536     ;; Non-portable SSE instructions are used so building fails on platforms
 537     ;; other than x86_64.
 538     (supported-systems '("x86_64-linux"))
 539     (home-page "http://shogun-toolbox.org/")
 540     (synopsis "Machine learning toolbox")
 541     (description
 542      "The Shogun Machine learning toolbox provides a wide range of unified and
 543 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 544 combine multiple data representations, algorithm classes, and general purpose
 545 tools.  This enables both rapid prototyping of data pipelines and extensibility
 546 in terms of new algorithms.")
 547     (license license:gpl3+)))
 548
 549 (define-public rxcpp
 550   (package
 551     (name "rxcpp")
 552     (version "4.0.0")
 553     (source
 554      (origin
 555        (method url-fetch)
 556        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 557                            version ".tar.gz"))
 558        (sha256
 559         (base32
 560          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 561        (file-name (string-append name "-" version ".tar.gz"))))
 562     (build-system cmake-build-system)
 563     (arguments
 564      `(#:phases
 565        (modify-phases %standard-phases
 566          (add-after 'unpack 'remove-werror
 567            (lambda _
 568              (substitute* (find-files ".")
 569                (("-Werror") ""))
 570              #t))
 571          (replace 'check
 572            (lambda _
 573              (invoke "ctest"))))))
 574     (native-inputs
 575      `(("catch" ,catch-framework)))
 576     (home-page "http://reactivex.io/")
 577     (synopsis "Reactive Extensions for C++")
 578     (description
 579      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 580 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 581 and event-based programs by using observable sequences.
 582
 583 It extends the observer pattern to support sequences of data and/or events and
 584 adds operators that allow you to compose sequences together declaratively while
 585 abstracting away concerns about things like low-level threading,
 586 synchronization, thread-safety, concurrent data structures, and non-blocking
 587 I/O.")
 588     (license license:asl2.0)))
 589
 590 (define-public r-adaptivesparsity
 591   (package
 592     (name "r-adaptivesparsity")
 593     (version "1.6")
 594     (source (origin
 595               (method url-fetch)
 596               (uri (cran-uri "AdaptiveSparsity" version))
 597               (sha256
 598                (base32
 599                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 600     (properties
 601      `((upstream-name . "AdaptiveSparsity")))
 602     (build-system r-build-system)
 603     (arguments
 604      `(#:phases
 605        (modify-phases %standard-phases
 606          (add-after 'unpack 'link-against-armadillo
 607            (lambda _
 608              (substitute* "src/Makevars"
 609                (("PKG_LIBS=" prefix)
 610                 (string-append prefix "-larmadillo"))))))))
 611     (propagated-inputs
 612      `(("r-mass" ,r-mass)
 613        ("r-matrix" ,r-matrix)
 614        ("r-rcpp" ,r-rcpp)
 615        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 616     (inputs
 617      `(("armadillo" ,armadillo)))
 618     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 619     (synopsis "Adaptive sparsity models")
 620     (description
 621      "This package implements the Figueiredo machine learning algorithm for
 622 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 623 geometric models.")
 624     (license license:lgpl3+)))
 625
 626 (define-public r-kernlab
 627   (package
 628     (name "r-kernlab")
 629     (version "0.9-27")
 630     (source
 631      (origin
 632        (method url-fetch)
 633        (uri (cran-uri "kernlab" version))
 634        (sha256
 635         (base32
 636          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 637     (build-system r-build-system)
 638     (home-page "https://cran.r-project.org/web/packages/kernlab")
 639     (synopsis "Kernel-based machine learning tools")
 640     (description
 641      "This package provides kernel-based machine learning methods for
 642 classification, regression, clustering, novelty detection, quantile regression
 643 and dimensionality reduction.  Among other methods @code{kernlab} includes
 644 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 645 and a QP solver.")
 646     (license license:gpl2)))
 647
 648 (define-public dlib
 649   (package
 650     (name "dlib")
 651     (version "19.7")
 652     (source (origin
 653               (method url-fetch)
 654               (uri (string-append
 655                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 656               (sha256
 657                (base32
 658                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 659               (modules '((guix build utils)))
 660               (snippet
 661                '(begin
 662                   ;; Delete ~13MB of bundled dependencies.
