gnu/packages/machine-learning.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2015, 2016, 2017, 2018 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 ;;;
  13 ;;; This file is part of GNU Guix.
  14 ;;;
  15 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  16 ;;; under the terms of the GNU General Public License as published by
  17 ;;; the Free Software Foundation; either version 3 of the License, or (at
  18 ;;; your option) any later version.
  19 ;;;
  20 ;;; GNU Guix is distributed in the hope that it will be useful, but
  21 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  22 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23 ;;; GNU General Public License for more details.
  24 ;;;
  25 ;;; You should have received a copy of the GNU General Public License
  26 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  27
  28 (define-module (gnu packages machine-learning)
  29   #:use-module ((guix licenses) #:prefix license:)
  30   #:use-module (guix packages)
  31   #:use-module (guix utils)
  32   #:use-module (guix download)
  33   #:use-module (guix svn-download)
  34   #:use-module (guix build-system cmake)
  35   #:use-module (guix build-system gnu)
  36   #:use-module (guix build-system ocaml)
  37   #:use-module (guix build-system python)
  38   #:use-module (guix build-system r)
  39   #:use-module (guix git-download)
  40   #:use-module (gnu packages)
  41   #:use-module (gnu packages algebra)
  42   #:use-module (gnu packages autotools)
  43   #:use-module (gnu packages boost)
  44   #:use-module (gnu packages check)
  45   #:use-module (gnu packages compression)
  46   #:use-module (gnu packages cran)
  47   #:use-module (gnu packages dejagnu)
  48   #:use-module (gnu packages gcc)
  49   #:use-module (gnu packages image)
  50   #:use-module (gnu packages maths)
  51   #:use-module (gnu packages mpi)
  52   #:use-module (gnu packages ocaml)
  53   #:use-module (gnu packages onc-rpc)
  54   #:use-module (gnu packages perl)
  55   #:use-module (gnu packages pkg-config)
  56   #:use-module (gnu packages python)
  57   #:use-module (gnu packages statistics)
  58   #:use-module (gnu packages swig)
  59   #:use-module (gnu packages xml)
  60   #:use-module (gnu packages xorg))
  61
  62 (define-public fann
  63   ;; The last release is >100 commits behind, so we package from git.
  64   (let ((commit "d71d54788bee56ba4cf7522801270152da5209d7"))
  65     (package
  66       (name "fann")
  67       (version (string-append "2.2.0-1." (string-take commit 8)))
  68       (source (origin
  69                 (method git-fetch)
  70                 (uri (git-reference
  71                       (url "https://github.com/libfann/fann.git")
  72                       (commit commit)))
  73                 (file-name (string-append name "-" version "-checkout"))
  74                 (sha256
  75                  (base32
  76                   "0ibwpfrjs6q2lijs8slxjgzb2llcl6rk3v2ski4r6215g5jjhg3x"))))
  77       (build-system cmake-build-system)
  78       (arguments
  79        `(#:phases
  80          (modify-phases %standard-phases
  81            (replace 'check
  82              (lambda* (#:key outputs #:allow-other-keys)
  83                (let* ((out (assoc-ref outputs "out")))
  84                  (with-directory-excursion (string-append (getcwd) "/tests")
  85                    (invoke "./fann_tests"))))))))
  86       (home-page "http://leenissen.dk/fann/wp/")
  87       (synopsis "Fast Artificial Neural Network")
  88       (description
  89        "FANN is a free open source neural network library, which implements
  90 multilayer artificial neural networks in C with support for both fully
  91 connected and sparsely connected networks.")
