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        (sha256
 338         (base32
 339          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 340     (build-system gnu-build-system)
 341     (arguments
 342      `(#:configure-flags
 343        (list (string-append "--with-boost="
 344                             (assoc-ref %build-inputs "boost")))
 345        #:phases
 346        (modify-phases %standard-phases
 347          (add-before
 348           'configure 'set-CXXFLAGS
 349           (lambda _
 350             (setenv "CXXFLAGS" "-fpermissive ")
 351             #t)))))
 352     (inputs
 353      `(("boost" ,boost)
 354        ("gsl" ,gsl)
 355        ("libxml2" ,libxml2)
 356        ("zlib" ,zlib)))
 357     (native-inputs
 358      `(("gfortran" ,gfortran)
 359        ("gfortran:lib" ,gfortran "lib")))
 360     ;; Non-portable assembly instructions are used so building fails on
 361     ;; platforms other than x86_64 or i686.
 362     (supported-systems '("x86_64-linux" "i686-linux"))
 363     (home-page "https://www.imbs.uni-luebeck.de/forschung/software/details.html#c224")
 364     (synopsis "Implementation of the Random Forests machine learning method")
 365     (description
 366      "Random Jungle is an implementation of Random Forests.  It is supposed to
 367 analyse high dimensional data.  In genetics, it can be used for analysing big
 368 Genome Wide Association (GWA) data.  Random Forests is a powerful machine
 369 learning method.  Most interesting features are variable selection, missing
 370 value imputation, classifier creation, generalization error estimation and
 371 sample proximities between pairs of cases.")
 372     (license license:gpl3+)))
 373
 374 (define-public shogun
 375   (package
 376     (name "shogun")
 377     (version "6.1.3")
 378     (source
 379      (origin
 380        (method url-fetch)
 381        (uri (string-append
 382              "ftp://shogun-toolbox.org/shogun/releases/"
 383              (version-major+minor version)
 384              "/sources/shogun-" version ".tar.bz2"))
 385        (sha256
 386         (base32
 387          "1rn9skm3nw6hr7mr3lgp2gfqhi7ii0lyxck7qmqnf8avq349s5jp"))
 388        (modules '((guix build utils)
 389                   (ice-9 rdelim)))
 390        (snippet
 391         '(begin
 392            ;; Remove non-free sources and files referencing them
 393            (for-each delete-file
 394                      (find-files "src/shogun/classifier/svm/"
 395                                  "SVMLight\\.(cpp|h)"))
 396            (for-each delete-file
 397                      (find-files "examples/undocumented/libshogun/"
 398                                  (string-append
 399                                   "(classifier_.*svmlight.*|"
 400                                   "evaluation_cross_validation_locked_comparison).cpp")))
 401            ;; Remove non-free functions.
 402            (define (delete-ifdefs file)
 403              (with-atomic-file-replacement file
 404                (lambda (in out)
 405                  (let loop ((line (read-line in 'concat))
 406                             (skipping? #f))
 407                    (if (eof-object? line)
 408                        #t
 409                        (let ((skip-next?
 410                               (or (and skipping?
 411                                        (not (string-prefix?
 412                                              "#endif //USE_SVMLIGHT" line)))
 413                                   (string-prefix?
 414                                    "#ifdef USE_SVMLIGHT" line))))
 415                          (when (or (not skipping?)
 416                                    (and skipping? (not skip-next?)))
 417                            (display line out))
 418                          (loop (read-line in 'concat) skip-next?)))))))