 663                   (delete-file-recursively "dlib/external")
 664                   (delete-file-recursively "docs/dlib/external")
 665                   #t))))
 666     (build-system cmake-build-system)
 667     (arguments
 668      `(#:phases
 669        (modify-phases %standard-phases
 670          (add-after 'unpack 'disable-asserts
 671            (lambda _
 672              ;; config.h recommends explicitly enabling or disabling asserts
 673              ;; when building as a shared library. By default neither is set.
 674              (substitute* "dlib/config.h"
 675                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 676              #t))
 677          (add-after 'disable-asserts 'disable-failing-tests
 678            (lambda _
 679              ;; One test times out on MIPS, so we need to disable it.
 680              ;; Others are flaky on some platforms.
 681              (let* ((system ,(or (%current-target-system)
 682                                  (%current-system)))
 683                     (disabled-tests (cond
 684                                      ((string-prefix? "mips64" system)
 685                                       '("object_detector" ; timeout
 686                                         "data_io"))
 687                                      ((string-prefix? "armhf" system)
 688                                       '("learning_to_track"))
 689                                      ((string-prefix? "i686" system)
 690                                       '("optimization"))
 691                                      (else '()))))
 692                (for-each
 693                 (lambda (test)
 694                   (substitute* "dlib/test/makefile"
 695                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 696                 disabled-tests)
 697                #t)))
 698          (replace 'check
 699            (lambda _
 700              ;; No test target, so we build and run the unit tests here.
 701              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 702                (with-directory-excursion test-dir
 703                  (invoke "make" "-j" (number->string (parallel-job-count)))
 704                  (invoke "./dtest" "--runall"))
 705                #t)))
 706          (add-after 'install 'delete-static-library
 707            (lambda* (#:key outputs #:allow-other-keys)
 708              (delete-file (string-append (assoc-ref outputs "out")
 709                                          "/lib/libdlib.a"))
 710              #t)))))
 711     (native-inputs
 712      `(("pkg-config" ,pkg-config)
 713        ;; For tests.
 714        ("libnsl" ,libnsl)))
 715     (inputs
 716      `(("giflib" ,giflib)
 717        ("lapack" ,lapack)
 718        ("libjpeg" ,libjpeg)
 719        ("libpng" ,libpng)
 720        ("libx11" ,libx11)
 721        ("openblas" ,openblas)
 722        ("zlib" ,zlib)))
 723     (synopsis
 724      "Toolkit for making machine learning and data analysis applications in C++")
 725     (description
 726      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 727 tools.  It is used in both industry and academia in a wide range of domains
 728 including robotics, embedded devices, mobile phones, and large high performance
 729 computing environments.")
 730     (home-page "http://dlib.net")
 731     (license license:boost1.0)))
 732
 733 (define-public python-scikit-learn
 734   (package
 735     (name "python-scikit-learn")
 736     (version "0.20.1")
 737     (source
 738      (origin
 739        (method git-fetch)
 740        (uri (git-reference
 741              (url "https://github.com/scikit-learn/scikit-learn.git")
 742              (commit version)))
 743        (file-name (git-file-name name version))
 744        (sha256
 745         (base32
 746          "0fkhwg3xn1s7ln9q1szq6kwc4jhwvjh8w4kmv9wcrqy7cq3lbv0d"))))
 747     (build-system python-build-system)
 748     (arguments
 749      `(#:phases
 750        (modify-phases %standard-phases
 751          (add-after 'build 'build-ext
 752            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 753          (replace 'check
 754            (lambda _
 755              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 756              (setenv "OPENBLAS_NUM_THREADS" "1")
 757
 758              ;; Some tests require write access to $HOME.