  92       (license license:lgpl2.1))))
  93
  94 (define-public libsvm
  95   (package
  96     (name "libsvm")
  97     (version "3.22")
  98     (source
  99      (origin
 100        (method url-fetch)
 101        (uri (string-append "https://www.csie.ntu.edu.tw/~cjlin/libsvm/"
 102                            name "-" version ".tar.gz"))
 103        (sha256
 104         (base32
 105          "0zd7s19y5vb7agczl6456bn45cj1y64739sslaskw1qk7dywd0bd"))))
 106     (build-system gnu-build-system)
 107     (arguments
 108      `(#:tests? #f ;no "check" target
 109        #:phases (modify-phases %standard-phases
 110                   (delete 'configure)
 111                   (replace
 112                    'install             ; no ‘install’ target
 113                    (lambda* (#:key outputs #:allow-other-keys)
 114                      (let* ((out (assoc-ref outputs "out"))
 115                             (bin (string-append out "/bin/")))
 116                        (mkdir-p bin)
 117                        (for-each (lambda (file)
 118                                    (copy-file file (string-append bin file)))
 119                                  '("svm-train"
 120                                    "svm-predict"
 121                                    "svm-scale")))
 122                      #t)))))
 123     (home-page "http://www.csie.ntu.edu.tw/~cjlin/libsvm/")
 124     (synopsis "Library for Support Vector Machines")
 125     (description
 126      "LIBSVM is a machine learning library for support vector
 127 classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and
 128 distribution estimation (one-class SVM).  It supports multi-class
 129 classification.")
 130     (license license:bsd-3)))
 131
 132 (define-public python-libsvm
 133   (package (inherit libsvm)
 134     (name "python-libsvm")
 135     (build-system gnu-build-system)
 136     (arguments
 137      `(#:tests? #f ;no "check" target
 138        #:make-flags '("-C" "python")
 139        #:phases
 140        (modify-phases %standard-phases
 141          (delete 'configure)
 142          (replace
 143           'install                      ; no ‘install’ target
 144           (lambda* (#:key inputs outputs #:allow-other-keys)
 145             (let ((site (string-append (assoc-ref outputs "out")
 146                                        "/lib/python"
 147                                        (string-take
 148                                         (string-take-right
 149                                          (assoc-ref inputs "python") 5) 3)
 150                                        "/site-packages/")))
 151               (substitute* "python/svm.py"
 152                 (("../libsvm.so.2") "libsvm.so.2"))
 153               (mkdir-p site)
 154               (for-each (lambda (file)
 155                           (copy-file file (string-append site (basename file))))
 156                         (find-files "python" "\\.py"))
 157               (copy-file "libsvm.so.2"
 158                          (string-append site "libsvm.so.2")))
 159             #t)))))
 160     (inputs
 161      `(("python" ,python)))
 162     (synopsis "Python bindings of libSVM")))
 163
 164 (define-public ghmm
 165   ;; The latest release candidate is several years and a couple of fixes have
 166   ;; been published since.  This is why we download the sources from the SVN
 167   ;; repository.
 168   (let ((svn-revision 2341))
 169     (package
 170       (name "ghmm")
 171       (version (string-append "0.9-rc3-0." (number->string svn-revision)))
 172       (source (origin
 173                 (method svn-fetch)
 174                 (uri (svn-reference
 175                       (url "http://svn.code.sf.net/p/ghmm/code/trunk")
 176                       (revision svn-revision)))
 177                 (file-name (string-append name "-" version))
 178                 (sha256
 179                  (base32
 180                   "0qbq1rqp94l530f043qzp8aw5lj7dng9wq0miffd7spd1ff638wq"))))
 181       (build-system gnu-build-system)
 182       (arguments
 183        `(#:imported-modules (,@%gnu-build-system-modules
 184                              (guix build python-build-system))
 185          #:phases
 186          (modify-phases %standard-phases
 187            (add-after 'unpack 'enter-dir
 188              (lambda _ (chdir "ghmm") #t))
 189            (delete 'check)
 190            (add-after 'install 'check
 191              (assoc-ref %standard-phases 'check))
 192            (add-before 'check 'fix-PYTHONPATH
 193              (lambda* (#:key inputs outputs #:allow-other-keys)
 194                (let ((python-version ((@@ (guix build python-build-system)
 195                                            get-python-version)
 196                                       (assoc-ref inputs "python"))))
 197                  (setenv "PYTHONPATH"
 198                          (string-append (getenv "PYTHONPATH")
 199                                         ":" (assoc-ref outputs "out")
 200                                         "/lib/python" python-version
 201                                         "/site-packages")))
 202                #t))
 203            (add-after 'enter-dir 'fix-runpath
 204              (lambda* (#:key outputs #:allow-other-keys)
 205                (substitute* "ghmmwrapper/setup.py"
 206                  (("^(.*)extra_compile_args = \\[" line indent)
 207                   (string-append indent
 208                                  "extra_link_args = [\"-Wl,-rpath="
 209                                  (assoc-ref outputs "out") "/lib\"],\n"
 210                                  line
 211                                  "\"-Wl,-rpath="
 212                                  (assoc-ref outputs "out")
 213                                  "/lib\", ")))
 214                #t))
 215            (add-after 'enter-dir 'disable-broken-tests
 216              (lambda _
 217                (substitute* "tests/Makefile.am"
 218                  ;; GHMM_SILENT_TESTS is assumed to be a command.
