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 ;;;
  12 ;;; This file is part of GNU Guix.
  13 ;;;
  14 ;;; GNU Guix is free software; you can redistribute it and/or modify it
  15 ;;; under the terms of the GNU General Public License as published by
  16 ;;; the Free Software Foundation; either version 3 of the License, or (at
  17 ;;; your option) any later version.
  18 ;;;
  19 ;;; GNU Guix is distributed in the hope that it will be useful, but
  20 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  21 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22 ;;; GNU General Public License for more details.
  23 ;;;
  24 ;;; You should have received a copy of the GNU General Public License
  25 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  26
  27 (define-module (gnu packages machine-learning)
  28   #:use-module ((guix licenses) #:prefix license:)
  29   #:use-module (guix packages)
  30   #:use-module (guix utils)
  31   #:use-module (guix download)
  32   #:use-module (guix svn-download)
  33   #:use-module (guix build-system cmake)
  34   #:use-module (guix build-system gnu)
  35   #:use-module (guix build-system ocaml)
  36   #:use-module (guix build-system python)
  37   #:use-module (guix build-system r)
  38   #:use-module (guix git-download)
  39   #:use-module (gnu packages)
  40   #:use-module (gnu packages algebra)
  41   #:use-module (gnu packages autotools)
  42   #:use-module (gnu packages boost)
  43   #:use-module (gnu packages check)
  44   #:use-module (gnu packages compression)
  45   #:use-module (gnu packages cran)
  46   #:use-module (gnu packages dejagnu)
  47   #:use-module (gnu packages gcc)
  48   #:use-module (gnu packages image)
  49   #:use-module (gnu packages maths)
  50   #:use-module (gnu packages mpi)
  51   #:use-module (gnu packages ocaml)
  52   #:use-module (gnu packages onc-rpc)
  53   #:use-module (gnu packages perl)
  54   #:use-module (gnu packages pkg-config)
  55   #:use-module (gnu packages python)
  56   #:use-module (gnu packages statistics)
  57   #:use-module (gnu packages swig)
  58   #:use-module (gnu packages xml)
  59   #:use-module (gnu packages xorg))
  60
  61 (define-public fann
  62   ;; The last release is >100 commits behind, so we package from git.
  63   (let ((commit "d71d54788bee56ba4cf7522801270152da5209d7"))
  64     (package
  65       (name "fann")
  66       (version (string-append "2.2.0-1." (string-take commit 8)))
  67       (source (origin
  68                 (method git-fetch)
  69                 (uri (git-reference
  70                       (url "https://github.com/libfann/fann.git")
  71                       (commit commit)))
  72                 (file-name (string-append name "-" version "-checkout"))
  73                 (sha256
  74                  (base32
  75                   "0ibwpfrjs6q2lijs8slxjgzb2llcl6rk3v2ski4r6215g5jjhg3x"))))
  76       (build-system cmake-build-system)
  77       (arguments
  78        `(#:phases
  79          (modify-phases %standard-phases
  80            (replace 'check
  81              (lambda* (#:key outputs #:allow-other-keys)
  82                (let* ((out (assoc-ref outputs "out")))
  83                  (with-directory-excursion (string-append (getcwd) "/tests")
  84                    (invoke "./fann_tests"))))))))
  85       (home-page "http://leenissen.dk/fann/wp/")
  86       (synopsis "Fast Artificial Neural Network")
  87       (description
  88        "FANN is a free open source neural network library, which implements
  89 multilayer artificial neural networks in C with support for both fully
  90 connected and sparsely connected networks.")
  91       (license license:lgpl2.1))))
  92
  93 (define-public libsvm
  94   (package
  95     (name "libsvm")
  96     (version "3.22")
  97     (source
  98      (origin
  99        (method url-fetch)
 100        (uri (string-append "https://www.csie.ntu.edu.tw/~cjlin/libsvm/"
 101                            name "-" version ".tar.gz"))
 102        (sha256
 103         (base32
 104          "0zd7s19y5vb7agczl6456bn45cj1y64739sslaskw1qk7dywd0bd"))))
 105     (build-system gnu-build-system)
 106     (arguments
 107      `(#:tests? #f ;no "check" target
 108        #:phases (modify-phases %standard-phases
 109                   (delete 'configure)
 110                   (replace
 111                    'install             ; no ‘install’ target
 112                    (lambda* (#:key outputs #:allow-other-keys)
 113                      (let* ((out (assoc-ref outputs "out"))
 114                             (bin (string-append out "/bin/")))
 115                        (mkdir-p bin)
 116                        (for-each (lambda (file)
 117                                    (copy-file file (string-append bin file)))
 118                                  '("svm-train"
 119                                    "svm-predict"
 120                                    "svm-scale")))
 121                      #t)))))
 122     (home-page "http://www.csie.ntu.edu.tw/~cjlin/libsvm/")
 123     (synopsis "Library for Support Vector Machines")
 124     (description
 125      "LIBSVM is a machine learning library for support vector
 126 classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and
 127 distribution estimation (one-class SVM).  It supports multi-class
 128 classification.")
