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 python-xyz)
  59   #:use-module (gnu packages statistics)
  60   #:use-module (gnu packages swig)
  61   #:use-module (gnu packages xml)
  62   #:use-module (gnu packages xorg))
  63
  64 (define-public fann
  65   ;; The last release is >100 commits behind, so we package from git.
  66   (let ((commit "d71d54788bee56ba4cf7522801270152da5209d7"))
  67     (package
  68       (name "fann")
  69       (version (string-append "2.2.0-1." (string-take commit 8)))
  70       (source (origin
  71                 (method git-fetch)
  72                 (uri (git-reference
  73                       (url "https://github.com/libfann/fann.git")
  74                       (commit commit)))
  75                 (file-name (string-append name "-" version "-checkout"))
  76                 (sha256
  77                  (base32
  78                   "0ibwpfrjs6q2lijs8slxjgzb2llcl6rk3v2ski4r6215g5jjhg3x"))))
  79       (build-system cmake-build-system)
  80       (arguments
  81        `(#:phases
  82          (modify-phases %standard-phases
  83            (replace 'check
  84              (lambda* (#:key outputs #:allow-other-keys)
  85                (let* ((out (assoc-ref outputs "out")))
  86                  (with-directory-excursion (string-append (getcwd) "/tests")
  87                    (invoke "./fann_tests"))))))))
  88       (home-page "http://leenissen.dk/fann/wp/")
  89       (synopsis "Fast Artificial Neural Network")
  90       (description
  91        "FANN is a free open source neural network library, which implements
  92 multilayer artificial neural networks in C with support for both fully
  93 connected and sparsely connected networks.")
  94       (license license:lgpl2.1))))
  95
  96 (define-public libsvm
  97   (package
  98     (name "libsvm")
  99     (version "3.22")
 100     (source
 101      (origin
 102        (method url-fetch)
 103        (uri (string-append "https://www.csie.ntu.edu.tw/~cjlin/libsvm/"
 104                            name "-" version ".tar.gz"))
 105        (sha256
 106         (base32
 107          "0zd7s19y5vb7agczl6456bn45cj1y64739sslaskw1qk7dywd0bd"))))
 108     (build-system gnu-build-system)
 109     (arguments
 110      `(#:tests? #f ;no "check" target
 111        #:phases (modify-phases %standard-phases
 112                   (delete 'configure)
 113                   (replace
 114                    'install             ; no ‘install’ target
 115                    (lambda* (#:key outputs #:allow-other-keys)
 116                      (let* ((out (assoc-ref outputs "out"))
 117                             (bin (string-append out "/bin/")))
 118                        (mkdir-p bin)
 119                        (for-each (lambda (file)
 120                                    (copy-file file (string-append bin file)))
 121                                  '("svm-train"
 122                                    "svm-predict"
 123                                    "svm-scale")))
 124                      #t)))))
 125     (home-page "http://www.csie.ntu.edu.tw/~cjlin/libsvm/")
 126     (synopsis "Library for Support Vector Machines")
 127     (description
 128      "LIBSVM is a machine learning library for support vector
 129 classification, (C-SVC, nu-SVC), regression (epsilon-SVR, nu-SVR) and
 130 distribution estimation (one-class SVM).  It supports multi-class
 131 classification.")
 132     (license license:bsd-3)))
 133
 134 (define-public python-libsvm
 135   (package (inherit libsvm)
 136     (name "python-libsvm")
 137     (build-system gnu-build-system)
 138     (arguments
 139      `(#:tests? #f ;no "check" target
 140        #:make-flags '("-C" "python")
 141        #:phases
 142        (modify-phases %standard-phases
 143          (delete 'configure)
 144          (replace
 145           'install                      ; no ‘install’ target
 146           (lambda* (#:key inputs outputs #:allow-other-keys)
 147             (let ((site (string-append (assoc-ref outputs "out")
 148                                        "/lib/python"
 149                                        (string-take
 150                                         (string-take-right
 151                                          (assoc-ref inputs "python") 5) 3)
 152                                        "/site-packages/")))
 153               (substitute* "python/svm.py"
 154                 (("../libsvm.so.2") "libsvm.so.2"))
 155               (mkdir-p site)
 156               (for-each (lambda (file)
 157                           (copy-file file (string-append site (basename file))))
 158                         (find-files "python" "\\.py"))
 159               (copy-file "libsvm.so.2"
 160                          (string-append site "libsvm.so.2")))
 161             #t)))))
 162     (inputs
 163      `(("python" ,python)))
 164     (synopsis "Python bindings of libSVM")))
 165
 166 (define-public ghmm
 167   ;; The latest release candidate is several years and a couple of fixes have
 168   ;; been published since.  This is why we download the sources from the SVN
 169   ;; repository.