 419            (for-each delete-ifdefs
 420                      (append
 421                       (find-files "src/shogun/classifier/mkl"
 422                                   "^MKLClassification\\.cpp")
 423                       (find-files "src/shogun/classifier/svm"
 424                                   "^SVMLightOneClass\\.(cpp|h)")
 425                       (find-files "src/shogun/multiclass"
 426                                   "^ScatterSVM\\.(cpp|h)")
 427                       (find-files "src/shogun/kernel/"
 428                                   "^(Kernel|CombinedKernel|ProductKernel)\\.(cpp|h)")
 429                       (find-files "src/shogun/regression/svr"
 430                                   "^(MKLRegression|SVRLight)\\.(cpp|h)")
 431                       (find-files "src/shogun/transfer/domain_adaptation"
 432                                   "^DomainAdaptationSVM\\.(cpp|h)")))
 433            #t))))
 434     (build-system cmake-build-system)
 435     (arguments
 436      '(#:tests? #f ;no check target
 437        #:phases
 438        (modify-phases %standard-phases
 439          (add-after 'unpack 'delete-broken-symlinks
 440            (lambda _
 441              (for-each delete-file '("applications/arts/data"
 442                                      "applications/asp/data"
 443                                      "applications/easysvm/data"
 444                                      "applications/msplicer/data"
 445                                      "applications/ocr/data"
 446                                      "examples/meta/data"
 447                                      "examples/undocumented/data"))
 448              #t))
 449          (add-after 'unpack 'change-R-target-path
 450            (lambda* (#:key outputs #:allow-other-keys)
 451              (substitute* '("src/interfaces/r/CMakeLists.txt"
 452                             "examples/meta/r/CMakeLists.txt")
 453                (("\\$\\{R_COMPONENT_LIB_PATH\\}")
 454                 (string-append (assoc-ref outputs "out")
 455                                "/lib/R/library/")))
 456              #t))
 457          (add-after 'unpack 'fix-octave-modules
 458            (lambda* (#:key outputs #:allow-other-keys)
 459              (substitute* "src/interfaces/octave/CMakeLists.txt"
 460                (("^include_directories\\(\\$\\{OCTAVE_INCLUDE_DIRS\\}")
 461                 "include_directories(${OCTAVE_INCLUDE_DIRS} ${OCTAVE_INCLUDE_DIRS}/octave")
 462                ;; change target directory
 463                (("\\$\\{OCTAVE_OCT_LOCAL_API_FILE_DIR\\}")
 464                 (string-append (assoc-ref outputs "out")
 465                                "/share/octave/packages")))
 466              (substitute* '("src/interfaces/octave/swig_typemaps.i"
 467                             "src/interfaces/octave/sg_print_functions.cpp")
 468                ;; "octave/config.h" and "octave/oct-obj.h" deprecated in Octave.
 469                (("octave/config\\.h") "octave/octave-config.h")
 470                (("octave/oct-obj.h") "octave/ovl.h"))
 471              #t))
 472          (add-after 'unpack 'move-rxcpp
 473            (lambda* (#:key inputs #:allow-other-keys)
 474              (let ((rxcpp-dir "shogun/third-party/rxcpp"))
 475                (mkdir-p rxcpp-dir)
 476                (install-file (assoc-ref inputs "rxcpp") rxcpp-dir)
 477                #t)))
 478          (add-before 'build 'set-HOME
 479            ;; $HOME needs to be set at some point during the build phase
 480            (lambda _ (setenv "HOME" "/tmp") #t)))
 481        #:configure-flags
 482        (list "-DCMAKE_BUILD_WITH_INSTALL_RPATH=TRUE"
 483              "-DUSE_SVMLIGHT=OFF" ;disable proprietary SVMLIGHT
 484              "-DBUILD_META_EXAMPLES=OFF" ;requires unpackaged ctags
 485              ;;"-DINTERFACE_JAVA=ON" ;requires unpackaged jblas
 486              ;;"-DINTERFACE_RUBY=ON" ;requires unpackaged ruby-narray
 487              ;;"-DINTERFACE_PERL=ON" ;"FindPerlLibs" does not exist
 488              ;;"-DINTERFACE_LUA=ON"  ;fails because lua doesn't build pkgconfig file
 489              "-DINTERFACE_OCTAVE=ON"
 490              "-DINTERFACE_PYTHON=ON"
 491              "-DINTERFACE_R=ON")))
 492     (inputs
 493      `(("python" ,python)
 494        ("numpy" ,python-numpy)
 495        ("r-minimal" ,r-minimal)
 496        ("octave" ,octave)
 497        ("swig" ,swig)
 498        ("eigen" ,eigen)
 499        ("hdf5" ,hdf5)
 500        ("atlas" ,atlas)
 501        ("arpack" ,arpack-ng)
 502        ("lapack" ,lapack)
 503        ("glpk" ,glpk)
 504        ("libxml2" ,libxml2)
 505        ("lzo" ,lzo)
 506        ("zlib" ,zlib)))
 507     (native-inputs
 508      `(("pkg-config" ,pkg-config)
 509        ("rxcpp" ,rxcpp)))
 510     ;; Non-portable SSE instructions are used so building fails on platforms
 511     ;; other than x86_64.