 129     (license license:bsd-3)))
 130
 131 (define-public python-libsvm
 132   (package (inherit libsvm)
 133     (name "python-libsvm")
 134     (build-system gnu-build-system)
 135     (arguments
 136      `(#:tests? #f ;no "check" target
 137        #:make-flags '("-C" "python")
 138        #:phases
 139        (modify-phases %standard-phases
 140          (delete 'configure)
 141          (replace
 142           'install                      ; no ‘install’ target
 143           (lambda* (#:key inputs outputs #:allow-other-keys)
 144             (let ((site (string-append (assoc-ref outputs "out")
 145                                        "/lib/python"
 146                                        (string-take
 147                                         (string-take-right
 148                                          (assoc-ref inputs "python") 5) 3)
 149                                        "/site-packages/")))
 150               (substitute* "python/svm.py"
 151                 (("../libsvm.so.2") "libsvm.so.2"))
 152               (mkdir-p site)
 153               (for-each (lambda (file)
 154                           (copy-file file (string-append site (basename file))))
 155                         (find-files "python" "\\.py"))
 156               (copy-file "libsvm.so.2"
 157                          (string-append site "libsvm.so.2")))
 158             #t)))))
 159     (inputs
 160      `(("python" ,python)))
 161     (synopsis "Python bindings of libSVM")))
 162
 163 (define-public ghmm
 164   ;; The latest release candidate is several years and a couple of fixes have
 165   ;; been published since.  This is why we download the sources from the SVN
 166   ;; repository.
 167   (let ((svn-revision 2341))
 168     (package
 169       (name "ghmm")
 170       (version (string-append "0.9-rc3-0." (number->string svn-revision)))
 171       (source (origin
 172                 (method svn-fetch)
 173                 (uri (svn-reference
 174                       (url "http://svn.code.sf.net/p/ghmm/code/trunk")
 175                       (revision svn-revision)))
 176                 (file-name (string-append name "-" version))
 177                 (sha256
 178                  (base32
 179                   "0qbq1rqp94l530f043qzp8aw5lj7dng9wq0miffd7spd1ff638wq"))))
 180       (build-system gnu-build-system)
 181       (arguments
 182        `(#:imported-modules (,@%gnu-build-system-modules
 183                              (guix build python-build-system))
 184          #:phases
 185          (modify-phases %standard-phases
 186            (add-after 'unpack 'enter-dir
 187              (lambda _ (chdir "ghmm") #t))
 188            (delete 'check)
 189            (add-after 'install 'check
 190              (assoc-ref %standard-phases 'check))
 191            (add-before 'check 'fix-PYTHONPATH
 192              (lambda* (#:key inputs outputs #:allow-other-keys)
 193                (let ((python-version ((@@ (guix build python-build-system)
 194                                            get-python-version)
 195                                       (assoc-ref inputs "python"))))
 196                  (setenv "PYTHONPATH"
 197                          (string-append (getenv "PYTHONPATH")
 198                                         ":" (assoc-ref outputs "out")
 199                                         "/lib/python" python-version
 200                                         "/site-packages")))
 201                #t))
 202            (add-after 'enter-dir 'fix-runpath
 203              (lambda* (#:key outputs #:allow-other-keys)
 204                (substitute* "ghmmwrapper/setup.py"
 205                  (("^(.*)extra_compile_args = \\[" line indent)
 206                   (string-append indent
 207                                  "extra_link_args = [\"-Wl,-rpath="
 208                                  (assoc-ref outputs "out") "/lib\"],\n"
 209                                  line
 210                                  "\"-Wl,-rpath="
 211                                  (assoc-ref outputs "out")
 212                                  "/lib\", ")))
 213                #t))
 214            (add-after 'enter-dir 'disable-broken-tests
 215              (lambda _
 216                (substitute* "tests/Makefile.am"
 217                  ;; GHMM_SILENT_TESTS is assumed to be a command.