 170   (let ((svn-revision 2341))
 171     (package
 172       (name "ghmm")
 173       (version (string-append "0.9-rc3-0." (number->string svn-revision)))
 174       (source (origin
 175                 (method svn-fetch)
 176                 (uri (svn-reference
 177                       (url "http://svn.code.sf.net/p/ghmm/code/trunk")
 178                       (revision svn-revision)))
 179                 (file-name (string-append name "-" version))
 180                 (sha256
 181                  (base32
 182                   "0qbq1rqp94l530f043qzp8aw5lj7dng9wq0miffd7spd1ff638wq"))))
 183       (build-system gnu-build-system)
 184       (arguments
 185        `(#:imported-modules (,@%gnu-build-system-modules
 186                              (guix build python-build-system))
 187          #:phases
 188          (modify-phases %standard-phases
 189            (add-after 'unpack 'enter-dir
 190              (lambda _ (chdir "ghmm") #t))
 191            (delete 'check)
 192            (add-after 'install 'check
 193              (assoc-ref %standard-phases 'check))
 194            (add-before 'check 'fix-PYTHONPATH
 195              (lambda* (#:key inputs outputs #:allow-other-keys)
 196                (let ((python-version ((@@ (guix build python-build-system)
 197                                            get-python-version)
 198                                       (assoc-ref inputs "python"))))
 199                  (setenv "PYTHONPATH"
 200                          (string-append (getenv "PYTHONPATH")
 201                                         ":" (assoc-ref outputs "out")
 202                                         "/lib/python" python-version
 203                                         "/site-packages")))
 204                #t))
 205            (add-after 'enter-dir 'fix-runpath
 206              (lambda* (#:key outputs #:allow-other-keys)
 207                (substitute* "ghmmwrapper/setup.py"
 208                  (("^(.*)extra_compile_args = \\[" line indent)
 209                   (string-append indent
 210                                  "extra_link_args = [\"-Wl,-rpath="
 211                                  (assoc-ref outputs "out") "/lib\"],\n"
 212                                  line
 213                                  "\"-Wl,-rpath="
 214                                  (assoc-ref outputs "out")
 215                                  "/lib\", ")))
 216                #t))
 217            (add-after 'enter-dir 'disable-broken-tests
 218              (lambda _
 219                (substitute* "tests/Makefile.am"
 220                  ;; GHMM_SILENT_TESTS is assumed to be a command.
 221                  (("TESTS_ENVIRONMENT.*") "")
 222                  ;; Do not build broken tests.
 223                  (("chmm .*") "")
 224                  (("read_fa .*") "")
 225                  (("mcmc .*") "")
 226                  (("label_higher_order_test.*$")
 227                   "label_higher_order_test\n"))
 228
 229                ;; These Python unittests are broken as there is no gato.
 230                ;; See https://sourceforge.net/p/ghmm/support-requests/3/
 231                (substitute* "ghmmwrapper/ghmmunittests.py"
 232                  (("^(.*)def (testNewXML|testMultipleTransitionClasses|testNewXML)"
 233                    line indent)
 234                   (string-append indent
 235                                  "@unittest.skip(\"Disabled by Guix\")\n"
 236                                  line)))
 237                #t))
 238            (add-after 'disable-broken-tests 'autogen
 239              (lambda _
 240                (invoke "bash" "autogen.sh"))))))
 241       (inputs
 242        `(("python" ,python-2) ; only Python 2 is supported
 243          ("libxml2" ,libxml2)))
 244       (native-inputs
 245        `(("pkg-config" ,pkg-config)
 246          ("dejagnu" ,dejagnu)
 247          ("swig" ,swig)
 248          ("autoconf" ,autoconf)
 249          ("automake" ,automake)
 250          ("libtool" ,libtool)))
 251       (home-page "http://ghmm.org")
 252       (synopsis "Hidden Markov Model library")
 253       (description
 254        "The General Hidden Markov Model library (GHMM) is a C library with
 255 additional Python bindings implementing a wide range of types of @dfn{Hidden
 256 Markov Models} (HMM) and algorithms: discrete, continuous emissions, basic
 257 training, HMM clustering, HMM mixtures.")
