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 randomjungle
 326   (package
 327     (name "randomjungle")
 328     (version "2.1.0")
 329     (source
 330      (origin
 331        (method url-fetch)
 332        (uri (string-append
 333              "https://www.imbs.uni-luebeck.de/fileadmin/files/Software"
 334              "/randomjungle/randomjungle-" version ".tar_.gz"))
 335        (patches (search-patches "randomjungle-disable-static-build.patch"))
 336        (sha256
 337         (base32
 338          "12c8rf30cla71swx2mf4ww9mfd8jbdw5lnxd7dxhyw1ygrvg6y4w"))))
 339     (build-system gnu-build-system)
 340     (arguments
 341      `(#:configure-flags
 342        (list "--disable-static"
 343              (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-cli)
 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 "0.20.1")
 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          "0fkhwg3xn1s7ln9q1szq6kwc4jhwvjh8w4kmv9wcrqy7cq3lbv0d"))))
 721     (build-system python-build-system)
 722     (arguments
 723      `(#:phases
 724        (modify-phases %standard-phases
 725          (add-after 'build 'build-ext
 726            (lambda _ (invoke "python" "setup.py" "build_ext" "--inplace") #t))
 727          (replace 'check
 728            (lambda _
 729              ;; Restrict OpenBLAS threads to prevent segfaults while testing!
 730              (setenv "OPENBLAS_NUM_THREADS" "1")
 731
 732              ;; Some tests require write access to $HOME.
 733              (setenv "HOME" "/tmp")
 734
 735              (invoke "pytest" "sklearn" "-m" "not network")))
 736          ;; FIXME: This fails with permission denied
 737          (delete 'reset-gzip-timestamps))))
 738     (inputs
 739      `(("openblas" ,openblas)))
 740     (native-inputs
 741      `(("python-pytest" ,python-pytest)
 742        ("python-pandas" ,python-pandas) ;for tests
 743        ("python-cython" ,python-cython)))
 744     (propagated-inputs
 745      `(("python-numpy" ,python-numpy)
 746        ("python-scipy" ,python-scipy)))
 747     (home-page "http://scikit-learn.org/")
 748     (synopsis "Machine Learning in Python")
 749     (description
 750      "Scikit-learn provides simple and efficient tools for data mining and
 751 data analysis.")
 752     (license license:bsd-3)))
 753
 754 (define-public python2-scikit-learn
 755   (package-with-python2 python-scikit-learn))
 756
 757 (define-public python-autograd
 758   (let* ((commit "442205dfefe407beffb33550846434baa90c4de7")
 759          (revision "0")
 760          (version (git-version "0.0.0" revision commit)))
 761     (package
 762       (name "python-autograd")
 763       (home-page "https://github.com/HIPS/autograd")
 764       (source (origin
 765                 (method git-fetch)
 766                 (uri (git-reference
 767                       (url home-page)
 768                       (commit commit)))
 769                 (sha256
 770                  (base32
 771                   "189sv2xb0mwnjawa9z7mrgdglc1miaq93pnck26r28fi1jdwg0z4"))
 772                 (file-name (git-file-name name version))))
 773       (version version)
 774       (build-system python-build-system)
 775       (native-inputs
 776        `(("python-nose" ,python-nose)
 777          ("python-pytest" ,python-pytest)))
 778       (propagated-inputs
 779        `(("python-future" ,python-future)
 780          ("python-numpy" ,python-numpy)))
 781       (arguments
 782        `(#:phases (modify-phases %standard-phases
 783                     (replace 'check
 784                       (lambda _
 785                         (invoke "py.test" "-v"))))))
 786       (synopsis "Efficiently computes derivatives of NumPy code")
 787       (description "Autograd can automatically differentiate native Python and
 788 NumPy code.  It can handle a large subset of Python's features, including loops,
 789 ifs, recursion and closures, and it can even take derivatives of derivatives
 790 of derivatives.  It supports reverse-mode differentiation
 791 (a.k.a. backpropagation), which means it can efficiently take gradients of
 792 scalar-valued functions with respect to array-valued arguments, as well as
 793 forward-mode differentiation, and the two can be composed arbitrarily.  The
 794 main intended application of Autograd is gradient-based optimization.")