 218                  (("TESTS_ENVIRONMENT.*") "")
 219                  ;; Do not build broken tests.
 220                  (("chmm .*") "")
 221                  (("read_fa .*") "")
 222                  (("mcmc .*") "")
 223                  (("label_higher_order_test.*$")
 224                   "label_higher_order_test\n"))
 225
 226                ;; These Python unittests are broken as there is no gato.
 227                ;; See https://sourceforge.net/p/ghmm/support-requests/3/
 228                (substitute* "ghmmwrapper/ghmmunittests.py"
 229                  (("^(.*)def (testNewXML|testMultipleTransitionClasses|testNewXML)"
 230                    line indent)
 231                   (string-append indent
 232                                  "@unittest.skip(\"Disabled by Guix\")\n"
 233                                  line)))
 234                #t))
 235            (add-after 'disable-broken-tests 'autogen
 236              (lambda _
 237                (invoke "bash" "autogen.sh"))))))
 238       (inputs
 239        `(("python" ,python-2) ; only Python 2 is supported
 240          ("libxml2" ,libxml2)))
 241       (native-inputs
 242        `(("pkg-config" ,pkg-config)
 243          ("dejagnu" ,dejagnu)
 244          ("swig" ,swig)
 245          ("autoconf" ,autoconf)
 246          ("automake" ,automake)
 247          ("libtool" ,libtool)))
 248       (home-page "http://ghmm.org")
 249       (synopsis "Hidden Markov Model library")
 250       (description
 251        "The General Hidden Markov Model library (GHMM) is a C library with
 252 additional Python bindings implementing a wide range of types of @dfn{Hidden
 253 Markov Models} (HMM) and algorithms: discrete, continuous emissions, basic
 254 training, HMM clustering, HMM mixtures.")
 255       (license license:lgpl2.0+))))
 256
 257 (define-public mcl
 258   (package
 259     (name "mcl")
 260     (version "14.137")
 261     (source (origin
 262               (method url-fetch)
 263               (uri (string-append
 264                     "http://micans.org/mcl/src/mcl-"
 265                     (string-replace-substring version "." "-")
 266                     ".tar.gz"))
 267               (sha256
 268                (base32
 269                 "15xlax3z31lsn62vlg94hkm75nm40q4679amnfg13jm8m2bnhy5m"))))
 270     (build-system gnu-build-system)
 271     (arguments
 272      `(#:configure-flags (list "--enable-blast")))
 273     (inputs
 274      `(("perl" ,perl)))
 275     (home-page "http://micans.org/mcl/")
 276     (synopsis "Clustering algorithm for graphs")
 277     (description
 278      "The MCL algorithm is short for the @dfn{Markov Cluster Algorithm}, a
 279 fast and scalable unsupervised cluster algorithm for graphs (also known as
 280 networks) based on simulation of (stochastic) flow in graphs.")
 281     ;; In the LICENCE file and web page it says "The software is licensed
 282     ;; under the GNU General Public License, version 3.", but in several of
 283     ;; the source code files it suggests GPL3 or later.
 284     ;; http://listserver.ebi.ac.uk/pipermail/mcl-users/2016/000376.html
 285     (license license:gpl3)))
 286
 287 (define-public ocaml-mcl
 288   (package
 289     (name "ocaml-mcl")
 290     (version "12-068oasis4")
 291     (source
 292      (origin
 293        (method url-fetch)
 294        (uri (string-append
 295              "https://github.com/fhcrc/mcl/archive/"
 296              version ".tar.gz"))
 297        (file-name (string-append name "-" version ".tar.gz"))
 298        (sha256
 299         (base32
 300          "1l5jbhwjpsj38x8b9698hfpkv75h8hn3kj0gihjhn8ym2cwwv110"))))
 301     (build-system ocaml-build-system)
 302     (arguments
 303      `(#:phases
 304        (modify-phases %standard-phases
 305          (add-before 'configure 'patch-paths
 306            (lambda _
 307              (substitute* "configure"
 308                (("SHELL = /bin/sh") (string-append "SHELL = "(which "sh"))))
 309              (substitute* "setup.ml"
 310                (("LDFLAGS=-fPIC")
 311                 (string-append "LDFLAGS=-fPIC\"; \"SHELL=" (which "sh"))))
 312              #t)))))
 313     (home-page "https://github.com/fhcrc/mcl")
 314     (synopsis "OCaml wrappers around MCL")
 315     (description
 316      "This package provides OCaml bindings for the MCL graph clustering
 317 algorithm.")