 258       (license license:lgpl2.0+))))
 259
 260 (define-public mcl
 261   (package
 262     (name "mcl")
 263     (version "14.137")
 264     (source (origin
 265               (method url-fetch)
 266               (uri (string-append
 267                     "http://micans.org/mcl/src/mcl-"
 268                     (string-replace-substring version "." "-")
 269                     ".tar.gz"))
 270               (sha256
 271                (base32
 272                 "15xlax3z31lsn62vlg94hkm75nm40q4679amnfg13jm8m2bnhy5m"))))
 273     (build-system gnu-build-system)
 274     (arguments
 275      `(#:configure-flags (list "--enable-blast")))
 276     (inputs
 277      `(("perl" ,perl)))
 278     (home-page "http://micans.org/mcl/")
 279     (synopsis "Clustering algorithm for graphs")
 280     (description
 281      "The MCL algorithm is short for the @dfn{Markov Cluster Algorithm}, a
 282 fast and scalable unsupervised cluster algorithm for graphs (also known as
 283 networks) based on simulation of (stochastic) flow in graphs.")
 284     ;; In the LICENCE file and web page it says "The software is licensed
 285     ;; under the GNU General Public License, version 3.", but in several of
 286     ;; the source code files it suggests GPL3 or later.
 287     ;; http://listserver.ebi.ac.uk/pipermail/mcl-users/2016/000376.html
 288     (license license:gpl3)))
 289
 290 (define-public ocaml-mcl
 291   (package
 292     (name "ocaml-mcl")
 293     (version "12-068oasis4")
 294     (source
 295      (origin
 296        (method url-fetch)
 297        (uri (string-append
 298              "https://github.com/fhcrc/mcl/archive/"
 299              version ".tar.gz"))
 300        (file-name (string-append name "-" version ".tar.gz"))
 301        (sha256
 302         (base32
 303          "1l5jbhwjpsj38x8b9698hfpkv75h8hn3kj0gihjhn8ym2cwwv110"))))
 304     (build-system ocaml-build-system)
 305     (arguments
 306      `(#:ocaml ,ocaml-4.02
 307        #:findlib ,ocaml4.02-findlib
 308        #:phases
 309        (modify-phases %standard-phases
 310          (add-before 'configure 'patch-paths
 311            (lambda _
 312              (substitute* "configure"
 313                (("SHELL = /bin/sh") (string-append "SHELL = "(which "sh"))))
 314              (substitute* "setup.ml"
 315                (("LDFLAGS=-fPIC")
 316                 (string-append "LDFLAGS=-fPIC\"; \"SHELL=" (which "sh"))))
 317              #t)))))
 318     (home-page "https://github.com/fhcrc/mcl")
 319     (synopsis "OCaml wrappers around MCL")
 320     (description
 321      "This package provides OCaml bindings for the MCL graph clustering
 322 algorithm.")
 323     (license license:gpl3)))
 324
 325 (define-public ocaml4.01-mcl
 326   (package-with-ocaml4.01 ocaml-mcl))
 327
 328 (define-public randomjungle
 329   (package
 330     (name "randomjungle")
 331     (version "2.1.0")
 332     (source
 333      (origin
 334        (method url-fetch)
 335        (uri (string-append
 336              "https://www.imbs.uni-luebeck.de/fileadmin/files/Software"
 337              "/randomjungle/randomjungle-" version ".tar_.gz"))
 338        (patches (search-patches "randomjungle-disable-static-build.patch"))
 339        (sha256
 340         (base32
 341          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 342     (build-system gnu-build-system)
 343     (arguments
 344      `(#:configure-flags
 345        (list "--disable-static"
 346              (string-append "--with-boost="
 347                             (assoc-ref %build-inputs "boost")))
 348        #:phases
 349        (modify-phases %standard-phases
 350          (add-before
 351           'configure 'set-CXXFLAGS
 352           (lambda _
 353             (setenv "CXXFLAGS" "-fpermissive ")
 354             #t)))))
 355     (inputs
 356      `(("boost" ,boost)
 357        ("gsl" ,gsl)
 358        ("libxml2" ,libxml2)
 359        ("zlib" ,zlib)))
 360     (native-inputs
 361      `(("gfortran" ,gfortran)
 362        ("gfortran:lib" ,gfortran "lib")))
 363     ;; Non-portable assembly instructions are used so building fails on
 364     ;; platforms other than x86_64 or i686.