 759              (setenv "HOME" "/tmp")
 760
 761              (invoke "pytest" "sklearn" "-m" "not network")))
 762          ;; FIXME: This fails with permission denied
 763          (delete 'reset-gzip-timestamps))))
 764     (inputs
 765      `(("openblas" ,openblas)))
 766     (native-inputs
 767      `(("python-pytest" ,python-pytest)
 768        ("python-pandas" ,python-pandas) ;for tests
 769        ("python-cython" ,python-cython)))
 770     (propagated-inputs
 771      `(("python-numpy" ,python-numpy)
 772        ("python-scipy" ,python-scipy)))
 773     (home-page "http://scikit-learn.org/")
 774     (synopsis "Machine Learning in Python")
 775     (description
 776      "Scikit-learn provides simple and efficient tools for data mining and
 777 data analysis.")
 778     (license license:bsd-3)))
 779
 780 (define-public python2-scikit-learn
 781   (package-with-python2 python-scikit-learn))
 782
 783 (define-public python-autograd
 784   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 785          (revision "0")
 786          (version (git-version "0.0.0" revision commit)))
 787     (package
 788       (name "python-autograd")
 789       (home-page "https://github.com/HIPS/autograd")
 790       (source (origin
 791                 (method git-fetch)
 792                 (uri (git-reference
 793                       (url home-page)
 794                       (commit commit)))
 795                 (sha256
 796                  (base32
 797                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 798                 (file-name (git-file-name name version))))
 799       (version version)
 800       (build-system python-build-system)
 801       (native-inputs
 802        `(("python-nose" ,python-nose)
 803          ("python-pytest" ,python-pytest)))
 804       (propagated-inputs
 805        `(("python-future" ,python-future)
 806          ("python-numpy" ,python-numpy)))
 807       (arguments
 808        `(#:phases (modify-phases %standard-phases
 809                     (replace 'check
 810                       (lambda _
 811                         (invoke "py.test" "-v"))))))
 812       (synopsis "Efficiently computes derivatives of NumPy code")
 813       (description "Autograd can automatically differentiate native Python and
 814 NumPy code.  It can handle a large subset of Python's features, including loops,
 815 ifs, recursion and closures, and it can even take derivatives of derivatives
 816 of derivatives.  It supports reverse-mode differentiation
 817 (a.k.a. backpropagation), which means it can efficiently take gradients of
 818 scalar-valued functions with respect to array-valued arguments, as well as
 819 forward-mode differentiation, and the two can be composed arbitrarily.  The
 820 main intended application of Autograd is gradient-based optimization.")
 821       (license license:expat))))
 822
 823 (define-public python2-autograd
 824   (package-with-python2 python-autograd))
 825
 826 (define-public lightgbm
 827   (package
 828     (name "lightgbm")
 829     (version "2.0.12")
 830     (source (origin
 831               (method url-fetch)
 832               (uri (string-append
 833                     "https://github.com/Microsoft/LightGBM/archive/v"
 834                     version ".tar.gz"))
 835               (sha256
 836                (base32
 837                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 838               (file-name (string-append name "-" version ".tar.gz"))))
 839     (native-inputs
 840      `(("python-pytest" ,python-pytest)
 841        ("python-nose" ,python-nose)))
 842     (inputs
 843      `(("openmpi" ,openmpi)))
 844     (propagated-inputs
 845      `(("python-numpy" ,python-numpy)
 846        ("python-scipy" ,python-scipy)))
 847     (arguments
 848      `(#:configure-flags
 849        '("-DUSE_MPI=ON")
 850        #:phases
 851        (modify-phases %standard-phases
 852          (replace 'check
 853            (lambda* (#:key outputs #:allow-other-keys)
 854              (with-directory-excursion ,(string-append "../LightGBM-" version)
 855                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 856     (build-system cmake-build-system)
 857     (home-page "https://github.com/Microsoft/LightGBM")
 858     (synopsis "Gradient boosting framework based on decision tree algorithms")
 859     (description "LightGBM is a gradient boosting framework that uses tree
 860 based learning algorithms.  It is designed to be distributed and efficient with
 861 the following advantages:
 862
 863 @itemize
 864 @item Faster training speed and higher efficiency
 865 @item Lower memory usage
 866 @item Better accuracy
 867 @item Parallel and GPU learning supported (not enabled in this package)
 868 @item Capable of handling large-scale data
 869 @end itemize\n")
 870     (license license:expat)))
 871
 872 (define-public vowpal-wabbit
 873   ;; Language bindings not included.