 318     (license license:gpl3)))
 319
 320 (define-public ocaml4.01-mcl
 321   (package-with-ocaml4.01 ocaml-mcl))
 322
 323 (define-public randomjungle
 324   (package
 325     (name "randomjungle")
 326     (version "2.1.0")
 327     (source
 328      (origin
 329        (method url-fetch)
 330        (uri (string-append
 331              "http://www.imbs-luebeck.de/imbs/sites/default/files/u59/"
 332              "randomjungle-" version ".tar_.gz"))
 333        (sha256
 334         (base32
 335          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 336     (build-system gnu-build-system)
 337     (arguments
 338      `(#:configure-flags
 339        (list (string-append "--with-boost="
 340                             (assoc-ref %build-inputs "boost")))
 341        #:phases
 342        (modify-phases %standard-phases
 343          (add-before
 344           'configure 'set-CXXFLAGS
 345           (lambda _
 346             (setenv "CXXFLAGS" "-fpermissive ")
 347             #t)))))
 348     (inputs
 349      `(("boost" ,boost)
 350        ("gsl" ,gsl)
 351        ("libxml2" ,libxml2)
 352        ("zlib" ,zlib)))
 353     (native-inputs
 354      `(("gfortran" ,gfortran)
 355        ("gfortran:lib" ,gfortran "lib")))
 356     ;; Non-portable assembly instructions are used so building fails on
 357     ;; platforms other than x86_64 or i686.
 358     (supported-systems '("x86_64-linux" "i686-linux"))
 359     (home-page "http://www.imbs-luebeck.de/imbs/de/node/227/")
 360     (synopsis "Implementation of the Random Forests machine learning method")
 361     (description
 362      "Random Jungle is an implementation of Random Forests.  It is supposed to
 363 analyse high dimensional data.  In genetics, it can be used for analysing big
 364 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 365 learning method.  Most interesting features are variable selection, missing
 366 value imputation, classifier creation, generalization error estimation and
 367 sample proximities between pairs of cases.")
 368     (license license:gpl3+)))
 369
 370 (define-public shogun
 371   (package
 372     (name "shogun")
 373     (version "6.1.3")
 374     (source
 375      (origin
 376        (method url-fetch)
 377        (uri (string-append
 378              "ftp://shogun-toolbox.org/shogun/releases/"
 379              (version-major+minor version)
 380              "/sources/shogun-" version ".tar.bz2"))
 381        (sha256
 382         (base32
 383          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 384        (modules '((guix build utils)
 385                   (ice-9 rdelim)))
 386        (snippet
 387         '(begin
 388            ;; Remove non-free sources and files referencing them
 389            (for-each delete-file
 390                      (find-files "src/shogun/classifier/svm/"
 391                                  "SVMLight\\.(cpp|h)"))
 392            (for-each delete-file
 393                      (find-files "examples/undocumented/libshogun/"
 394                                  (string-append
 395                                   "(classifier_.*svmlight.*|"
 396                                   "evaluation_cross_validation_locked_comparison).cpp")))
 397            ;; Remove non-free functions.
 398            (define (delete-ifdefs file)
 399              (with-atomic-file-replacement file
 400                (lambda (in out)
 401                  (let loop ((line (read-line in 'concat))
 402                             (skipping? #f))
 403                    (if (eof-object? line)
 404                        #t
 405                        (let ((skip-next?
 406                               (or (and skipping?
 407                                        (not (string-prefix?
 408                                              "#endif //USE_SVMLIGHT" line)))
 409                                   (string-prefix?
 410                                    "#ifdef USE_SVMLIGHT" line))))
 411                          (when (or (not skipping?)
 412                                    (and skipping? (not skip-next?)))
 413                            (display line out))
 414                          (loop (read-line in 'concat) skip-next?)))))))