 365     (supported-systems '("x86_64-linux" "i686-linux"))
 366     (home-page "https://www.imbs.uni-luebeck.de/forschung/software/details.html#c224")
 367     (synopsis "Implementation of the Random Forests machine learning method")
 368     (description
 369      "Random Jungle is an implementation of Random Forests.  It is supposed to
 370 analyse high dimensional data.  In genetics, it can be used for analysing big
 371 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 372 learning method.  Most interesting features are variable selection, missing
 373 value imputation, classifier creation, generalization error estimation and
 374 sample proximities between pairs of cases.")
 375     (license license:gpl3+)))
 376
 377 (define-public shogun
 378   (package
 379     (name "shogun")
 380     (version "6.1.3")
 381     (source
 382      (origin
 383        (method url-fetch)
 384        (uri (string-append
 385              "ftp://shogun-toolbox.org/shogun/releases/"
 386              (version-major+minor version)
 387              "/sources/shogun-" version ".tar.bz2"))
 388        (sha256
 389         (base32
 390          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 391        (modules '((guix build utils)
 392                   (ice-9 rdelim)))
 393        (snippet
 394         '(begin
 395            ;; Remove non-free sources and files referencing them
 396            (for-each delete-file
 397                      (find-files "src/shogun/classifier/svm/"
 398                                  "SVMLight\\.(cpp|h)"))
 399            (for-each delete-file
 400                      (find-files "examples/undocumented/libshogun/"
 401                                  (string-append
 402                                   "(classifier_.*svmlight.*|"
 403                                   "evaluation_cross_validation_locked_comparison).cpp")))
 404            ;; Remove non-free functions.
 405            (define (delete-ifdefs file)
 406              (with-atomic-file-replacement file
 407                (lambda (in out)
 408                  (let loop ((line (read-line in 'concat))
 409                             (skipping? #f))
 410                    (if (eof-object? line)
 411                        #t
 412                        (let ((skip-next?
 413                               (or (and skipping?
 414                                        (not (string-prefix?
 415                                              "#endif //USE_SVMLIGHT" line)))
 416                                   (string-prefix?
 417                                    "#ifdef USE_SVMLIGHT" line))))
 418                          (when (or (not skipping?)
 419                                    (and skipping? (not skip-next?)))
 420                            (display line out))
 421                          (loop (read-line in 'concat) skip-next?)))))))
 422            (for-each delete-ifdefs
 423                      (append
 424                       (find-files "src/shogun/classifier/mkl"
 425                                   "^MKLClassification\\.cpp")
 426                       (find-files "src/shogun/classifier/svm"
 427                                   "^SVMLightOneClass\\.(cpp|h)")
 428                       (find-files "src/shogun/multiclass"
 429                                   "^ScatterSVM\\.(cpp|h)")
 430                       (find-files "src/shogun/kernel/"
 431                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 432                       (find-files "src/shogun/regression/svr"
 433                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 434                       (find-files "src/shogun/transfer/domain_adaptation"
 435                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 436            #t))))
 437     (build-system cmake-build-system)
 438     (arguments
 439      '(#:tests? #f ;no check target
 440        #:phases
 441        (modify-phases %standard-phases
 442          (add-after 'unpack 'delete-broken-symlinks
 443            (lambda _
 444              (for-each delete-file '("applications/arts/data"
 445                                      "applications/asp/data"
 446                                      "applications/easysvm/data"
 447                                      "applications/msplicer/data"
 448                                      "applications/ocr/data"
 449                                      "examples/meta/data"
 450                                      "examples/undocumented/data"))
 451              #t))
 452          (add-after 'unpack 'change-R-target-path
 453            (lambda* (#:key outputs #:allow-other-keys)
 454              (substitute* '("src/interfaces/r/CMakeLists.txt"
