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 ;;; 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 autotools)
  44   #:use-module (gnu packages boost)
  45   #:use-module (gnu packages check)
  46   #:use-module (gnu packages compression)
  47   #:use-module (gnu packages cran)
  48   #:use-module (gnu packages dejagnu)
  49   #:use-module (gnu packages gcc)
  50   #:use-module (gnu packages image)
  51   #:use-module (gnu packages maths)
  52   #:use-module (gnu packages mpi)
  53   #:use-module (gnu packages ocaml)
  54   #:use-module (gnu packages onc-rpc)
  55   #:use-module (gnu packages perl)
  56   #:use-module (gnu packages pkg-config)
  57   #:use-module (gnu packages python)
  58   #:use-module (gnu packages statistics)
  59   #:use-module (gnu packages swig)
  60   #:use-module (gnu packages xml)
  61   #:use-module (gnu packages xorg))
  62
  63 (define-public fann
  64   ;; The last release is >100 commits behind, so we package from git.
  65   (let ((commit "d71d54788bee56ba4cf7522801270152da5209d7"))
  66     (package
  67       (name "fann")
  68       (version (string-append "2.2.0-1." (string-take commit 8)))
  69       (source (origin
  70                 (method git-fetch)
  71                 (uri (git-reference
  72                       (url "https://github.com/libfann/fann.git")
  73                       (commit commit)))
  74                 (file-name (string-append name "-" version "-checkout"))
  75                 (sha256
  76                  (base32
  77                   "0ibwpfrjs6q2lijs8slxjgzb2llcl6rk3v2ski4r6215g5jjhg3x"))))
  78       (build-system cmake-build-system)
  79       (arguments
  80        `(#:phases
  81          (modify-phases %standard-phases
  82            (replace 'check
  83              (lambda* (#:key outputs #:allow-other-keys)
  84                (let* ((out (assoc-ref outputs "out")))
  85                  (with-directory-excursion (string-append (getcwd) "/tests")
  86                    (invoke "./fann_tests"))))))))
  87       (home-page "http://leenissen.dk/fann/wp/")
  88       (synopsis "Fast Artificial Neural Network")
  89       (description
  90        "FANN is a free open source neural network library, which implements
  91 multilayer artificial neural networks in C with support for both fully
  92 connected and sparsely connected networks.")
  93       (license license:lgpl2.1))))
  94
  95 (define-public libsvm
  96   (package
  97     (name "libsvm")
  98     (version "3.22")
  99     (source
 100      (origin
 101        (method url-fetch)
 102        (uri (string-append "https://www.csie.ntu.edu.tw/~cjlin/libsvm/"
 103                            name "-" version ".tar.gz"))
 104        (sha256
 105         (base32
 106          "0zd7s19y5vb7agczl6456bn45cj1y64739sslaskw1qk7dywd0bd"))))
 107     (build-system gnu-build-system)
 108     (arguments
 109      `(#:tests? #f ;no "check" target
 110        #:phases (modify-phases %standard-phases
 111                   (delete 'configure)
 112                   (replace
 113                    'install             ; no ‘install’ target
 114                    (lambda* (#:key outputs #:allow-other-keys)
 115                      (let* ((out (assoc-ref outputs "out"))
 116                             (bin (string-append out "/bin/")))
 117                        (mkdir-p bin)
 118                        (for-each (lambda (file)
 119                                    (copy-file file (string-append bin file)))
 120                                  '("svm-train"
 121                                    "svm-predict"
 122                                    "svm-scale")))
 123                      #t)))))
 124     (home-page "http://www.csie.ntu.edu.tw/~cjlin/libsvm/")
 125     (synopsis "Library for Support Vector Machines")
 126     (description
 127      "LIBSVM is a machine learning library for support vector
 128 classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and
 129 distribution estimation (one-class SVM).  It supports multi-class
 130 classification.")