 512     (supported-systems '("x86_64-linux"))
 513     (home-page "http://shogun-toolbox.org/")
 514     (synopsis "Machine learning toolbox")
 515     (description
 516      "The Shogun Machine learning toolbox provides a wide range of unified and
 517 efficient Machine Learning (ML) methods.  The toolbox seamlessly allows to
 518 combine multiple data representations, algorithm classes, and general purpose
 519 tools.  This enables both rapid prototyping of data pipelines and extensibility
 520 in terms of new algorithms.")
 521     (license license:gpl3+)))
 522
 523 (define-public rxcpp
 524   (package
 525     (name "rxcpp")
 526     (version "4.0.0")
 527     (source
 528      (origin
 529        (method url-fetch)
 530        (uri (string-append "https://github.com/ReactiveX/RxCpp/archive/v"
 531                            version ".tar.gz"))
 532        (sha256
 533         (base32
 534          "0y2isr8dy2n1yjr9c5570kpc9lvdlch6jv0jvw000amwn5d3krsh"))
 535        (file-name (string-append name "-" version ".tar.gz"))))
 536     (build-system cmake-build-system)
 537     (arguments
 538      `(#:phases
 539        (modify-phases %standard-phases
 540          (add-after 'unpack 'remove-werror
 541            (lambda _
 542              (substitute* (find-files ".")
 543                (("-Werror") ""))
 544              #t))
 545          (replace 'check
 546            (lambda _
 547              (invoke "ctest"))))))
 548     (native-inputs
 549      `(("catch" ,catch-framework)))
 550     (home-page "http://reactivex.io/")
 551     (synopsis "Reactive Extensions for C++")
 552     (description
 553      "The Reactive Extensions for C++ (RxCpp) is a library of algorithms for
 554 values-distributed-in-time.  ReactiveX is a library for composing asynchronous
 555 and event-based programs by using observable sequences.
 556
 557 It extends the observer pattern to support sequences of data and/or events and
 558 adds operators that allow you to compose sequences together declaratively while