 219                  (("TESTS_ENVIRONMENT.*") "")
 220                  ;; Do not build broken tests.
 221                  (("chmm .*") "")
 222                  (("read_fa .*") "")
 223                  (("mcmc .*") "")
 224                  (("label_higher_order_test.*$")
 225                   "label_higher_order_test\n"))
 226
 227                ;; These Python unittests are broken as there is no gato.
 228                ;; See https://sourceforge.net/p/ghmm/support-requests/3/
 229                (substitute* "ghmmwrapper/ghmmunittests.py"
 230                  (("^(.*)def (testNewXML|testMultipleTransitionClasses|testNewXML)"
 231                    line indent)
 232                   (string-append indent
 233                                  "@unittest.skip(\"Disabled by Guix\")\n"
 234                                  line)))
 235                #t))
 236            (add-after 'disable-broken-tests 'autogen
 237              (lambda _
 238                (invoke "bash" "autogen.sh"))))))
 239       (inputs
 240        `(("python" ,python-2) ; only Python 2 is supported
 241          ("libxml2" ,libxml2)))
 242       (native-inputs
 243        `(("pkg-config" ,pkg-config)
 244          ("dejagnu" ,dejagnu)
 245          ("swig" ,swig)
 246          ("autoconf" ,autoconf)
 247          ("automake" ,automake)
 248          ("libtool" ,libtool)))
 249       (home-page "http://ghmm.org")
 250       (synopsis "Hidden Markov Model library")
 251       (description
 252        "The General Hidden Markov Model library (GHMM) is a C library with
 253 additional Python bindings implementing a wide range of types of @dfn{Hidden
 254 Markov Models} (HMM) and algorithms: discrete, continuous emissions, basic
 255 training, HMM clustering, HMM mixtures.")
 256       (license license:lgpl2.0+))))
 257
 258 (define-public mcl
 259   (package
 260     (name "mcl")
 261     (version "14.137")
 262     (source (origin
 263               (method url-fetch)
 264               (uri (string-append
 265                     "http://micans.org/mcl/src/mcl-"
 266                     (string-replace-substring version "." "-")
 267                     ".tar.gz"))
 268               (sha256
 269                (base32
 270                 "15xlax3z31lsn62vlg94hkm75nm40q4679amnfg13jm8m2bnhy5m"))))
 271     (build-system gnu-build-system)
 272     (arguments
 273      `(#:configure-flags (list "--enable-blast")))
 274     (inputs
 275      `(("perl" ,perl)))
 276     (home-page "http://micans.org/mcl/")
 277     (synopsis "Clustering algorithm for graphs")
 278     (description
 279      "The MCL algorithm is short for the @dfn{Markov Cluster Algorithm}, a
 280 fast and scalable unsupervised cluster algorithm for graphs (also known as
 281 networks) based on simulation of (stochastic) flow in graphs.")
 282     ;; In the LICENCE file and web page it says "The software is licensed
 283     ;; under the GNU General Public License, version 3.", but in several of
 284     ;; the source code files it suggests GPL3 or later.
 285     ;; http://listserver.ebi.ac.uk/pipermail/mcl-users/2016/000376.html
 286     (license license:gpl3)))
 287
 288 (define-public ocaml-mcl
 289   (package
 290     (name "ocaml-mcl")
 291     (version "12-068oasis4")
 292     (source
 293      (origin
 294        (method url-fetch)
 295        (uri (string-append
 296              "https://github.com/fhcrc/mcl/archive/"
 297              version ".tar.gz"))
 298        (file-name (string-append name "-" version ".tar.gz"))
 299        (sha256
 300         (base32
 301          "1l5jbhwjpsj38x8b9698hfpkv75h8hn3kj0gihjhn8ym2cwwv110"))))
 302     (build-system ocaml-build-system)
 303     (arguments
 304      `(#:ocaml ,ocaml-4.02
 305        #:findlib ,ocaml4.02-findlib
 306        #:phases
 307        (modify-phases %standard-phases
 308          (add-before 'configure 'patch-paths
 309            (lambda _
 310              (substitute* "configure"
 311                (("SHELL = /bin/sh") (string-append "SHELL = "(which "sh"))))
 312              (substitute* "setup.ml"
 313                (("LDFLAGS=-fPIC")
 314                 (string-append "LDFLAGS=-fPIC\"; \"SHELL=" (which "sh"))))
 315              #t)))))
 316     (home-page "https://github.com/fhcrc/mcl")
 317     (synopsis "OCaml wrappers around MCL")
 318     (description
 319      "This package provides OCaml bindings for the MCL graph clustering
 320 algorithm.")