 874   (package
 875     (name "vowpal-wabbit")
 876     (version "8.5.0")
 877     (source (origin
 878               (method url-fetch)
 879               (uri (string-append
 880                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 881                     version ".tar.gz"))
 882               (sha256
 883                (base32
 884                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 885               (file-name (string-append name "-" version ".tar.gz"))))
 886     (inputs
 887      `(("boost" ,boost)
 888        ("zlib" ,zlib)))
 889     (arguments
 890      `(#:configure-flags
 891        (list (string-append "--with-boost="
 892                             (assoc-ref %build-inputs "boost")))))
 893     (build-system gnu-build-system)
 894     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 895     (synopsis "Fast machine learning library for online learning")
 896     (description "Vowpal Wabbit is a machine learning system with techniques
 897 such as online, hashing, allreduce, reductions, learning2search, active, and
 898 interactive learning.")
 899     (license license:bsd-3)))
 900
 901 (define-public python2-fastlmm
 902   (package
 903     (name "python2-fastlmm")
 904     (version "0.2.21")
 905     (source
 906      (origin
 907        (method url-fetch)
 908        (uri (pypi-uri "fastlmm" version ".zip"))
 909        (sha256
 910         (base32
 911          "1q8c34rpmwkfy3r4d5172pzdkpfryj561897z9r3x22gq7813x1m"))))
 912     (build-system python-build-system)
 913     (arguments
 914      `(#:tests? #f ; some test files are missing
 915        #:python ,python-2)) ; only Python 2.7 is supported
 916     (propagated-inputs
 917      `(("python2-numpy" ,python2-numpy)
 918        ("python2-scipy" ,python2-scipy)
 919        ("python2-matplotlib" ,python2-matplotlib)
 920        ("python2-pandas" ,python2-pandas)
 921        ("python2-scikit-learn" ,python2-scikit-learn)
 922        ("python2-pysnptools" ,python2-pysnptools)))
 923     (native-inputs
 924      `(("unzip" ,unzip)
 925        ("python2-cython" ,python2-cython)
 926        ("python2-mock" ,python2-mock)
 927        ("python2-nose" ,python2-nose)))
 928     (home-page "http://research.microsoft.com/en-us/um/redmond/projects/mscompbio/fastlmm/")
 929     (synopsis "Perform genome-wide association studies on large data sets")
 930     (description
 931      "FaST-LMM, which stands for Factored Spectrally Transformed Linear Mixed
 932 Models, is a program for performing both single-SNP and SNP-set genome-wide
 933 association studies (GWAS) on extremely large data sets.")
 934     (license license:asl2.0)))
 935
 936 ;; There have been no proper releases yet.
 937 (define-public kaldi
 938   (let ((commit "2f95609f0bb085bd3a1dc5eb0a39f3edea59e606")
 939         (revision "1"))
 940     (package
 941       (name "kaldi")
 942       (version (git-version "0" revision commit))
 943       (source (origin
 944                 (method git-fetch)
 945                 (uri (git-reference
 946                       (url "https://github.com/kaldi-asr/kaldi.git")
 947                       (commit commit)))
 948                 (file-name (git-file-name name version))
 949                 (sha256
 950                  (base32
 951                   "082qh3pfi7hvncylp4xsmkfahbd7gb0whdfa4rwrx7fxk9rdh3kz"))))
 952       (build-system gnu-build-system)
 953       (arguments
 954        `(#:test-target "test"
 955          #:phases
 956          (modify-phases %standard-phases
 957            (add-after 'unpack 'chdir
 958              (lambda _ (chdir "src") #t))
 959            (replace 'configure
 960              (lambda* (#:key build system inputs outputs #:allow-other-keys)
 961                (when (not (or (string-prefix? "x86_64" system)
 962                               (string-prefix? "i686" system)))
 963                  (substitute* "makefiles/linux_openblas.mk"
 964                    (("-msse -msse2") "")))
 965                (substitute* "makefiles/default_rules.mk"
 966                  (("/bin/bash") (which "bash")))
 967                (substitute* "Makefile"
 968                  (("ext_depend: check_portaudio")
 969                   "ext_depend:"))
 970                (substitute* '("online/Makefile"
 971                               "onlinebin/Makefile"
 972                               "gst-plugin/Makefile")
 973                  (("../../tools/portaudio/install")
 974                   (assoc-ref inputs "portaudio")))
 975
 976                ;; This `configure' script doesn't support variables passed as
 977                ;; arguments, nor does it support "prefix".