 415            (for-each delete-ifdefs
 416                      (append
 417                       (find-files "src/shogun/classifier/mkl"
 418                                   "^MKLClassification\\.cpp")
 419                       (find-files "src/shogun/classifier/svm"
 420                                   "^SVMLightOneClass\\.(cpp|h)")
 421                       (find-files "src/shogun/multiclass"
 422                                   "^ScatterSVM\\.(cpp|h)")
 423                       (find-files "src/shogun/kernel/"
 424                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 425                       (find-files "src/shogun/regression/svr"
 426                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 427                       (find-files "src/shogun/transfer/domain_adaptation"
 428                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 429            #t))))
 430     (build-system cmake-build-system)
 431     (arguments
 432      '(#:tests? #f ;no check target
 433        #:phases
 434        (modify-phases %standard-phases
 435          (add-after 'unpack 'delete-broken-symlinks
 436            (lambda _
 437              (for-each delete-file '("applications/arts/data"
 438                                      "applications/asp/data"
 439                                      "applications/easysvm/data"
 440                                      "applications/msplicer/data"
 441                                      "applications/ocr/data"
 442                                      "examples/meta/data"
 443                                      "examples/undocumented/data"))
 444              #t))
 445          (add-after 'unpack 'change-R-target-path
 446            (lambda* (#:key outputs #:allow-other-keys)
 447              (substitute* '("src/interfaces/r/CMakeLists.txt"
 448                             "examples/meta/r/CMakeLists.txt")
 449                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 450                 (string-append (assoc-ref outputs "out")
 451                                "/lib/R/library/")))
 452              #t))
 453          (add-after 'unpack 'fix-octave-modules
 454            (lambda* (#:key outputs #:allow-other-keys)
 455              (substitute* "src/interfaces/octave/CMakeLists.txt"
 456                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 457                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 458                ;; change target directory
 459                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 460                 (string-append (assoc-ref outputs "out")
 461                                "/share/octave/packages")))
 462              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 463                             "src/interfaces/octave/sg_print_functions.cpp")
 464                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 465                (("octave/config\\.h") "octave/octave-config.h")
 466                (("octave/oct-obj.h") "octave/ovl.h"))
 467              #t))
 468          (add-after 'unpack 'move-rxcpp
 469            (lambda* (#:key inputs #:allow-other-keys)
 470              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 471                (mkdir-p rxcpp-dir)
 472                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 473                #t)))
 474          (add-before 'build 'set-HOME
 475            ;; $HOME needs to be set at some point during the build phase
 476            (lambda _ (setenv "HOME" "/tmp") #t)))
 477        #:configure-flags
 478        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 479              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 480              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 481              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 482              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 483              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 484              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 485              "-DINTERFACE_OCTAVE=ON"
 486              "-DINTERFACE_PYTHON=ON"
 487              "-DINTERFACE_R=ON")))
 488     (inputs
 489      `(("python" ,python)
 490        ("numpy" ,python-numpy)
 491        ("r-minimal" ,r-minimal)
 492        ("octave" ,octave)
 493        ("swig" ,swig)
 494        ("eigen" ,eigen)
 495        ("hdf5" ,hdf5)
 496        ("atlas" ,atlas)
 497        ("arpack" ,arpack-ng)
 498        ("lapack" ,lapack)
 499        ("glpk" ,glpk)
 500        ("libxml2" ,libxml2)
 501        ("lzo" ,lzo)
 502        ("zlib" ,zlib)))
 503     (native-inputs
 504      `(("pkg-config" ,pkg-config)
 505        ("rxcpp" ,rxcpp)))
 506     ;; Non-portable SSE instructions are used so building fails on platforms
 507     ;; other than x86_64.
 508     (supported-systems '("x86_64-linux"))
 509     (home-page "http://shogun-toolbox.org/")
 510     (synopsis "Machine learning toolbox")
 511     (description
 512      "The Shogun Machine learning toolbox provides a wide range of unified and
 513 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 514 combine multiple data representations, algorithm classes, and general purpose
 515 tools.  This enables both rapid prototyping of data pipelines and extensibility
 516 in terms of new algorithms.")
 517     (license license:gpl3+)))
 518
 519 (define-public rxcpp
 520   (package
 521     (name "rxcpp")
 522     (version "4.0.0")
 523     (source
 524      (origin
 525        (method url-fetch)
 526        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 527                            version ".tar.gz"))
 528        (sha256
 529         (base32
 530          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 531        (file-name (string-append name "-" version ".tar.gz"))))
 532     (build-system cmake-build-system)
 533     (arguments
 534      `(#:phases
 535        (modify-phases %standard-phases
 536          (add-after 'unpack 'remove-werror
 537            (lambda _
 538              (substitute* (find-files ".")