 321     (license license:gpl3)))
 322
 323 (define-public ocaml4.01-mcl
 324   (package-with-ocaml4.01 ocaml-mcl))
 325
 326 (define-public randomjungle
 327   (package
 328     (name "randomjungle")
 329     (version "2.1.0")
 330     (source
 331      (origin
 332        (method url-fetch)
 333        (uri (string-append
 334              "http://www.imbs-luebeck.de/imbs/sites/default/files/u59/"
 335              "randomjungle-" version ".tar_.gz"))
 336        (sha256
 337         (base32
 338          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 339     (build-system gnu-build-system)
 340     (arguments
 341      `(#:configure-flags
 342        (list (string-append "--with-boost="
 343                             (assoc-ref %build-inputs "boost")))
 344        #:phases
 345        (modify-phases %standard-phases
 346          (add-before
 347           'configure 'set-CXXFLAGS
 348           (lambda _
 349             (setenv "CXXFLAGS" "-fpermissive ")
 350             #t)))))
 351     (inputs
 352      `(("boost" ,boost)
 353        ("gsl" ,gsl)
 354        ("libxml2" ,libxml2)
 355        ("zlib" ,zlib)))
 356     (native-inputs
 357      `(("gfortran" ,gfortran)
 358        ("gfortran:lib" ,gfortran "lib")))
 359     ;; Non-portable assembly instructions are used so building fails on
 360     ;; platforms other than x86_64 or i686.
 361     (supported-systems '("x86_64-linux" "i686-linux"))
 362     (home-page "http://www.imbs-luebeck.de/imbs/de/node/227/")
 363     (synopsis "Implementation of the Random Forests machine learning method")
 364     (description
 365      "Random Jungle is an implementation of Random Forests.  It is supposed to
 366 analyse high dimensional data.  In genetics, it can be used for analysing big
 367 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 368 learning method.  Most interesting features are variable selection, missing
 369 value imputation, classifier creation, generalization error estimation and
 370 sample proximities between pairs of cases.")
 371     (license license:gpl3+)))
 372
 373 (define-public shogun
 374   (package
 375     (name "shogun")
 376     (version "6.1.3")
 377     (source
 378      (origin
 379        (method url-fetch)
 380        (uri (string-append
 381              "ftp://shogun-toolbox.org/shogun/releases/"
 382              (version-major+minor version)
 383              "/sources/shogun-" version ".tar.bz2"))
 384        (sha256
 385         (base32
 386          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 387        (modules '((guix build utils)
 388                   (ice-9 rdelim)))
 389        (snippet
 390         '(begin
 391            ;; Remove non-free sources and files referencing them
 392            (for-each delete-file
 393                      (find-files "src/shogun/classifier/svm/"
 394                                  "SVMLight\\.(cpp|h)"))
 395            (for-each delete-file
 396                      (find-files "examples/undocumented/libshogun/"
 397                                  (string-append
 398                                   "(classifier_.*svmlight.*|"
 399                                   "evaluation_cross_validation_locked_comparison).cpp")))
 400            ;; Remove non-free functions.
 401            (define (delete-ifdefs file)
 402              (with-atomic-file-replacement file
 403                (lambda (in out)
 404                  (let loop ((line (read-line in 'concat))
 405                             (skipping? #f))
 406                    (if (eof-object? line)
 407                        #t
 408                        (let ((skip-next?
 409                               (or (and skipping?
 410                                        (not (string-prefix?
 411                                              "#endif //USE_SVMLIGHT" line)))
 412                                   (string-prefix?