 131     (license license:bsd-3)))
 132
 133 (define-public python-libsvm
 134   (package (inherit libsvm)
 135     (name "python-libsvm")
 136     (build-system gnu-build-system)
 137     (arguments
 138      `(#:tests? #f ;no "check" target
 139        #:make-flags '("-C" "python")
 140        #:phases
 141        (modify-phases %standard-phases
 142          (delete 'configure)
 143          (replace
 144           'install                      ; no ‘install’ target
 145           (lambda* (#:key inputs outputs #:allow-other-keys)
 146             (let ((site (string-append (assoc-ref outputs "out")
 147                                        "/lib/python"
 148                                        (string-take
 149                                         (string-take-right
 150                                          (assoc-ref inputs "python") 5) 3)
 151                                        "/site-packages/")))
 152               (substitute* "python/svm.py"
 153                 (("../libsvm.so.2") "libsvm.so.2"))
 154               (mkdir-p site)
 155               (for-each (lambda (file)
 156                           (copy-file file (string-append site (basename file))))
 157                         (find-files "python" "\\.py"))
 158               (copy-file "libsvm.so.2"
 159                          (string-append site "libsvm.so.2")))
 160             #t)))))
 161     (inputs
 162      `(("python" ,python)))
 163     (synopsis "Python bindings of libSVM")))
 164
 165 (define-public ghmm
 166   ;; The latest release candidate is several years and a couple of fixes have
 167   ;; been published since.  This is why we download the sources from the SVN
 168   ;; repository.
 169   (let ((svn-revision 2341))
 170     (package
 171       (name "ghmm")
 172       (version (string-append "0.9-rc3-0." (number->string svn-revision)))
 173       (source (origin
 174                 (method svn-fetch)
 175                 (uri (svn-reference
 176                       (url "http://svn.code.sf.net/p/ghmm/code/trunk")
 177                       (revision svn-revision)))
 178                 (file-name (string-append name "-" version))
 179                 (sha256
 180                  (base32
 181                   "0qbq1rqp94l530f043qzp8aw5lj7dng9wq0miffd7spd1ff638wq"))))
 182       (build-system gnu-build-system)
 183       (arguments
 184        `(#:imported-modules (,@%gnu-build-system-modules
 185                              (guix build python-build-system))
 186          #:phases
 187          (modify-phases %standard-phases
 188            (add-after 'unpack 'enter-dir
 189              (lambda _ (chdir "ghmm") #t))
 190            (delete 'check)
 191            (add-after 'install 'check
 192              (assoc-ref %standard-phases 'check))
 193            (add-before 'check 'fix-PYTHONPATH
 194              (lambda* (#:key inputs outputs #:allow-other-keys)
 195                (let ((python-version ((@@ (guix build python-build-system)
 196                                            get-python-version)
 197                                       (assoc-ref inputs "python"))))
 198                  (setenv "PYTHONPATH"
 199                          (string-append (getenv "PYTHONPATH")
 200                                         ":" (assoc-ref outputs "out")
 201                                         "/lib/python" python-version
 202                                         "/site-packages")))
 203                #t))
 204            (add-after 'enter-dir 'fix-runpath
 205              (lambda* (#:key outputs #:allow-other-keys)
 206                (substitute* "ghmmwrapper/setup.py"
 207                  (("^(.*)extra_compile_args = \\[" line indent)
 208                   (string-append indent
 209                                  "extra_link_args = [\"-Wl,-rpath="
 210                                  (assoc-ref outputs "out") "/lib\"],\n"
 211                                  line
 212                                  "\"-Wl,-rpath="
 213                                  (assoc-ref outputs "out")
 214                                  "/lib\", ")))
 215                #t))
 216            (add-after 'enter-dir 'disable-broken-tests
 217              (lambda _
 218                (substitute* "tests/Makefile.am"
 219                  ;; GHMM_SILENT_TESTS is assumed to be a command.
 220                  (("TESTS_ENVIRONMENT.*") "")
 221                  ;; Do not build broken tests.
 222                  (("chmm .*") "")
 223                  (("read_fa .*") "")
 224                  (("mcmc .*") "")
 225                  (("label_higher_order_test.*$")
 226                   "label_higher_order_test\n"))
 227
 228                ;; These Python unittests are broken as there is no gato.