 978                (let ((out (assoc-ref outputs "out"))
 979                      (openblas (assoc-ref inputs "openblas"))
 980                      (openfst (assoc-ref inputs "openfst")))
 981                  (substitute* "configure"
 982                    (("check_for_slow_expf;") "")
 983                    ;; This affects the RPATH and also serves as the installation
 984                    ;; directory.
 985                    (("KALDILIBDIR=`pwd`/lib")
 986                     (string-append "KALDILIBDIR=" out "/lib")))
 987                  (mkdir-p out) ; must exist
 988                  (setenv "CONFIG_SHELL" (which "bash"))
 989                  (setenv "OPENFST_VER" ,(package-version openfst))
 990                  (invoke "./configure"
 991                          "--use-cuda=no"
 992                          "--shared"
 993                          (string-append "--openblas-root=" openblas)
 994                          (string-append "--fst-root=" openfst)))))
 995            (add-after 'build 'build-ext-and-gstreamer-plugin
 996              (lambda _
 997                (invoke "make" "-C" "online" "depend")
 998                (invoke "make" "-C" "online")
 999                (invoke "make" "-C" "onlinebin" "depend")
1000                (invoke "make" "-C" "onlinebin")
1001                (invoke "make" "-C" "gst-plugin" "depend")
1002                (invoke "make" "-C" "gst-plugin")
1003                #t))
1004            ;; TODO: also install the executables.
1005            (replace 'install
1006              (lambda* (#:key outputs #:allow-other-keys)
1007                (let* ((out (assoc-ref outputs "out"))
1008                       (inc (string-append out "/include"))
1009                       (lib (string-append out "/lib")))
1010                  (mkdir-p lib)
1011                  ;; The build phase installed symlinks to the actual
1012                  ;; libraries.  Install the actual targets.
1013                  (for-each (lambda (file)
1014                              (let ((target (readlink file)))
1015                                (delete-file file)
1016                                (install-file target lib)))
1017                            (find-files lib "\\.so"))
1018                  ;; Install headers
1019                  (for-each (lambda (file)
1020                              (let ((target-dir (string-append inc "/" (dirname file))))
1021                                (install-file file target-dir)))
1022                            (find-files "." "\\.h"))
1023                  (install-file "gst-plugin/libgstonlinegmmdecodefaster.so"
1024                                (string-append lib "/gstreamer-1.0"))
1025                  #t))))))
1026       (inputs
1027        `(("alsa-lib" ,alsa-lib)
1028          ("gfortran" ,gfortran "lib")
1029          ("glib" ,glib)
1030          ("gstreamer" ,gstreamer)
1031          ("jack" ,jack-1)
1032          ("openblas" ,openblas)
1033          ("openfst" ,openfst)
1034          ("portaudio" ,portaudio)
1035          ("python" ,python)))
1036       (native-inputs
1037        `(("glib" ,glib "bin")             ; glib-genmarshal
1038          ("grep" ,grep)
1039          ("sed" ,sed)
1040          ("pkg-config" ,pkg-config)
1041          ("which" ,which)))
1042       (home-page "https://kaldi-asr.org/")
1043       (synopsis "Speech recognition toolkit")
1044       (description "Kaldi is an extensible toolkit for speech recognition
1045 written in C++.")