 455                             "examples/meta/r/CMakeLists.txt")
 456                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 457                 (string-append (assoc-ref outputs "out")
 458                                "/lib/R/library/")))
 459              #t))
 460          (add-after 'unpack 'fix-octave-modules
 461            (lambda* (#:key outputs #:allow-other-keys)
 462              (substitute* "src/interfaces/octave/CMakeLists.txt"
 463                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 464                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 465                ;; change target directory
 466                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 467                 (string-append (assoc-ref outputs "out")
 468                                "/share/octave/packages")))
 469              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 470                             "src/interfaces/octave/sg_print_functions.cpp")
 471                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 472                (("octave/config\\.h") "octave/octave-config.h")
 473                (("octave/oct-obj.h") "octave/ovl.h"))
 474              #t))
 475          (add-after 'unpack 'move-rxcpp
 476            (lambda* (#:key inputs #:allow-other-keys)
 477              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 478                (mkdir-p rxcpp-dir)
 479                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 480                #t)))
 481          (add-before 'build 'set-HOME
 482            ;; $HOME needs to be set at some point during the build phase
 483            (lambda _ (setenv "HOME" "/tmp") #t)))
 484        #:configure-flags
 485        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 486              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 487              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 488              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 489              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 490              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 491              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 492              "-DINTERFACE_OCTAVE=ON"
 493              "-DINTERFACE_PYTHON=ON"
 494              "-DINTERFACE_R=ON")))
 495     (inputs
 496      `(("python" ,python)
 497        ("numpy" ,python-numpy)
 498        ("r-minimal" ,r-minimal)
 499        ("octave" ,octave-cli)
 500        ("swig" ,swig)
 501        ("eigen" ,eigen)
 502        ("hdf5" ,hdf5)
 503        ("atlas" ,atlas)
 504        ("arpack" ,arpack-ng)
 505        ("lapack" ,lapack)
 506        ("glpk" ,glpk)
 507        ("libxml2" ,libxml2)
 508        ("lzo" ,lzo)
 509        ("zlib" ,zlib)))
 510     (native-inputs
 511      `(("pkg-config" ,pkg-config)
 512        ("rxcpp" ,rxcpp)))
 513     ;; Non-portable SSE instructions are used so building fails on platforms
 514     ;; other than x86_64.
 515     (supported-systems '("x86_64-linux"))
 516     (home-page "http://shogun-toolbox.org/")
 517     (synopsis "Machine learning toolbox")
 518     (description
 519      "The Shogun Machine learning toolbox provides a wide range of unified and
 520 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 521 combine multiple data representations, algorithm classes, and general purpose
 522 tools.  This enables both rapid prototyping of data pipelines and extensibility
 523 in terms of new algorithms.")
 524     (license license:gpl3+)))
 525
 526 (define-public rxcpp
 527   (package
 528     (name "rxcpp")
 529     (version "4.0.0")
 530     (source
 531      (origin
 532        (method url-fetch)
 533        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 534                            version ".tar.gz"))
 535        (sha256
 536         (base32
 537          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 538        (file-name (string-append name "-" version ".tar.gz"))))
 539     (build-system cmake-build-system)
 540     (arguments
 541      `(#:phases
 542        (modify-phases %standard-phases
 543          (add-after 'unpack 'remove-werror
 544            (lambda _
 545              (substitute* (find-files ".")
 546                (("-Werror") ""))
 547              #t))
 548          (replace 'check
 549            (lambda _
 550              (invoke "ctest"))))))
 551     (native-inputs
 552      `(("catch" ,catch-framework)))
 553     (home-page "http://reactivex.io/")
 554     (synopsis "Reactive Extensions for C++")
 555     (description
 556      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 557 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 558 and event-based programs by using observable sequences.