 229                ;; See https://sourceforge.net/p/ghmm/support-requests/3/
 230                (substitute* "ghmmwrapper/ghmmunittests.py"
 231                  (("^(.*)def (testNewXML|testMultipleTransitionClasses|testNewXML)"
 232                    line indent)
 233                   (string-append indent
 234                                  "@unittest.skip(\"Disabled by Guix\")\n"
 235                                  line)))
 236                #t))
 237            (add-after 'disable-broken-tests 'autogen
 238              (lambda _
 239                (invoke "bash" "autogen.sh"))))))
 240       (inputs
 241        `(("python" ,python-2) ; only Python 2 is supported
 242          ("libxml2" ,libxml2)))
 243       (native-inputs
 244        `(("pkg-config" ,pkg-config)
 245          ("dejagnu" ,dejagnu)
 246          ("swig" ,swig)
 247          ("autoconf" ,autoconf)
 248          ("automake" ,automake)
 249          ("libtool" ,libtool)))
 250       (home-page "http://ghmm.org")
 251       (synopsis "Hidden Markov Model library")
 252       (description
 253        "The General Hidden Markov Model library (GHMM) is a C library with
 254 additional Python bindings implementing a wide range of types of @dfn{Hidden
 255 Markov Models} (HMM) and algorithms: discrete, continuous emissions, basic
 256 training, HMM clustering, HMM mixtures.")
 257       (license license:lgpl2.0+))))
 258
 259 (define-public mcl
 260   (package
 261     (name "mcl")
 262     (version "14.137")
 263     (source (origin
 264               (method url-fetch)
 265               (uri (string-append
 266                     "http://micans.org/mcl/src/mcl-"
 267                     (string-replace-substring version "." "-")
 268                     ".tar.gz"))
 269               (sha256
 270                (base32
 271                 "15xlax3z31lsn62vlg94hkm75nm40q4679amnfg13jm8m2bnhy5m"))))
 272     (build-system gnu-build-system)
 273     (arguments
 274      `(#:configure-flags (list "--enable-blast")))
 275     (inputs
 276      `(("perl" ,perl)))
 277     (home-page "http://micans.org/mcl/")
 278     (synopsis "Clustering algorithm for graphs")
 279     (description
 280      "The MCL algorithm is short for the @dfn{Markov Cluster Algorithm}, a
 281 fast and scalable unsupervised cluster algorithm for graphs (also known as
 282 networks) based on simulation of (stochastic) flow in graphs.")
 283     ;; In the LICENCE file and web page it says "The software is licensed
 284     ;; under the GNU General Public License, version 3.", but in several of
 285     ;; the source code files it suggests GPL3 or later.
 286     ;; http://listserver.ebi.ac.uk/pipermail/mcl-users/2016/000376.html
 287     (license license:gpl3)))
 288
 289 (define-public ocaml-mcl
 290   (package
 291     (name "ocaml-mcl")
 292     (version "12-068oasis4")
 293     (source
 294      (origin
 295        (method url-fetch)
 296        (uri (string-append
 297              "https://github.com/fhcrc/mcl/archive/"
 298              version ".tar.gz"))
 299        (file-name (string-append name "-" version ".tar.gz"))
 300        (sha256
 301         (base32
 302          "1l5jbhwjpsj38x8b9698hfpkv75h8hn3kj0gihjhn8ym2cwwv110"))))
 303     (build-system ocaml-build-system)
 304     (arguments
 305      `(#:ocaml ,ocaml-4.02
 306        #:findlib ,ocaml4.02-findlib
 307        #:phases
 308        (modify-phases %standard-phases
 309          (add-before 'configure 'patch-paths
 310            (lambda _
 311              (substitute* "configure"
 312                (("SHELL = /bin/sh") (string-append "SHELL = "(which "sh"))))
 313              (substitute* "setup.ml"
 314                (("LDFLAGS=-fPIC")
 315                 (string-append "LDFLAGS=-fPIC\"; \"SHELL=" (which "sh"))))
 316              #t)))))
 317     (home-page "https://github.com/fhcrc/mcl")
 318     (synopsis "OCaml wrappers around MCL")
 319     (description
 320      "This package provides OCaml bindings for the MCL graph clustering
 321 algorithm.")