1046       (license license:asl2.0))))
1047
1048 (define-public gst-kaldi-nnet2-online
1049   (let ((commit "617e43e73c7cc45eb9119028c02bd4178f738c4a")
1050         (revision "1"))
1051     (package
1052       (name "gst-kaldi-nnet2-online")
1053       (version (git-version "0" revision commit))
1054       (source (origin
1055                 (method git-fetch)
1056                 (uri (git-reference
1057                       (url "https://github.com/alumae/gst-kaldi-nnet2-online.git")
1058                       (commit commit)))
1059                 (file-name (git-file-name name version))
1060                 (sha256
1061                  (base32
1062                   "0xh3w67b69818s6ib02ara4lw7wamjdmh4jznvkpzrs4skbs9jx9"))))
1063       (build-system gnu-build-system)
1064       (arguments
1065        `(#:tests? #f                    ; there are none
1066          #:make-flags
1067          (list (string-append "SHELL="
1068                               (assoc-ref %build-inputs "bash") "/bin/bash")
1069                (string-append "KALDI_ROOT="
1070                               (assoc-ref %build-inputs "kaldi-src"))
1071                (string-append "KALDILIBDIR="
1072                               (assoc-ref %build-inputs "kaldi") "/lib")
1073                "KALDI_FLAVOR=dynamic")
1074          #:phases
1075          (modify-phases %standard-phases
1076            (add-after 'unpack 'chdir
1077              (lambda _ (chdir "src") #t))
1078            (replace 'configure
1079              (lambda* (#:key inputs #:allow-other-keys)
1080                (let ((glib (assoc-ref inputs "glib")))
1081                  (setenv "CXXFLAGS" "-std=c++11 -fPIC")
1082                  (setenv "CPLUS_INCLUDE_PATH"
1083                          (string-append glib "/include/glib-2.0:"
1084                                         glib "/lib/glib-2.0/include:"
1085                                         (assoc-ref inputs "gstreamer")
1086                                         "/include/gstreamer-1.0:"
1087                                         (getenv "CPLUS_INCLUDE_PATH"))))
1088                (substitute* "Makefile"
1089                  (("include \\$\\(KALDI_ROOT\\)/src/kaldi.mk") "")
1090                  (("\\$\\(error Cannot find") "#"))))
1091            (add-before 'build 'build-depend
1092              (lambda* (#:key make-flags #:allow-other-keys)
1093                (apply invoke "make" "depend" make-flags)))
1094            (replace 'install
1095              (lambda* (#:key outputs #:allow-other-keys)
1096                (let* ((out (assoc-ref outputs "out"))
1097                       (lib (string-append out "/lib/gstreamer-1.0")))
1098                  (install-file "libgstkaldinnet2onlinedecoder.so" lib)
1099                  #t))))))
1100       (inputs
1101        `(("glib" ,glib)
1102          ("gstreamer" ,gstreamer)
1103          ("jansson" ,jansson)
1104          ("openfst" ,openfst)
1105          ("kaldi" ,kaldi)))
1106       (native-inputs
1107        `(("bash" ,bash)
1108          ("glib:bin" ,glib "bin")       ; glib-genmarshal
1109          ("kaldi-src" ,(package-source kaldi))
1110          ("pkg-config" ,pkg-config)))
1111       (home-page "https://kaldi-asr.org/")
1112       (synopsis "Gstreamer plugin for decoding speech")
1113       (description "This package provides a GStreamer plugin that wraps
1114 Kaldi's @code{SingleUtteranceNnet2Decoder}.  It requires iVector-adapted DNN
1115 acoustic models.  The iVectors are adapted to the current audio stream
1116 automatically.")