 539                (("-Werror") ""))
 540              #t))
 541          (replace 'check
 542            (lambda _
 543              (invoke "ctest"))))))
 544     (native-inputs
 545      `(("catch" ,catch-framework)))
 546     (home-page "http://reactivex.io/")
 547     (synopsis "Reactive Extensions for C++")
 548     (description
 549      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 550 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 551 and event-based programs by using observable sequences.
 552
 553 It extends the observer pattern to support sequences of data and/or events and
 554 adds operators that allow you to compose sequences together declaratively while
 555 abstracting away concerns about things like low-level threading,
 556 synchronization, thread-safety, concurrent data structures, and non-blocking
 557 I/O.")
 558     (license license:asl2.0)))
 559
 560 (define-public r-adaptivesparsity
 561   (package
 562     (name "r-adaptivesparsity")
 563     (version "1.6")
 564     (source (origin
 565               (method url-fetch)
 566               (uri (cran-uri "AdaptiveSparsity" version))
 567               (sha256
 568                (base32
 569                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 570     (properties
 571      `((upstream-name . "AdaptiveSparsity")))
 572     (build-system r-build-system)
 573     (arguments
 574      `(#:phases
 575        (modify-phases %standard-phases
 576          (add-after 'unpack 'link-against-armadillo
 577            (lambda _
 578              (substitute* "src/Makevars"
 579                (("PKG_LIBS=" prefix)
 580                 (string-append prefix "-larmadillo"))))))))
 581     (propagated-inputs
 582      `(("r-mass" ,r-mass)
 583        ("r-matrix" ,r-matrix)
 584        ("r-rcpp" ,r-rcpp)
 585        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 586     (inputs
 587      `(("armadillo" ,armadillo)))
 588     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 589     (synopsis "Adaptive sparsity models")
 590     (description
 591      "This package implements the Figueiredo machine learning algorithm for
 592 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 593 geometric models.")
 594     (license license:lgpl3+)))
 595
 596 (define-public r-kernlab
 597   (package
 598     (name "r-kernlab")
 599     (version "0.9-27")
 600     (source
 601      (origin
 602        (method url-fetch)
 603        (uri (cran-uri "kernlab" version))
 604        (sha256
 605         (base32
 606          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 607     (build-system r-build-system)
 608     (home-page "https://cran.r-project.org/web/packages/kernlab")
 609     (synopsis "Kernel-based machine learning tools")
 610     (description
 611      "This package provides kernel-based machine learning methods for
 612 classification, regression, clustering, novelty detection, quantile regression
 613 and dimensionality reduction.  Among other methods @code{kernlab} includes
 614 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 615 and a QP solver.")
 616     (license license:gpl2)))
 617
 618 (define-public dlib
 619   (package
 620     (name "dlib")
 621     (version "19.7")
 622     (source (origin
 623               (method url-fetch)
 624               (uri (string-append
 625                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 626               (sha256
 627                (base32
 628                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 629               (modules '((guix build utils)))
 630               (snippet
 631                '(begin
 632                   ;; Delete ~13MB of bundled dependencies.
 633                   (delete-file-recursively "dlib/external")
 634                   (delete-file-recursively "docs/dlib/external")
 635                   #t))))
 636     (build-system cmake-build-system)
 637     (arguments
 638      `(#:phases
 639        (modify-phases %standard-phases
 640          (add-after 'unpack 'disable-asserts
 641            (lambda _
 642              ;; config.h recommends explicitly enabling or disabling asserts
 643              ;; when building as a shared library. By default neither is set.
 644              (substitute* "dlib/config.h"
 645                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 646              #t))
 647          (add-after 'disable-asserts 'disable-failing-tests
 648            (lambda _
 649              ;; One test times out on MIPS, so we need to disable it.
 650              ;; Others are flaky on some platforms.
 651              (let* ((system ,(or (%current-target-system)
 652                                  (%current-system)))
 653                     (disabled-tests (cond
 654                                      ((string-prefix? "mips64" system)
 655                                       '("object_detector" ; timeout
 656                                         "data_io"))
 657                                      ((string-prefix? "armhf" system)
 658                                       '("learning_to_track"))
 659                                      ((string-prefix? "i686" system)
 660                                       '("optimization"))
 661                                      (else '()))))
 662                (for-each
 663                 (lambda (test)
 664                   (substitute* "dlib/test/makefile"
 665                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 666                 disabled-tests)
 667                #t)))
 668          (replace 'check
 669            (lambda _
 670              ;; No test target, so we build and run the unit tests here.