 322     (license license:gpl3)))
 323
 324 (define-public ocaml4.01-mcl
 325   (package-with-ocaml4.01 ocaml-mcl))
 326
 327 (define-public randomjungle
 328   (package
 329     (name "randomjungle")
 330     (version "2.1.0")
 331     (source
 332      (origin
 333        (method url-fetch)
 334        (uri (string-append
 335              "https://www.imbs.uni-luebeck.de/fileadmin/files/Software"
 336              "/randomjungle/randomjungle-" version ".tar_.gz"))
 337        (patches (search-patches "randomjungle-disable-static-build.patch"))
 338        (sha256
 339         (base32
 340          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 341     (build-system gnu-build-system)
 342     (arguments
 343      `(#:configure-flags
 344        (list "--disable-static"
 345              (string-append "--with-boost="
 346                             (assoc-ref %build-inputs "boost")))
 347        #:phases
 348        (modify-phases %standard-phases
 349          (add-before
 350           'configure 'set-CXXFLAGS
 351           (lambda _
 352             (setenv "CXXFLAGS" "-fpermissive ")
 353             #t)))))
 354     (inputs
 355      `(("boost" ,boost)
 356        ("gsl" ,gsl)
 357        ("libxml2" ,libxml2)
 358        ("zlib" ,zlib)))
 359     (native-inputs
 360      `(("gfortran" ,gfortran)
 361        ("gfortran:lib" ,gfortran "lib")))
 362     ;; Non-portable assembly instructions are used so building fails on
 363     ;; platforms other than x86_64 or i686.
 364     (supported-systems '("x86_64-linux" "i686-linux"))
 365     (home-page "https://www.imbs.uni-luebeck.de/forschung/software/details.html#c224")
 366     (synopsis "Implementation of the Random Forests machine learning method")
 367     (description
 368      "Random Jungle is an implementation of Random Forests.  It is supposed to
 369 analyse high dimensional data.  In genetics, it can be used for analysing big
 370 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 371 learning method.  Most interesting features are variable selection, missing
 372 value imputation, classifier creation, generalization error estimation and
 373 sample proximities between pairs of cases.")
 374     (license license:gpl3+)))
 375
 376 (define-public shogun
 377   (package
 378     (name "shogun")
 379     (version "6.1.3")
 380     (source
 381      (origin
 382        (method url-fetch)
 383        (uri (string-append
 384              "ftp://shogun-toolbox.org/shogun/releases/"
 385              (version-major+minor version)
 386              "/sources/shogun-" version ".tar.bz2"))
 387        (sha256
 388         (base32
 389          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 390        (modules '((guix build utils)
 391                   (ice-9 rdelim)))
 392        (snippet
 393         '(begin
 394            ;; Remove non-free sources and files referencing them
 395            (for-each delete-file
 396                      (find-files "src/shogun/classifier/svm/"
 397                                  "SVMLight\\.(cpp|h)"))
 398            (for-each delete-file
 399                      (find-files "examples/undocumented/libshogun/"
 400                                  (string-append
 401                                   "(classifier_.*svmlight.*|"
 402                                   "evaluation_cross_validation_locked_comparison).cpp")))
 403            ;; Remove non-free functions.
 404            (define (delete-ifdefs file)
 405              (with-atomic-file-replacement file
 406                (lambda (in out)
 407                  (let loop ((line (read-line in 'concat))
 408                             (skipping? #f))
 409                    (if (eof-object? line)
 410                        #t
 411                        (let ((skip-next?
 412                               (or (and skipping?
 413                                        (not (string-prefix?
 414                                              "#endif //USE_SVMLIGHT" line)))
 415                                   (string-prefix?
 416                                    "#ifdef USE_SVMLIGHT" line))))
 417                          (when (or (not skipping?)
 418                                    (and skipping? (not skip-next?)))
 419                            (display line out))
 420                          (loop (read-line in 'concat) skip-next?)))))))