1117       (license license:asl2.0))))
1118
1119 (define-public kaldi-gstreamer-server
1120   (let ((commit "1735ba49c5dc0ebfc184e45105fc600cd9f1f508")
1121         (revision "1"))
1122     (package
1123       (name "kaldi-gstreamer-server")
1124       (version (git-version "0" revision commit))
1125       (source (origin
1126                 (method git-fetch)
1127                 (uri (git-reference
1128                       (url "https://github.com/alumae/kaldi-gstreamer-server.git")
1129                       (commit commit)))
1130                 (file-name (git-file-name name version))
1131                 (sha256
1132                  (base32
1133                   "0j701m7lbwmzqxsfanj882v7881hrbmpqybbczbxqpcbg8q34w0k"))))
1134       (build-system gnu-build-system)
1135       (arguments
1136        `(#:tests? #f ; there are no tests that can be run automatically
1137          #:modules ((guix build utils)
1138                     (guix build gnu-build-system)
1139                     (srfi srfi-26))
1140          #:phases
1141          (modify-phases %standard-phases
1142            (delete 'configure)
1143            (replace 'build
1144              (lambda* (#:key outputs #:allow-other-keys)
1145                ;; Disable hash randomization to ensure the generated .pycs
1146                ;; are reproducible.
1147                (setenv "PYTHONHASHSEED" "0")
1148                (with-directory-excursion "kaldigstserver"
1149                  (for-each (lambda (file)
1150                              (apply invoke
1151                                     `("python"
1152                                       "-m" "compileall"
1153                                       "-f" ; force rebuild
1154                                       ,file)))
1155                            (find-files "." "\\.py$")))
1156                #t))
1157            (replace 'install
1158              (lambda* (#:key inputs outputs #:allow-other-keys)
1159                (let* ((out (assoc-ref outputs "out"))
1160                       (bin (string-append out "/bin"))
1161                       (share (string-append out "/share/kaldi-gstreamer-server/")))
1162                  ;; Install Python files
1163                  (with-directory-excursion "kaldigstserver"
1164                    (for-each (cut install-file <> share)
1165                              (find-files "." ".*")))
1166
1167                  ;; Install sample configuration files
1168                  (for-each (cut install-file <> share)
1169                            (find-files "." "\\.yaml"))
1170
1171                  ;; Install executables
1172                  (mkdir-p bin)
1173                  (let* ((server (string-append bin "/kaldi-gst-server"))
1174                         (client (string-append bin "/kaldi-gst-client"))
1175                         (worker (string-append bin "/kaldi-gst-worker"))
1176                         (PYTHONPATH (getenv "PYTHONPATH"))
1177                         (GST_PLUGIN_PATH (string-append
1178                                           (assoc-ref inputs "gst-kaldi-nnet2-online")
1179                                           "/lib/gstreamer-1.0:${GST_PLUGIN_PATH}"))
1180                         (wrap (lambda (wrapper what)
1181                                 (with-output-to-file wrapper
1182                                   (lambda _
1183                                     (format #t
1184                                             "#!~a
1185 export PYTHONPATH=~a
1186 export GST_PLUGIN_PATH=~a
1187 exec ~a ~a/~a \"$@\"~%"
1188                                             (which "bash") PYTHONPATH GST_PLUGIN_PATH
1189                                             (which "python") share what)))
1190                                 (chmod wrapper #o555))))
1191                    (for-each wrap
1192                              (list server client worker)
1193                              (list "master_server.py"
1194                                    "client.py"
1195                                    "worker.py")))
1196                  #t))))))
1197       (inputs
1198        `(("gst-kaldi-nnet2-online" ,gst-kaldi-nnet2-online)
1199          ("python2" ,python-2)
1200          ("python2-futures" ,python2-futures)
1201          ("python2-pygobject" ,python2-pygobject)
1202          ("python2-pyyaml" ,python2-pyyaml)
1203          ("python2-tornado" ,python2-tornado)
1204          ("python2-ws4py" ,python2-ws4py-for-kaldi-gstreamer-server)))
1205       (home-page "https://github.com/alumae/kaldi-gstreamer-server")
1206       (synopsis "Real-time full-duplex speech recognition server")
1207       (description "This is a real-time full-duplex speech recognition server,
1208 based on the Kaldi toolkit and the GStreamer framework and implemented in
1209 Python.")
1210       (license license:bsd-2))))