 671              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 672                (with-directory-excursion test-dir
 673                  (invoke "make" "-j" (number->string (parallel-job-count)))
 674                  (invoke "./dtest" "--runall"))
 675                #t)))
 676          (add-after 'install 'delete-static-library
 677            (lambda* (#:key outputs #:allow-other-keys)
 678              (delete-file (string-append (assoc-ref outputs "out")
 679                                          "/lib/libdlib.a"))
 680              #t)))))
 681     (native-inputs
 682      `(("pkg-config" ,pkg-config)
 683        ;; For tests.
 684        ("libnsl" ,libnsl)))
 685     (inputs
 686      `(("giflib" ,giflib)
 687        ("lapack" ,lapack)
 688        ("libjpeg" ,libjpeg)
 689        ("libpng" ,libpng)
 690        ("libx11" ,libx11)
 691        ("openblas" ,openblas)
 692        ("zlib" ,zlib)))
 693     (synopsis
 694      "Toolkit for making machine learning and data analysis applications in C++")
 695     (description
 696      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 697 tools.  It is used in both industry and academia in a wide range of domains
 698 including robotics, embedded devices, mobile phones, and large high performance
 699 computing environments.")
 700     (home-page "http://dlib.net")
 701     (license license:boost1.0)))
 702
 703 (define-public python-scikit-learn
 704   (package
 705     (name "python-scikit-learn")
 706     (version "0.19.2")
 707     (source
 708      (origin
 709        (method git-fetch)
 710        (uri (git-reference
 711              (url "https://github.com/scikit-learn/scikit-learn.git")
 712              (commit version)))
 713        (file-name (git-file-name name version))
 714        (sha256
 715         (base32
 716          "1dk9hdj01c0bny4ps78b7869fjw9gr6qklxf6wyql8h6nh4k19xm"))
 717        (patches (search-patches
 718                  "python-scikit-learn-fix-test-non-determinism.patch"))))
 719     (build-system python-build-system)
 720     (arguments
 721      `(#:phases
 722        (modify-phases %standard-phases
 723          (add-after 'build 'build-ext
 724            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 725          (replace 'check
 726            (lambda _
 727              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 728              (setenv "OPENBLAS_NUM_THREADS" "1")
 729              ;; Disable tests that require network access
 730              (delete-file "sklearn/datasets/tests/test_mldata.py")
 731              (delete-file "sklearn/datasets/tests/test_rcv1.py")
 732              (invoke "pytest" "sklearn")
 733              #t))
 734          ;; FIXME: This fails with permission denied
 735          (delete 'reset-gzip-timestamps))))
 736     (inputs
 737      `(("openblas" ,openblas)))
 738     (native-inputs
 739      `(("python-pytest" ,python-pytest)
 740        ("python-pandas" ,python-pandas) ;for tests
 741        ("python-cython" ,python-cython)))
 742     (propagated-inputs
 743      `(("python-numpy" ,python-numpy)
 744        ("python-scipy" ,python-scipy)))
 745     (home-page "http://scikit-learn.org/")
 746     (synopsis "Machine Learning in Python")
 747     (description
 748      "Scikit-learn provides simple and efficient tools for data mining and
 749 data analysis.")
 750     (license license:bsd-3)))
 751
 752 (define-public python2-scikit-learn
 753   (let ((parent (package-with-python2 python-scikit-learn)))
 754     (package (inherit parent)
 755       (arguments
 756        (substitute-keyword-arguments (package-arguments parent)
 757          ((#:phases phases)
 758           `(modify-phases ,phases
 759              (replace 'check
 760                (lambda _
 761                  ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 762                  (setenv "OPENBLAS_NUM_THREADS" "1")
 763                  ;; Some tests expect to be able to write to HOME.