 421            (for-each delete-ifdefs
 422                      (append
 423                       (find-files "src/shogun/classifier/mkl"
 424                                   "^MKLClassification\\.cpp")
 425                       (find-files "src/shogun/classifier/svm"
 426                                   "^SVMLightOneClass\\.(cpp|h)")
 427                       (find-files "src/shogun/multiclass"
 428                                   "^ScatterSVM\\.(cpp|h)")
 429                       (find-files "src/shogun/kernel/"
 430                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 431                       (find-files "src/shogun/regression/svr"
 432                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 433                       (find-files "src/shogun/transfer/domain_adaptation"
 434                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 435            #t))))
 436     (build-system cmake-build-system)
 437     (arguments
 438      '(#:tests? #f ;no check target
 439        #:phases
 440        (modify-phases %standard-phases
 441          (add-after 'unpack 'delete-broken-symlinks
 442            (lambda _
 443              (for-each delete-file '("applications/arts/data"
 444                                      "applications/asp/data"
 445                                      "applications/easysvm/data"
 446                                      "applications/msplicer/data"
 447                                      "applications/ocr/data"
 448                                      "examples/meta/data"
 449                                      "examples/undocumented/data"))
 450              #t))
 451          (add-after 'unpack 'change-R-target-path
 452            (lambda* (#:key outputs #:allow-other-keys)
 453              (substitute* '("src/interfaces/r/CMakeLists.txt"
 454                             "examples/meta/r/CMakeLists.txt")
 455                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 456                 (string-append (assoc-ref outputs "out")
 457                                "/lib/R/library/")))
 458              #t))
 459          (add-after 'unpack 'fix-octave-modules
 460            (lambda* (#:key outputs #:allow-other-keys)
 461              (substitute* "src/interfaces/octave/CMakeLists.txt"
 462                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 463                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 464                ;; change target directory
 465                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 466                 (string-append (assoc-ref outputs "out")
 467                                "/share/octave/packages")))
 468              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 469                             "src/interfaces/octave/sg_print_functions.cpp")
 470                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 471                (("octave/config\\.h") "octave/octave-config.h")
 472                (("octave/oct-obj.h") "octave/ovl.h"))
 473              #t))
 474          (add-after 'unpack 'move-rxcpp
 475            (lambda* (#:key inputs #:allow-other-keys)
 476              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 477                (mkdir-p rxcpp-dir)
 478                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 479                #t)))
 480          (add-before 'build 'set-HOME
 481            ;; $HOME needs to be set at some point during the build phase
 482            (lambda _ (setenv "HOME" "/tmp") #t)))
 483        #:configure-flags
 484        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 485              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 486              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 487              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 488              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 489              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 490              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 491              "-DINTERFACE_OCTAVE=ON"
 492              "-DINTERFACE_PYTHON=ON"
 493              "-DINTERFACE_R=ON")))
 494     (inputs
 495      `(("python" ,python)
 496        ("numpy" ,python-numpy)
 497        ("r-minimal" ,r-minimal)
 498        ("octave" ,octave-cli)
 499        ("swig" ,swig)
 500        ("eigen" ,eigen)
 501        ("hdf5" ,hdf5)
 502        ("atlas" ,atlas)
 503        ("arpack" ,arpack-ng)
 504        ("lapack" ,lapack)
 505        ("glpk" ,glpk)
 506        ("libxml2" ,libxml2)
 507        ("lzo" ,lzo)
 508        ("zlib" ,zlib)))
 509     (native-inputs
 510      `(("pkg-config" ,pkg-config)
 511        ("rxcpp" ,rxcpp)))
 512     ;; Non-portable SSE instructions are used so building fails on platforms
 513     ;; other than x86_64.
 514     (supported-systems '("x86_64-linux"))
 515     (home-page "http://shogun-toolbox.org/")
 516     (synopsis "Machine learning toolbox")
 517     (description
 518      "The Shogun Machine learning toolbox provides a wide range of unified and
 519 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 520 combine multiple data representations, algorithm classes, and general purpose
 521 tools.  This enables both rapid prototyping of data pipelines and extensibility
 522 in terms of new algorithms.")
 523     (license license:gpl3+)))
 524
 525 (define-public rxcpp
 526   (package
 527     (name "rxcpp")
 528     (version "4.0.0")
 529     (source
 530      (origin
 531        (method url-fetch)
 532        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 533                            version ".tar.gz"))
 534        (sha256
 535         (base32
 536          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 537        (file-name (string-append name "-" version ".tar.gz"))))
 538     (build-system cmake-build-system)
 539     (arguments
 540      `(#:phases
 541        (modify-phases %standard-phases
 542          (add-after 'unpack 'remove-werror
 543            (lambda _
 544              (substitute* (find-files ".")