 413                                    "#ifdef USE_SVMLIGHT" line))))
 414                          (when (or (not skipping?)
 415                                    (and skipping? (not skip-next?)))
 416                            (display line out))
 417                          (loop (read-line in 'concat) skip-next?)))))))
 418            (for-each delete-ifdefs
 419                      (append
 420                       (find-files "src/shogun/classifier/mkl"
 421                                   "^MKLClassification\\.cpp")
 422                       (find-files "src/shogun/classifier/svm"
 423                                   "^SVMLightOneClass\\.(cpp|h)")
 424                       (find-files "src/shogun/multiclass"
 425                                   "^ScatterSVM\\.(cpp|h)")
 426                       (find-files "src/shogun/kernel/"
 427                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 428                       (find-files "src/shogun/regression/svr"
 429                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 430                       (find-files "src/shogun/transfer/domain_adaptation"
 431                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 432            #t))))
 433     (build-system cmake-build-system)
 434     (arguments
 435      '(#:tests? #f ;no check target
 436        #:phases
 437        (modify-phases %standard-phases
 438          (add-after 'unpack 'delete-broken-symlinks
 439            (lambda _
 440              (for-each delete-file '("applications/arts/data"
 441                                      "applications/asp/data"
 442                                      "applications/easysvm/data"
 443                                      "applications/msplicer/data"
 444                                      "applications/ocr/data"
 445                                      "examples/meta/data"
 446                                      "examples/undocumented/data"))
 447              #t))
 448          (add-after 'unpack 'change-R-target-path
 449            (lambda* (#:key outputs #:allow-other-keys)
 450              (substitute* '("src/interfaces/r/CMakeLists.txt"
 451                             "examples/meta/r/CMakeLists.txt")
 452                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 453                 (string-append (assoc-ref outputs "out")
 454                                "/lib/R/library/")))
 455              #t))
 456          (add-after 'unpack 'fix-octave-modules
 457            (lambda* (#:key outputs #:allow-other-keys)
 458              (substitute* "src/interfaces/octave/CMakeLists.txt"
 459                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 460                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 461                ;; change target directory
 462                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 463                 (string-append (assoc-ref outputs "out")
 464                                "/share/octave/packages")))
 465              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 466                             "src/interfaces/octave/sg_print_functions.cpp")
 467                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 468                (("octave/config\\.h") "octave/octave-config.h")
 469                (("octave/oct-obj.h") "octave/ovl.h"))
 470              #t))
 471          (add-after 'unpack 'move-rxcpp
 472            (lambda* (#:key inputs #:allow-other-keys)
 473              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 474                (mkdir-p rxcpp-dir)
 475                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 476                #t)))
 477          (add-before 'build 'set-HOME
 478            ;; $HOME needs to be set at some point during the build phase
 479            (lambda _ (setenv "HOME" "/tmp") #t)))
 480        #:configure-flags
 481        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 482              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 483              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 484              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 485              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 486              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 487              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 488              "-DINTERFACE_OCTAVE=ON"
 489              "-DINTERFACE_PYTHON=ON"
 490              "-DINTERFACE_R=ON")))
 491     (inputs
 492      `(("python" ,python)
 493        ("numpy" ,python-numpy)
 494        ("r-minimal" ,r-minimal)
 495        ("octave" ,octave)
 496        ("swig" ,swig)
 497        ("eigen" ,eigen)
 498        ("hdf5" ,hdf5)
 499        ("atlas" ,atlas)
 500        ("arpack" ,arpack-ng)
 501        ("lapack" ,lapack)
 502        ("glpk" ,glpk)
 503        ("libxml2" ,libxml2)
 504        ("lzo" ,lzo)
 505        ("zlib" ,zlib)))
 506     (native-inputs
 507      `(("pkg-config" ,pkg-config)
 508        ("rxcpp" ,rxcpp)))
 509     ;; Non-portable SSE instructions are used so building fails on platforms
 510     ;; other than x86_64.
 511     (supported-systems '("x86_64-linux"))
 512     (home-page "http://shogun-toolbox.org/")
 513     (synopsis "Machine learning toolbox")
 514     (description
 515      "The Shogun Machine learning toolbox provides a wide range of unified and
 516 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 517 combine multiple data representations, algorithm classes, and general purpose
 518 tools.  This enables both rapid prototyping of data pipelines and extensibility
 519 in terms of new algorithms.")