 559 abstracting away concerns about things like low-level threading,
 560 synchronization, thread-safety, concurrent data structures, and non-blocking
 561 I/O.")
 562     (license license:asl2.0)))
 563
 564 (define-public r-adaptivesparsity
 565   (package
 566     (name "r-adaptivesparsity")
 567     (version "1.6")
 568     (source (origin
 569               (method url-fetch)
 570               (uri (cran-uri "AdaptiveSparsity" version))
 571               (sha256
 572                (base32
 573                 "0imr5m8mll9j6n4icsv6z9rl5kbnwsp9wvzrg7n90nnmcxq2cz91"))))
 574     (properties
 575      `((upstream-name . "AdaptiveSparsity")))
 576     (build-system r-build-system)
 577     (arguments
 578      `(#:phases
 579        (modify-phases %standard-phases
 580          (add-after 'unpack 'link-against-armadillo
 581            (lambda _
 582              (substitute* "src/Makevars"
 583                (("PKG_LIBS=" prefix)
 584                 (string-append prefix "-larmadillo"))))))))
 585     (propagated-inputs
 586      `(("r-mass" ,r-mass)
 587        ("r-matrix" ,r-matrix)
 588        ("r-rcpp" ,r-rcpp)
 589        ("r-rcpparmadillo" ,r-rcpparmadillo)))
 590     (inputs
 591      `(("armadillo" ,armadillo)))
 592     (home-page "https://cran.r-project.org/web/packages/AdaptiveSparsity")
 593     (synopsis "Adaptive sparsity models")
 594     (description
 595      "This package implements the Figueiredo machine learning algorithm for
 596 adaptive sparsity and the Wong algorithm for adaptively sparse gaussian
 597 geometric models.")
 598     (license license:lgpl3+)))
 599
 600 (define-public r-kernlab
 601   (package
 602     (name "r-kernlab")
 603     (version "0.9-27")
 604     (source
 605      (origin
 606        (method url-fetch)
 607        (uri (cran-uri "kernlab" version))
 608        (sha256
 609         (base32
 610          "1m0xqf6gyvwayz7w3c83y32ayvnlz0jicj8ijk808zq9sh7dbbgn"))))
 611     (build-system r-build-system)
 612     (home-page "https://cran.r-project.org/web/packages/kernlab")
 613     (synopsis "Kernel-based machine learning tools")
 614     (description
 615      "This package provides kernel-based machine learning methods for
 616 classification, regression, clustering, novelty detection, quantile regression
 617 and dimensionality reduction.  Among other methods @code{kernlab} includes
 618 Support Vector Machines, Spectral Clustering, Kernel PCA, Gaussian Processes
 619 and a QP solver.")
 620     (license license:gpl2)))
 621
 622 (define-public dlib
 623   (package
 624     (name "dlib")
 625     (version "19.7")
 626     (source (origin
 627               (method url-fetch)
 628               (uri (string-append
 629                     "http://dlib.net/files/dlib-" version ".tar.bz2"))
 630               (sha256
 631                (base32
 632                 "1mljz02kwkrbggyncxv5fpnyjdybw2qihaacb3js8yfkw12vwpc2"))
 633               (modules '((guix build utils)))
 634               (snippet
 635                '(begin
 636                   ;; Delete ~13MB of bundled dependencies.
 637                   (delete-file-recursively "dlib/external")
 638                   (delete-file-recursively "docs/dlib/external")
 639                   #t))))
 640     (build-system cmake-build-system)
 641     (arguments
 642      `(#:phases
 643        (modify-phases %standard-phases
 644          (add-after 'unpack 'disable-asserts
 645            (lambda _
 646              ;; config.h recommends explicitly enabling or disabling asserts
 647              ;; when building as a shared library. By default neither is set.
 648              (substitute* "dlib/config.h"
 649                (("^//#define DLIB_DISABLE_ASSERTS") "#define DLIB_DISABLE_ASSERTS"))
 650              #t))
 651          (add-after 'disable-asserts 'disable-failing-tests
 652            (lambda _
 653              ;; One test times out on MIPS, so we need to disable it.
 654              ;; Others are flaky on some platforms.
 655              (let* ((system ,(or (%current-target-system)
 656                                  (%current-system)))
 657                     (disabled-tests (cond
 658                                      ((string-prefix? "mips64" system)
 659                                       '("object_detector" ; timeout
 660                                         "data_io"))
 661                                      ((string-prefix? "armhf" system)
 662                                       '("learning_to_track"))
 663                                      ((string-prefix? "i686" system)
 664                                       '("optimization"))
 665                                      (else '()))))
 666                (for-each
 667                 (lambda (test)
 668                   (substitute* "dlib/test/makefile"
 669                     (((string-append "SRC \\+= " test "\\.cpp")) "")))
 670                 disabled-tests)
 671                #t)))
 672          (replace 'check
 673            (lambda _
 674              ;; No test target, so we build and run the unit tests here.