 764                  (setenv "HOME" "/tmp")
 765                  ;; Disable tests that require network access
 766                  (delete-file "sklearn/datasets/tests/test_kddcup99.py")
 767                  (delete-file "sklearn/datasets/tests/test_mldata.py")
 768                  (delete-file "sklearn/datasets/tests/test_rcv1.py")
 769                  (invoke "pytest" "sklearn")
 770                  #t)))))))))
 771
 772 (define-public python-autograd
 773   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 774          (revision "0")
 775          (version (git-version "0.0.0" revision commit)))
 776     (package
 777       (name "python-autograd")
 778       (home-page "https://github.com/HIPS/autograd")
 779       (source (origin
 780                 (method git-fetch)
 781                 (uri (git-reference
 782                       (url home-page)
 783                       (commit commit)))
 784                 (sha256
 785                  (base32
 786                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 787                 (file-name (git-file-name name version))))
 788       (version version)
 789       (build-system python-build-system)
 790       (native-inputs
 791        `(("python-nose" ,python-nose)
 792          ("python-pytest" ,python-pytest)))
 793       (propagated-inputs
 794        `(("python-future" ,python-future)
 795          ("python-numpy" ,python-numpy)))
 796       (arguments
 797        `(#:phases (modify-phases %standard-phases
 798                     (replace 'check
 799                       (lambda _
 800                         (invoke "py.test" "-v"))))))
 801       (synopsis "Efficiently computes derivatives of NumPy code")
 802       (description "Autograd can automatically differentiate native Python and
 803 NumPy code.  It can handle a large subset of Python's features, including loops,
 804 ifs, recursion and closures, and it can even take derivatives of derivatives
 805 of derivatives.  It supports reverse-mode differentiation
 806 (a.k.a. backpropagation), which means it can efficiently take gradients of
 807 scalar-valued functions with respect to array-valued arguments, as well as
 808 forward-mode differentiation, and the two can be composed arbitrarily.  The
 809 main intended application of Autograd is gradient-based optimization.")
 810       (license license:expat))))
 811
 812 (define-public python2-autograd
 813   (package-with-python2 python-autograd))
 814
 815 (define-public lightgbm
 816   (package
 817     (name "lightgbm")
 818     (version "2.0.12")
 819     (source (origin
 820               (method url-fetch)
 821               (uri (string-append
 822                     "https://github.com/Microsoft/LightGBM/archive/v"
 823                     version ".tar.gz"))
 824               (sha256
 825                (base32
 826                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 827               (file-name (string-append name "-" version ".tar.gz"))))
 828     (native-inputs
 829      `(("python-pytest" ,python-pytest)
 830        ("python-nose" ,python-nose)))
 831     (inputs
 832      `(("openmpi" ,openmpi)))
 833     (propagated-inputs
 834      `(("python-numpy" ,python-numpy)
 835        ("python-scipy" ,python-scipy)))
 836     (arguments
 837      `(#:configure-flags
 838        '("-DUSE_MPI=ON")
 839        #:phases
 840        (modify-phases %standard-phases
 841          (replace 'check
 842            (lambda* (#:key outputs #:allow-other-keys)
 843              (with-directory-excursion ,(string-append "../LightGBM-" version)
 844                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 845     (build-system cmake-build-system)
 846     (home-page "https://github.com/Microsoft/LightGBM")
 847     (synopsis "Gradient boosting framework based on decision tree algorithms")
 848     (description "LightGBM is a gradient boosting framework that uses tree
 849 based learning algorithms.  It is designed to be distributed and efficient with
 850 the following advantages:
 851
 852 @itemize
 853 @item Faster training speed and higher efficiency
 854 @item Lower memory usage
 855 @item Better accuracy
 856 @item Parallel and GPU learning supported (not enabled in this package)
 857 @item Capable of handling large-scale data
 858 @end itemize\n")
 859     (license license:expat)))
 860
 861 (define-public vowpal-wabbit
 862   ;; Language bindings not included.
 863   (package
 864     (name "vowpal-wabbit")
 865     (version "8.5.0")
 866     (source (origin
 867               (method url-fetch)
 868               (uri (string-append
 869                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 870                     version ".tar.gz"))
 871               (sha256
 872                (base32
 873                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 874               (file-name (string-append name "-" version ".tar.gz"))))
 875     (inputs
 876      `(("boost" ,boost)
 877        ("zlib" ,zlib)))
 878     (arguments
 879      `(#:configure-flags
 880        (list (string-append "--with-boost="
 881                             (assoc-ref %build-inputs "boost")))))
 882     (build-system gnu-build-system)
 883     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 884     (synopsis "Fast machine learning library for online learning")
 885     (description "Vowpal Wabbit is a machine learning system with techniques
 886 such as online, hashing, allreduce, reductions, learning2search, active, and
 887 interactive learning.")
 888     (license license:bsd-3)))