 520     (license license:gpl3+)))
 521
 522 (define-public rxcpp
 523   (package
 524     (name "rxcpp")
 525     (version "4.0.0")
 526     (source
 527      (origin
 528        (method url-fetch)
 529        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 530                            version ".tar.gz"))
 531        (sha256
 532         (base32
 533          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 534        (file-name (string-append name "-" version ".tar.gz"))))
 535     (build-system cmake-build-system)
 536     (arguments
 537      `(#:phases
 538        (modify-phases %standard-phases
 539          (add-after 'unpack 'remove-werror
 540            (lambda _
 541              (substitute* (find-files ".")
 542                (("-Werror") ""))
 543              #t))
 544          (replace 'check
 545            (lambda _
 546              (invoke "ctest"))))))
 547     (native-inputs
 548      `(("catch" ,catch-framework)))
 549     (home-page "http://reactivex.io/")
 550     (synopsis "Reactive Extensions for C++")
 551     (description
 552      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 553 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 554 and event-based programs by using observable sequences.
 555
 556 It extends the observer pattern to support sequences of data and/or events and
 557 adds operators that allow you to compose sequences together declaratively while
 558 abstracting away concerns about things like low-level threading,
 559 synchronization, thread-safety, concurrent data structures, and non-blocking
 560 I/O.")
 561     (license license:asl2.0)))
 562
 563 (define-public r-adaptivesparsity
 564   (package
 565     (name "r-adaptivesparsity")
 566     (version "1.6")
 567     (source (origin
 568               (method url-fetch)
 569               (uri (cran-uri "AdaptiveSparsity" version))
 570               (sha256
 571                (base32
 572                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 573     (properties
 574      `((upstream-name . "AdaptiveSparsity")))
 575     (build-system r-build-system)
 576     (arguments
 577      `(#:phases
 578        (modify-phases %standard-phases
 579          (add-after 'unpack 'link-against-armadillo
 580            (lambda _
 581              (substitute* "src/Makevars"
 582                (("PKG_LIBS=" prefix)
 583                 (string-append prefix "-larmadillo"))))))))
 584     (propagated-inputs
 585      `(("r-mass" ,r-mass)
 586        ("r-matrix" ,r-matrix)
 587        ("r-rcpp" ,r-rcpp)
 588        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 589     (inputs
 590      `(("armadillo" ,armadillo)))
 591     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 592     (synopsis "Adaptive sparsity models")
 593     (description
 594      "This package implements the Figueiredo machine learning algorithm for
 595 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 596 geometric models.")
 597     (license license:lgpl3+)))
 598
 599 (define-public r-kernlab
 600   (package
 601     (name "r-kernlab")
 602     (version "0.9-27")
 603     (source
 604      (origin
 605        (method url-fetch)
 606        (uri (cran-uri "kernlab" version))
 607        (sha256
 608         (base32
 609          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 610     (build-system r-build-system)
 611     (home-page "https://cran.r-project.org/web/packages/kernlab")
 612     (synopsis "Kernel-based machine learning tools")
 613     (description
 614      "This package provides kernel-based machine learning methods for
 615 classification, regression, clustering, novelty detection, quantile regression
 616 and dimensionality reduction.  Among other methods @code{kernlab} includes
 617 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 618 and a QP solver.")
 619     (license license:gpl2)))
 620
 621 (define-public dlib
 622   (package
 623     (name "dlib")
 624     (version "19.7")
 625     (source (origin
 626               (method url-fetch)
 627               (uri (string-append
 628                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 629               (sha256
 630                (base32
 631                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 632               (modules '((guix build utils)))
 633               (snippet
 634                '(begin
 635                   ;; Delete ~13MB of bundled dependencies.
 636                   (delete-file-recursively "dlib/external")
 637                   (delete-file-recursively "docs/dlib/external")
 638                   #t))))
 639     (build-system cmake-build-system)
 640     (arguments
 641      `(#:phases
 642        (modify-phases %standard-phases
 643          (add-after 'unpack 'disable-asserts
 644            (lambda _
 645              ;; config.h recommends explicitly enabling or disabling asserts
 646              ;; when building as a shared library. By default neither is set.