 675              (let ((test-dir (string-append "../dlib-" ,version "/dlib/test")))
 676                (with-directory-excursion test-dir
 677                  (invoke "make" "-j" (number->string (parallel-job-count)))
 678                  (invoke "./dtest" "--runall"))
 679                #t)))
 680          (add-after 'install 'delete-static-library
 681            (lambda* (#:key outputs #:allow-other-keys)
 682              (delete-file (string-append (assoc-ref outputs "out")
 683                                          "/lib/libdlib.a"))
 684              #t)))))
 685     (native-inputs
 686      `(("pkg-config" ,pkg-config)
 687        ;; For tests.
 688        ("libnsl" ,libnsl)))
 689     (inputs
 690      `(("giflib" ,giflib)
 691        ("lapack" ,lapack)
 692        ("libjpeg" ,libjpeg)
 693        ("libpng" ,libpng)
 694        ("libx11" ,libx11)
 695        ("openblas" ,openblas)
 696        ("zlib" ,zlib)))
 697     (synopsis
 698      "Toolkit for making machine learning and data analysis applications in C++")
 699     (description
 700      "Dlib is a modern C++ toolkit containing machine learning algorithms and
 701 tools.  It is used in both industry and academia in a wide range of domains
 702 including robotics, embedded devices, mobile phones, and large high performance
 703 computing environments.")
 704     (home-page "http://dlib.net")
 705     (license license:boost1.0)))
 706
 707 (define-public python-scikit-learn
 708   (package
 709     (name "python-scikit-learn")
 710     (version "0.19.2")
 711     (source
 712      (origin
 713        (method git-fetch)
 714        (uri (git-reference
 715              (url "https://github.com/scikit-learn/scikit-learn.git")
 716              (commit version)))
 717        (file-name (git-file-name name version))
 718        (sha256
 719         (base32
 720          "1dk9hdj01c0bny4ps78b7869fjw9gr6qklxf6wyql8h6nh4k19xm"))
 721        (patches (search-patches
 722                  "python-scikit-learn-fix-test-non-determinism.patch"))))
 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              ;; Disable tests that require network access
 734              (delete-file "sklearn/datasets/tests/test_mldata.py")
 735              (delete-file "sklearn/datasets/tests/test_rcv1.py")
 736              (invoke "pytest" "sklearn")
 737              #t))
 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   (let ((parent (package-with-python2 python-scikit-learn)))
 758     (package (inherit parent)
 759       (arguments
 760        (substitute-keyword-arguments (package-arguments parent)
 761          ((#:phases phases)
 762           `(modify-phases ,phases
 763              (replace 'check
 764                (lambda _
 765                  ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 766                  (setenv "OPENBLAS_NUM_THREADS" "1")
 767                  ;; Some tests expect to be able to write to HOME.