 559
 560 It extends the observer pattern to support sequences of data and/or events and
 561 adds operators that allow you to compose sequences together declaratively while
 562 abstracting away concerns about things like low-level threading,
 563 synchronization, thread-safety, concurrent data structures, and non-blocking
 564 I/O.")
 565     (license license:asl2.0)))
 566
 567 (define-public r-adaptivesparsity
 568   (package
 569     (name "r-adaptivesparsity")
 570     (version "1.6")
 571     (source (origin
 572               (method url-fetch)
 573               (uri (cran-uri "AdaptiveSparsity" version))
 574               (sha256
 575                (base32
 576                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 577     (properties
 578      `((upstream-name . "AdaptiveSparsity")))
 579     (build-system r-build-system)
 580     (arguments
 581      `(#:phases
 582        (modify-phases %standard-phases
 583          (add-after 'unpack 'link-against-armadillo
 584            (lambda _
 585              (substitute* "src/Makevars"
 586                (("PKG_LIBS=" prefix)
 587                 (string-append prefix "-larmadillo"))))))))
 588     (propagated-inputs
 589      `(("r-mass" ,r-mass)
 590        ("r-matrix" ,r-matrix)
 591        ("r-rcpp" ,r-rcpp)
 592        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 593     (inputs
 594      `(("armadillo" ,armadillo)))
 595     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 596     (synopsis "Adaptive sparsity models")
 597     (description
 598      "This package implements the Figueiredo machine learning algorithm for
 599 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 600 geometric models.")
 601     (license license:lgpl3+)))
 602
 603 (define-public r-kernlab
 604   (package
 605     (name "r-kernlab")
 606     (version "0.9-27")
 607     (source
 608      (origin
 609        (method url-fetch)
 610        (uri (cran-uri "kernlab" version))
 611        (sha256
 612         (base32
 613          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 614     (build-system r-build-system)
 615     (home-page "https://cran.r-project.org/web/packages/kernlab")
 616     (synopsis "Kernel-based machine learning tools")
 617     (description
 618      "This package provides kernel-based machine learning methods for
 619 classification, regression, clustering, novelty detection, quantile regression
 620 and dimensionality reduction.  Among other methods @code{kernlab} includes
 621 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 622 and a QP solver.")
 623     (license license:gpl2)))
 624
 625 (define-public dlib
 626   (package
 627     (name "dlib")
 628     (version "19.7")
 629     (source (origin
 630               (method url-fetch)
 631               (uri (string-append
 632                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 633               (sha256
 634                (base32
 635                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 636               (modules '((guix build utils)))
 637               (snippet
 638                '(begin
 639                   ;; Delete ~13MB of bundled dependencies.
 640                   (delete-file-recursively "dlib/external")
 641                   (delete-file-recursively "docs/dlib/external")
 642                   #t))))
 643     (build-system cmake-build-system)
 644     (arguments
 645      `(#:phases
 646        (modify-phases %standard-phases
 647          (add-after 'unpack 'disable-asserts
 648            (lambda _
 649              ;; config.h recommends explicitly enabling or disabling asserts
 650              ;; when building as a shared library. By default neither is set.
 651              (substitute* "dlib/config.h"
 652                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 653              #t))
 654          (add-after 'disable-asserts 'disable-failing-tests
 655            (lambda _
 656              ;; One test times out on MIPS, so we need to disable it.
 657              ;; Others are flaky on some platforms.
 658              (let* ((system ,(or (%current-target-system)
 659                                  (%current-system)))
 660                     (disabled-tests (cond
 661                                      ((string-prefix? "mips64" system)
 662                                       '("object_detector" ; timeout
 663                                         "data_io"))
 664                                      ((string-prefix? "armhf" system)
 665                                       '("learning_to_track"))
 666                                      ((string-prefix? "i686" system)
 667                                       '("optimization"))
 668                                      (else '()))))
 669                (for-each
 670                 (lambda (test)
 671                   (substitute* "dlib/test/makefile"
 672                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 673                 disabled-tests)
 674                #t)))
 675          (replace 'check
 676            (lambda _
 677              ;; No test target, so we build and run the unit tests here.
 678              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 679                (with-directory-excursion test-dir
 680                  (invoke "make" "-j" (number->string (parallel-job-count)))
 681                  (invoke "./dtest" "--runall"))
 682                #t)))
 683          (add-after 'install 'delete-static-library
 684            (lambda* (#:key outputs #:allow-other-keys)
 685              (delete-file (string-append (assoc-ref outputs "out")
 686                                          "/lib/libdlib.a"))
 687              #t)))))
 688     (native-inputs
 689      `(("pkg-config" ,pkg-config)
 690        ;; For tests.