 647              (substitute* "dlib/config.h"
 648                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 649              #t))
 650          (add-after 'disable-asserts 'disable-failing-tests
 651            (lambda _
 652              ;; One test times out on MIPS, so we need to disable it.
 653              ;; Others are flaky on some platforms.
 654              (let* ((system ,(or (%current-target-system)
 655                                  (%current-system)))
 656                     (disabled-tests (cond
 657                                      ((string-prefix? "mips64" system)
 658                                       '("object_detector" ; timeout
 659                                         "data_io"))
 660                                      ((string-prefix? "armhf" system)
 661                                       '("learning_to_track"))
 662                                      ((string-prefix? "i686" system)
 663                                       '("optimization"))
 664                                      (else '()))))
 665                (for-each
 666                 (lambda (test)
 667                   (substitute* "dlib/test/makefile"
 668                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 669                 disabled-tests)
 670                #t)))
 671          (replace 'check
 672            (lambda _
 673              ;; No test target, so we build and run the unit tests here.
 674              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 675                (with-directory-excursion test-dir
 676                  (invoke "make" "-j" (number->string (parallel-job-count)))
 677                  (invoke "./dtest" "--runall"))
 678                #t)))
 679          (add-after 'install 'delete-static-library
 680            (lambda* (#:key outputs #:allow-other-keys)
 681              (delete-file (string-append (assoc-ref outputs "out")
 682                                          "/lib/libdlib.a"))
 683              #t)))))
 684     (native-inputs
 685      `(("pkg-config" ,pkg-config)
 686        ;; For tests.
 687        ("libnsl" ,libnsl)))
 688     (inputs
 689      `(("giflib" ,giflib)
 690        ("lapack" ,lapack)
 691        ("libjpeg" ,libjpeg)
 692        ("libpng" ,libpng)
 693        ("libx11" ,libx11)
 694        ("openblas" ,openblas)
 695        ("zlib" ,zlib)))
 696     (synopsis
 697      "Toolkit for making machine learning and data analysis applications in C++")
 698     (description
 699      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 700 tools.  It is used in both industry and academia in a wide range of domains
 701 including robotics, embedded devices, mobile phones, and large high performance
 702 computing environments.")
 703     (home-page "http://dlib.net")
 704     (license license:boost1.0)))
 705
 706 (define-public python-scikit-learn
 707   (package
 708     (name "python-scikit-learn")
 709     (version "0.20.1")
 710     (source
 711      (origin
 712        (method git-fetch)
 713        (uri (git-reference
 714              (url "https://github.com/scikit-learn/scikit-learn.git")
 715              (commit version)))
 716        (file-name (git-file-name name version))
 717        (sha256
 718         (base32
 719          "0qv7ir1fy9vjar3llc72yxmfja3gxm5icdf0y3q57vsn3wcdglkz"))))
 720     (build-system python-build-system)
 721     (arguments
 722      `(#:phases
 723        (modify-phases %standard-phases
 724          (add-after 'build 'build-ext
 725            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 726          (replace 'check
 727            (lambda _
 728              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 729              (setenv "OPENBLAS_NUM_THREADS" "1")
 730
 731              ;; Some tests require write access to $HOME.
 732              (setenv "HOME" "/tmp")
 733
 734              (invoke "pytest" "sklearn" "-m" "not network")))
 735          ;; FIXME: This fails with permission denied
 736          (delete 'reset-gzip-timestamps))))
 737     (inputs
 738      `(("openblas" ,openblas)))
 739     (native-inputs
 740      `(("python-pytest" ,python-pytest)
 741        ("python-pandas" ,python-pandas) ;for tests
 742        ("python-cython" ,python-cython)))
 743     (propagated-inputs
 744      `(("python-numpy" ,python-numpy)
 745        ("python-scipy" ,python-scipy)))
 746     (home-page "http://scikit-learn.org/")
 747     (synopsis "Machine Learning in Python")
 748     (description
 749      "Scikit-learn provides simple and efficient tools for data mining and
 750 data analysis.")