 768                  (setenv "HOME" "/tmp")
 769                  ;; Disable tests that require network access
 770                  (delete-file "sklearn/datasets/tests/test_kddcup99.py")
 771                  (delete-file "sklearn/datasets/tests/test_mldata.py")
 772                  (delete-file "sklearn/datasets/tests/test_rcv1.py")
 773                  (invoke "pytest" "sklearn")
 774                  #t)))))))))
 775
 776 (define-public python-autograd
 777   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 778          (revision "0")
 779          (version (git-version "0.0.0" revision commit)))
 780     (package
 781       (name "python-autograd")
 782       (home-page "https://github.com/HIPS/autograd")
 783       (source (origin
 784                 (method git-fetch)
 785                 (uri (git-reference
 786                       (url home-page)
 787                       (commit commit)))
 788                 (sha256
 789                  (base32
 790                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 791                 (file-name (git-file-name name version))))
 792       (version version)
 793       (build-system python-build-system)
 794       (native-inputs
 795        `(("python-nose" ,python-nose)
 796          ("python-pytest" ,python-pytest)))
 797       (propagated-inputs
 798        `(("python-future" ,python-future)
 799          ("python-numpy" ,python-numpy)))
 800       (arguments
 801        `(#:phases (modify-phases %standard-phases
 802                     (replace 'check
 803                       (lambda _
 804                         (invoke "py.test" "-v"))))))
 805       (synopsis "Efficiently computes derivatives of NumPy code")
 806       (description "Autograd can automatically differentiate native Python and
 807 NumPy code.  It can handle a large subset of Python's features, including loops,
 808 ifs, recursion and closures, and it can even take derivatives of derivatives
 809 of derivatives.  It supports reverse-mode differentiation
 810 (a.k.a. backpropagation), which means it can efficiently take gradients of
 811 scalar-valued functions with respect to array-valued arguments, as well as
 812 forward-mode differentiation, and the two can be composed arbitrarily.  The
 813 main intended application of Autograd is gradient-based optimization.")
 814       (license license:expat))))
 815
 816 (define-public python2-autograd
 817   (package-with-python2 python-autograd))
 818
 819 (define-public lightgbm
 820   (package
 821     (name "lightgbm")
 822     (version "2.0.12")
 823     (source (origin
 824               (method url-fetch)
 825               (uri (string-append
 826                     "https://github.com/Microsoft/LightGBM/archive/v"
 827                     version ".tar.gz"))
 828               (sha256
 829                (base32
 830                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 831               (file-name (string-append name "-" version ".tar.gz"))))
 832     (native-inputs
 833      `(("python-pytest" ,python-pytest)
 834        ("python-nose" ,python-nose)))
 835     (inputs
 836      `(("openmpi" ,openmpi)))
 837     (propagated-inputs
 838      `(("python-numpy" ,python-numpy)
 839        ("python-scipy" ,python-scipy)))
 840     (arguments
 841      `(#:configure-flags
 842        '("-DUSE_MPI=ON")
 843        #:phases
 844        (modify-phases %standard-phases
 845          (replace 'check
 846            (lambda* (#:key outputs #:allow-other-keys)
 847              (with-directory-excursion ,(string-append "../LightGBM-" version)
 848                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 849     (build-system cmake-build-system)
 850     (home-page "https://github.com/Microsoft/LightGBM")
 851     (synopsis "Gradient boosting framework based on decision tree algorithms")
 852     (description "LightGBM is a gradient boosting framework that uses tree
 853 based learning algorithms.  It is designed to be distributed and efficient with
 854 the following advantages:
 855
 856 @itemize
 857 @item Faster training speed and higher efficiency
 858 @item Lower memory usage
 859 @item Better accuracy
 860 @item Parallel and GPU learning supported (not enabled in this package)
 861 @item Capable of handling large-scale data
 862 @end itemize\n")
 863     (license license:expat)))
 864
 865 (define-public vowpal-wabbit
 866   ;; Language bindings not included.
 867   (package
 868     (name "vowpal-wabbit")
 869     (version "8.5.0")
 870     (source (origin
 871               (method url-fetch)
 872               (uri (string-append
 873                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 874                     version ".tar.gz"))
 875               (sha256
 876                (base32
 877                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 878               (file-name (string-append name "-" version ".tar.gz"))))
 879     (inputs
 880      `(("boost" ,boost)
 881        ("zlib" ,zlib)))
 882     (arguments
 883      `(#:configure-flags
 884        (list (string-append "--with-boost="
 885                             (assoc-ref %build-inputs "boost")))))
 886     (build-system gnu-build-system)
 887     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 888     (synopsis "Fast machine learning library for online learning")
 889     (description "Vowpal Wabbit is a machine learning system with techniques
 890 such as online, hashing, allreduce, reductions, learning2search, active, and
 891 interactive learning.")
 892     (license license:bsd-3)))