 691        ("libnsl" ,libnsl)))
 692     (inputs
 693      `(("giflib" ,giflib)
 694        ("lapack" ,lapack)
 695        ("libjpeg" ,libjpeg)
 696        ("libpng" ,libpng)
 697        ("libx11" ,libx11)
 698        ("openblas" ,openblas)
 699        ("zlib" ,zlib)))
 700     (synopsis
 701      "Toolkit for making machine learning and data analysis applications in C++")
 702     (description
 703      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 704 tools.  It is used in both industry and academia in a wide range of domains
 705 including robotics, embedded devices, mobile phones, and large high performance
 706 computing environments.")
 707     (home-page "http://dlib.net")
 708     (license license:boost1.0)))
 709
 710 (define-public python-scikit-learn
 711   (package
 712     (name "python-scikit-learn")
 713     (version "0.20.1")
 714     (source
 715      (origin
 716        (method git-fetch)
 717        (uri (git-reference
 718              (url "https://github.com/scikit-learn/scikit-learn.git")
 719              (commit version)))
 720        (file-name (git-file-name name version))
 721        (sha256
 722         (base32
 723          "0fkhwg3xn1s7ln9q1szq6kwc4jhwvjh8w4kmv9wcrqy7cq3lbv0d"))))
 724     (build-system python-build-system)
 725     (arguments
 726      `(#:phases
 727        (modify-phases %standard-phases
 728          (add-after 'build 'build-ext
 729            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 730          (replace 'check
 731            (lambda _
 732              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 733              (setenv "OPENBLAS_NUM_THREADS" "1")
 734
 735              ;; Some tests require write access to $HOME.
 736              (setenv "HOME" "/tmp")
 737
 738              (invoke "pytest" "sklearn" "-m" "not network")))
 739          ;; FIXME: This fails with permission denied
 740          (delete 'reset-gzip-timestamps))))
 741     (inputs
 742      `(("openblas" ,openblas)))
 743     (native-inputs
 744      `(("python-pytest" ,python-pytest)
 745        ("python-pandas" ,python-pandas) ;for tests
 746        ("python-cython" ,python-cython)))
 747     (propagated-inputs
 748      `(("python-numpy" ,python-numpy)
 749        ("python-scipy" ,python-scipy)))
 750     (home-page "http://scikit-learn.org/")
 751     (synopsis "Machine Learning in Python")
 752     (description
 753      "Scikit-learn provides simple and efficient tools for data mining and
 754 data analysis.")
 755     (license license:bsd-3)))
 756
 757 (define-public python2-scikit-learn
 758   (package-with-python2 python-scikit-learn))
 759
 760 (define-public python-autograd
 761   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 762          (revision "0")
 763          (version (git-version "0.0.0" revision commit)))
 764     (package
 765       (name "python-autograd")
 766       (home-page "https://github.com/HIPS/autograd")
 767       (source (origin
 768                 (method git-fetch)
 769                 (uri (git-reference
 770                       (url home-page)
 771                       (commit commit)))
 772                 (sha256
 773                  (base32
 774                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 775                 (file-name (git-file-name name version))))
 776       (version version)
 777       (build-system python-build-system)
 778       (native-inputs
 779        `(("python-nose" ,python-nose)
 780          ("python-pytest" ,python-pytest)))
 781       (propagated-inputs
 782        `(("python-future" ,python-future)
 783          ("python-numpy" ,python-numpy)))
 784       (arguments
 785        `(#:phases (modify-phases %standard-phases
 786                     (replace 'check
 787                       (lambda _
 788                         (invoke "py.test" "-v"))))))
 789       (synopsis "Efficiently computes derivatives of NumPy code")
 790       (description "Autograd can automatically differentiate native Python and
 791 NumPy code.  It can handle a large subset of Python's features, including loops,
 792 ifs, recursion and closures, and it can even take derivatives of derivatives
 793 of derivatives.  It supports reverse-mode differentiation
 794 (a.k.a. backpropagation), which means it can efficiently take gradients of
 795 scalar-valued functions with respect to array-valued arguments, as well as
 796 forward-mode differentiation, and the two can be composed arbitrarily.  The
 797 main intended application of Autograd is gradient-based optimization.")