 545                (("-Werror") ""))
 546              #t))
 547          (replace 'check
 548            (lambda _
 549              (invoke "ctest"))))))
 550     (native-inputs
 551      `(("catch" ,catch-framework)))
 552     (home-page "http://reactivex.io/")
 553     (synopsis "Reactive Extensions for C++")
 554     (description
 555      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 556 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 557 and event-based programs by using observable sequences.
 558
 559 It extends the observer pattern to support sequences of data and/or events and
 560 adds operators that allow you to compose sequences together declaratively while
 561 abstracting away concerns about things like low-level threading,
 562 synchronization, thread-safety, concurrent data structures, and non-blocking
 563 I/O.")
 564     (license license:asl2.0)))
 565
 566 (define-public r-adaptivesparsity
 567   (package
 568     (name "r-adaptivesparsity")
 569     (version "1.6")
 570     (source (origin
 571               (method url-fetch)
 572               (uri (cran-uri "AdaptiveSparsity" version))
 573               (sha256
 574                (base32
 575                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 576     (properties
 577      `((upstream-name . "AdaptiveSparsity")))
 578     (build-system r-build-system)
 579     (arguments
 580      `(#:phases
 581        (modify-phases %standard-phases
 582          (add-after 'unpack 'link-against-armadillo
 583            (lambda _
 584              (substitute* "src/Makevars"
 585                (("PKG_LIBS=" prefix)
 586                 (string-append prefix "-larmadillo"))))))))
 587     (propagated-inputs
 588      `(("r-mass" ,r-mass)
 589        ("r-matrix" ,r-matrix)
 590        ("r-rcpp" ,r-rcpp)
 591        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 592     (inputs
 593      `(("armadillo" ,armadillo)))
 594     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 595     (synopsis "Adaptive sparsity models")
 596     (description
 597      "This package implements the Figueiredo machine learning algorithm for
 598 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 599 geometric models.")
 600     (license license:lgpl3+)))
 601
 602 (define-public r-kernlab
 603   (package
 604     (name "r-kernlab")
 605     (version "0.9-27")
 606     (source
 607      (origin
 608        (method url-fetch)
 609        (uri (cran-uri "kernlab" version))
 610        (sha256
 611         (base32
 612          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 613     (build-system r-build-system)
 614     (home-page "https://cran.r-project.org/web/packages/kernlab")
 615     (synopsis "Kernel-based machine learning tools")
 616     (description
 617      "This package provides kernel-based machine learning methods for
 618 classification, regression, clustering, novelty detection, quantile regression
 619 and dimensionality reduction.  Among other methods @code{kernlab} includes
 620 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 621 and a QP solver.")
 622     (license license:gpl2)))
 623
 624 (define-public dlib
 625   (package
 626     (name "dlib")
 627     (version "19.7")
 628     (source (origin
 629               (method url-fetch)
 630               (uri (string-append
 631                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 632               (sha256
 633                (base32
 634                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 635               (modules '((guix build utils)))
 636               (snippet
 637                '(begin
 638                   ;; Delete ~13MB of bundled dependencies.
 639                   (delete-file-recursively "dlib/external")
 640                   (delete-file-recursively "docs/dlib/external")
 641                   #t))))
 642     (build-system cmake-build-system)
 643     (arguments
 644      `(#:phases
 645        (modify-phases %standard-phases
 646          (add-after 'unpack 'disable-asserts
 647            (lambda _
 648              ;; config.h recommends explicitly enabling or disabling asserts
 649              ;; when building as a shared library. By default neither is set.
 650              (substitute* "dlib/config.h"
 651                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 652              #t))
 653          (add-after 'disable-asserts 'disable-failing-tests
 654            (lambda _
 655              ;; One test times out on MIPS, so we need to disable it.
 656              ;; Others are flaky on some platforms.