 751     (license license:bsd-3)))
 752
 753 (define-public python2-scikit-learn
 754   (package-with-python2 python-scikit-learn))
 755
 756 (define-public python-autograd
 757   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 758          (revision "0")
 759          (version (git-version "0.0.0" revision commit)))
 760     (package
 761       (name "python-autograd")
 762       (home-page "https://github.com/HIPS/autograd")
 763       (source (origin
 764                 (method git-fetch)
 765                 (uri (git-reference
 766                       (url home-page)
 767                       (commit commit)))
 768                 (sha256
 769                  (base32
 770                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 771                 (file-name (git-file-name name version))))
 772       (version version)
 773       (build-system python-build-system)
 774       (native-inputs
 775        `(("python-nose" ,python-nose)
 776          ("python-pytest" ,python-pytest)))
 777       (propagated-inputs
 778        `(("python-future" ,python-future)
 779          ("python-numpy" ,python-numpy)))
 780       (arguments
 781        `(#:phases (modify-phases %standard-phases
 782                     (replace 'check
 783                       (lambda _
 784                         (invoke "py.test" "-v"))))))
 785       (synopsis "Efficiently computes derivatives of NumPy code")
 786       (description "Autograd can automatically differentiate native Python and
 787 NumPy code.  It can handle a large subset of Python's features, including loops,
 788 ifs, recursion and closures, and it can even take derivatives of derivatives
 789 of derivatives.  It supports reverse-mode differentiation
 790 (a.k.a. backpropagation), which means it can efficiently take gradients of
 791 scalar-valued functions with respect to array-valued arguments, as well as
 792 forward-mode differentiation, and the two can be composed arbitrarily.  The
 793 main intended application of Autograd is gradient-based optimization.")
 794       (license license:expat))))
 795
 796 (define-public python2-autograd
 797   (package-with-python2 python-autograd))
 798
 799 (define-public lightgbm
 800   (package
 801     (name "lightgbm")
 802     (version "2.0.12")
 803     (source (origin
 804               (method url-fetch)
 805               (uri (string-append
 806                     "https://github.com/Microsoft/LightGBM/archive/v"
 807                     version ".tar.gz"))
 808               (sha256
 809                (base32
 810                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 811               (file-name (string-append name "-" version ".tar.gz"))))
 812     (native-inputs
 813      `(("python-pytest" ,python-pytest)
 814        ("python-nose" ,python-nose)))
 815     (inputs
 816      `(("openmpi" ,openmpi)))
 817     (propagated-inputs
 818      `(("python-numpy" ,python-numpy)
 819        ("python-scipy" ,python-scipy)))
 820     (arguments
 821      `(#:configure-flags
 822        '("-DUSE_MPI=ON")
 823        #:phases
 824        (modify-phases %standard-phases
 825          (replace 'check
 826            (lambda* (#:key outputs #:allow-other-keys)
 827              (with-directory-excursion ,(string-append "../LightGBM-" version)
 828                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 829     (build-system cmake-build-system)
 830     (home-page "https://github.com/Microsoft/LightGBM")
 831     (synopsis "Gradient boosting framework based on decision tree algorithms")
 832     (description "LightGBM is a gradient boosting framework that uses tree
 833 based learning algorithms.  It is designed to be distributed and efficient with
 834 the following advantages:
 835
 836 @itemize
 837 @item Faster training speed and higher efficiency
 838 @item Lower memory usage
 839 @item Better accuracy
 840 @item Parallel and GPU learning supported (not enabled in this package)
 841 @item Capable of handling large-scale data
 842 @end itemize\n")
 843     (license license:expat)))
 844
 845 (define-public vowpal-wabbit
 846   ;; Language bindings not included.
 847   (package
 848     (name "vowpal-wabbit")
 849     (version "8.5.0")
 850     (source (origin
 851               (method url-fetch)
 852               (uri (string-append
 853                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 854                     version ".tar.gz"))
 855               (sha256
 856                (base32
 857                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 858               (file-name (string-append name "-" version ".tar.gz"))))
 859     (inputs
 860      `(("boost" ,boost)
 861        ("zlib" ,zlib)))
 862     (arguments
 863      `(#:configure-flags
 864        (list (string-append "--with-boost="
 865                             (assoc-ref %build-inputs "boost")))))
 866     (build-system gnu-build-system)
 867     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 868     (synopsis "Fast machine learning library for online learning")
 869     (description "Vowpal Wabbit is a machine learning system with techniques
 870 such as online, hashing, allreduce, reductions, learning2search, active, and
 871 interactive learning.")
 872     (license license:bsd-3)))