 798       (license license:expat))))
 799
 800 (define-public python2-autograd
 801   (package-with-python2 python-autograd))
 802
 803 (define-public lightgbm
 804   (package
 805     (name "lightgbm")
 806     (version "2.0.12")
 807     (source (origin
 808               (method url-fetch)
 809               (uri (string-append
 810                     "https://github.com/Microsoft/LightGBM/archive/v"
 811                     version ".tar.gz"))
 812               (sha256
 813                (base32
 814                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 815               (file-name (string-append name "-" version ".tar.gz"))))
 816     (native-inputs
 817      `(("python-pytest" ,python-pytest)
 818        ("python-nose" ,python-nose)))
 819     (inputs
 820      `(("openmpi" ,openmpi)))
 821     (propagated-inputs
 822      `(("python-numpy" ,python-numpy)
 823        ("python-scipy" ,python-scipy)))
 824     (arguments
 825      `(#:configure-flags
 826        '("-DUSE_MPI=ON")
 827        #:phases
 828        (modify-phases %standard-phases
 829          (replace 'check
 830            (lambda* (#:key outputs #:allow-other-keys)
 831              (with-directory-excursion ,(string-append "../LightGBM-" version)
 832                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 833     (build-system cmake-build-system)
 834     (home-page "https://github.com/Microsoft/LightGBM")
 835     (synopsis "Gradient boosting framework based on decision tree algorithms")
 836     (description "LightGBM is a gradient boosting framework that uses tree
 837 based learning algorithms.  It is designed to be distributed and efficient with
 838 the following advantages:
 839
 840 @itemize
 841 @item Faster training speed and higher efficiency
 842 @item Lower memory usage
 843 @item Better accuracy
 844 @item Parallel and GPU learning supported (not enabled in this package)
 845 @item Capable of handling large-scale data
 846 @end itemize\n")
 847     (license license:expat)))
 848
 849 (define-public vowpal-wabbit
 850   ;; Language bindings not included.
 851   (package
 852     (name "vowpal-wabbit")
 853     (version "8.5.0")
 854     (source (origin
 855               (method url-fetch)
 856               (uri (string-append
 857                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 858                     version ".tar.gz"))
 859               (sha256
 860                (base32
 861                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 862               (file-name (string-append name "-" version ".tar.gz"))))
 863     (inputs
 864      `(("boost" ,boost)
 865        ("zlib" ,zlib)))
 866     (arguments
 867      `(#:configure-flags
 868        (list (string-append "--with-boost="
 869                             (assoc-ref %build-inputs "boost")))))
 870     (build-system gnu-build-system)
 871     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 872     (synopsis "Fast machine learning library for online learning")
 873     (description "Vowpal Wabbit is a machine learning system with techniques
 874 such as online, hashing, allreduce, reductions, learning2search, active, and
 875 interactive learning.")
 876     (license license:bsd-3)))
 877
 878 (define-public python2-fastlmm
 879   (package
 880     (name "python2-fastlmm")
 881     (version "0.2.21")
 882     (source
 883      (origin
 884        (method url-fetch)
 885        (uri (pypi-uri "fastlmm" version ".zip"))
 886        (sha256
 887         (base32
 888          "1q8c34rpmwkfy3r4d5172pzdkpfryj561897z9r3x22gq7813x1m"))))
 889     (build-system python-build-system)
 890     (arguments
 891      `(#:python ,python-2)) ; only Python 2.7 is supported
 892     (propagated-inputs
 893      `(("python2-numpy" ,python2-numpy)
 894        ("python2-scipy" ,python2-scipy)
 895        ("python2-matplotlib" ,python2-matplotlib)
 896        ("python2-pandas" ,python2-pandas)
 897        ("python2-scikit-learn" ,python2-scikit-learn)
 898        ("python2-pysnptools" ,python2-pysnptools)))
 899     (native-inputs
 900      `(("unzip" ,unzip)
 901        ("python2-cython" ,python2-cython)
 902        ("python2-mock" ,python2-mock)
 903        ("python2-nose" ,python2-nose)))
 904     (home-page "http://research.microsoft.com/en-us/um/redmond/projects/mscompbio/fastlmm/")
 905     (synopsis "Perform genome-wide association studies on large data sets")
 906     (description
 907      "FaST-LMM, which stands for Factored Spectrally Transformed Linear Mixed
 908 Models, is a program for performing both single-SNP and SNP-set genome-wide
 909 association studies (GWAS) on extremely large data sets.")
 910     (license license:asl2.0)))