 657              (let* ((system ,(or (%current-target-system)
 658                                  (%current-system)))
 659                     (disabled-tests (cond
 660                                      ((string-prefix? "mips64" system)
 661                                       '("object_detector" ; timeout
 662                                         "data_io"))
 663                                      ((string-prefix? "armhf" system)
 664                                       '("learning_to_track"))
 665                                      ((string-prefix? "i686" system)
 666                                       '("optimization"))
 667                                      (else '()))))
 668                (for-each
 669                 (lambda (test)
 670                   (substitute* "dlib/test/makefile"
 671                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 672                 disabled-tests)
 673                #t)))
 674          (replace 'check
 675            (lambda _
 676              ;; No test target, so we build and run the unit tests here.
 677              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 678                (with-directory-excursion test-dir
 679                  (invoke "make" "-j" (number->string (parallel-job-count)))
 680                  (invoke "./dtest" "--runall"))
 681                #t)))
 682          (add-after 'install 'delete-static-library
 683            (lambda* (#:key outputs #:allow-other-keys)
 684              (delete-file (string-append (assoc-ref outputs "out")
 685                                          "/lib/libdlib.a"))
 686              #t)))))
 687     (native-inputs
 688      `(("pkg-config" ,pkg-config)
 689        ;; For tests.
 690        ("libnsl" ,libnsl)))
 691     (inputs
 692      `(("giflib" ,giflib)
 693        ("lapack" ,lapack)
 694        ("libjpeg" ,libjpeg)
 695        ("libpng" ,libpng)
 696        ("libx11" ,libx11)
 697        ("openblas" ,openblas)
 698        ("zlib" ,zlib)))
 699     (synopsis
 700      "Toolkit for making machine learning and data analysis applications in C++")
 701     (description
 702      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 703 tools.  It is used in both industry and academia in a wide range of domains
 704 including robotics, embedded devices, mobile phones, and large high performance
 705 computing environments.")
 706     (home-page "http://dlib.net")
 707     (license license:boost1.0)))
 708
 709 (define-public python-scikit-learn
 710   (package
 711     (name "python-scikit-learn")
 712     (version "0.20.1")
 713     (source
 714      (origin
 715        (method git-fetch)
 716        (uri (git-reference
 717              (url "https://github.com/scikit-learn/scikit-learn.git")
 718              (commit version)))
 719        (file-name (git-file-name name version))
 720        (sha256
 721         (base32
 722          "0fkhwg3xn1s7ln9q1szq6kwc4jhwvjh8w4kmv9wcrqy7cq3lbv0d"))))
 723     (build-system python-build-system)
 724     (arguments
 725      `(#:phases
 726        (modify-phases %standard-phases
 727          (add-after 'build 'build-ext
 728            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 729          (replace 'check
 730            (lambda _
 731              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 732              (setenv "OPENBLAS_NUM_THREADS" "1")
 733
 734              ;; Some tests require write access to $HOME.
 735              (setenv "HOME" "/tmp")
 736
 737              (invoke "pytest" "sklearn" "-m" "not network")))
 738          ;; FIXME: This fails with permission denied
 739          (delete 'reset-gzip-timestamps))))
 740     (inputs
 741      `(("openblas" ,openblas)))
 742     (native-inputs
 743      `(("python-pytest" ,python-pytest)
 744        ("python-pandas" ,python-pandas) ;for tests
 745        ("python-cython" ,python-cython)))
 746     (propagated-inputs
 747      `(("python-numpy" ,python-numpy)
 748        ("python-scipy" ,python-scipy)))
 749     (home-page "http://scikit-learn.org/")
 750     (synopsis "Machine Learning in Python")
 751     (description
 752      "Scikit-learn provides simple and efficient tools for data mining and
 753 data analysis.")
 754     (license license:bsd-3)))
 755
 756 (define-public python2-scikit-learn
 757   (package-with-python2 python-scikit-learn))
 758
 759 (define-public python-autograd
 760   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 761          (revision "0")
 762          (version (git-version "0.0.0" revision commit)))
 763     (package
 764       (name "python-autograd")
 765       (home-page "https://github.com/HIPS/autograd")
 766       (source (origin
 767                 (method git-fetch)
 768                 (uri (git-reference
 769                       (url home-page)
 770                       (commit commit)))
 771                 (sha256
 772                  (base32
 773                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 774                 (file-name (git-file-name name version))))
 775       (version version)
 776       (build-system python-build-system)
 777       (native-inputs
 778        `(("python-nose" ,python-nose)
 779          ("python-pytest" ,python-pytest)))
 780       (propagated-inputs
 781        `(("python-future" ,python-future)
 782          ("python-numpy" ,python-numpy)))
 783       (arguments
 784        `(#:phases (modify-phases %standard-phases
 785                     (replace 'check
 786                       (lambda _
 787                         (invoke "py.test" "-v"))))))
 788       (synopsis "Efficiently computes derivatives of NumPy code")
 789       (description "Autograd can automatically differentiate native Python and
 790 NumPy code.  It can handle a large subset of Python's features, including loops,
 791 ifs, recursion and closures, and it can even take derivatives of derivatives
 792 of derivatives.  It supports reverse-mode differentiation
 793 (a.k.a. backpropagation), which means it can efficiently take gradients of
 794 scalar-valued functions with respect to array-valued arguments, as well as
 795 forward-mode differentiation, and the two can be composed arbitrarily.  The
 796 main intended application of Autograd is gradient-based optimization.")
 797       (license license:expat))))
 798
 799 (define-public python2-autograd
 800   (package-with-python2 python-autograd))
 801
 802 (define-public lightgbm
 803   (package
 804     (name "lightgbm")
 805     (version "2.0.12")
 806     (source (origin
 807               (method url-fetch)
 808               (uri (string-append
 809                     "https://github.com/Microsoft/LightGBM/archive/v"
 810                     version ".tar.gz"))
 811               (sha256
 812                (base32
 813                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 814               (file-name (string-append name "-" version ".tar.gz"))))
 815     (native-inputs
 816      `(("python-pytest" ,python-pytest)
 817        ("python-nose" ,python-nose)))
 818     (inputs
 819      `(("openmpi" ,openmpi)))
 820     (propagated-inputs
 821      `(("python-numpy" ,python-numpy)
 822        ("python-scipy" ,python-scipy)))
 823     (arguments
 824      `(#:configure-flags
 825        '("-DUSE_MPI=ON")
 826        #:phases
 827        (modify-phases %standard-phases
 828          (replace 'check
 829            (lambda* (#:key outputs #:allow-other-keys)
 830              (with-directory-excursion ,(string-append "../LightGBM-" version)
 831                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 832     (build-system cmake-build-system)
 833     (home-page "https://github.com/Microsoft/LightGBM")
 834     (synopsis "Gradient boosting framework based on decision tree algorithms")
 835     (description "LightGBM is a gradient boosting framework that uses tree
 836 based learning algorithms.  It is designed to be distributed and efficient with
 837 the following advantages:
 838
 839 @itemize
 840 @item Faster training speed and higher efficiency
 841 @item Lower memory usage
 842 @item Better accuracy
 843 @item Parallel and GPU learning supported (not enabled in this package)
 844 @item Capable of handling large-scale data
 845 @end itemize\n")
 846     (license license:expat)))
 847
 848 (define-public vowpal-wabbit
 849   ;; Language bindings not included.
 850   (package
 851     (name "vowpal-wabbit")
 852     (version "8.5.0")
 853     (source (origin
 854               (method url-fetch)
 855               (uri (string-append
 856                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 857                     version ".tar.gz"))
 858               (sha256
 859                (base32
 860                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 861               (file-name (string-append name "-" version ".tar.gz"))))
 862     (inputs
 863      `(("boost" ,boost)
 864        ("zlib" ,zlib)))
 865     (arguments
 866      `(#:configure-flags
 867        (list (string-append "--with-boost="
 868                             (assoc-ref %build-inputs "boost")))))
 869     (build-system gnu-build-system)
 870     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 871     (synopsis "Fast machine learning library for online learning")
 872     (description "Vowpal Wabbit is a machine learning system with techniques
 873 such as online, hashing, allreduce, reductions, learning2search, active, and
 874 interactive learning.")
 875     (license license:bsd-3)))