 795       (license license:expat))))
 796
 797 (define-public python2-autograd
 798   (package-with-python2 python-autograd))
 799
 800 (define-public lightgbm
 801   (package
 802     (name "lightgbm")
 803     (version "2.0.12")
 804     (source (origin
 805               (method url-fetch)
 806               (uri (string-append
 807                     "https://github.com/Microsoft/LightGBM/archive/v"
 808                     version ".tar.gz"))
 809               (sha256
 810                (base32
 811                 "132zf0yk0545mg72hyzxm102g3hpb6ixx9hnf8zd2k55gas6cjj1"))
 812               (file-name (string-append name "-" version ".tar.gz"))))
 813     (native-inputs
 814      `(("python-pytest" ,python-pytest)
 815        ("python-nose" ,python-nose)))
 816     (inputs
 817      `(("openmpi" ,openmpi)))
 818     (propagated-inputs
 819      `(("python-numpy" ,python-numpy)
 820        ("python-scipy" ,python-scipy)))
 821     (arguments
 822      `(#:configure-flags
 823        '("-DUSE_MPI=ON")
 824        #:phases
 825        (modify-phases %standard-phases
 826          (replace 'check
 827            (lambda* (#:key outputs #:allow-other-keys)
 828              (with-directory-excursion ,(string-append "../LightGBM-" version)
 829                (invoke "pytest" "tests/c_api_test/test_.py")))))))
 830     (build-system cmake-build-system)
 831     (home-page "https://github.com/Microsoft/LightGBM")
 832     (synopsis "Gradient boosting framework based on decision tree algorithms")
 833     (description "LightGBM is a gradient boosting framework that uses tree
 834 based learning algorithms.  It is designed to be distributed and efficient with
 835 the following advantages:
 836
 837 @itemize
 838 @item Faster training speed and higher efficiency
 839 @item Lower memory usage
 840 @item Better accuracy
 841 @item Parallel and GPU learning supported (not enabled in this package)
 842 @item Capable of handling large-scale data
 843 @end itemize\n")
 844     (license license:expat)))
 845
 846 (define-public vowpal-wabbit
 847   ;; Language bindings not included.
 848   (package
 849     (name "vowpal-wabbit")
 850     (version "8.5.0")
 851     (source (origin
 852               (method url-fetch)
 853               (uri (string-append
 854                     "https://github.com/JohnLangford/vowpal_wabbit/archive/"
 855                     version ".tar.gz"))
 856               (sha256
 857                (base32
 858                 "0clp2kb7rk5sckhllxjr5a651awf4s8dgzg4659yh4hf5cqnf0gr"))
 859               (file-name (string-append name "-" version ".tar.gz"))))
 860     (inputs
 861      `(("boost" ,boost)
 862        ("zlib" ,zlib)))
 863     (arguments
 864      `(#:configure-flags
 865        (list (string-append "--with-boost="
 866                             (assoc-ref %build-inputs "boost")))))
 867     (build-system gnu-build-system)
 868     (home-page "https://github.com/JohnLangford/vowpal_wabbit")
 869     (synopsis "Fast machine learning library for online learning")
 870     (description "Vowpal Wabbit is a machine learning system with techniques
 871 such as online, hashing, allreduce, reductions, learning2search, active, and
 872 interactive learning.")
 873     (license license:bsd-3)))
 874
 875 (define-public python2-fastlmm
 876   (package
 877     (name "python2-fastlmm")
 878     (version "0.2.21")
 879     (source
 880      (origin
 881        (method url-fetch)
 882        (uri (pypi-uri "fastlmm" version ".zip"))
 883        (sha256
 884         (base32
 885          "1q8c34rpmwkfy3r4d5172pzdkpfryj561897z9r3x22gq7813x1m"))))
 886     (build-system python-build-system)
 887     (arguments
 888      `(#:tests? #f ; some test files are missing
 889        #:python ,python-2)) ; only Python 2.7 is supported
 890     (propagated-inputs
 891      `(("python2-numpy" ,python2-numpy)
 892        ("python2-scipy" ,python2-scipy)
 893        ("python2-matplotlib" ,python2-matplotlib)
 894        ("python2-pandas" ,python2-pandas)
 895        ("python2-scikit-learn" ,python2-scikit-learn)
 896        ("python2-pysnptools" ,python2-pysnptools)))
 897     (native-inputs
 898      `(("unzip" ,unzip)
 899        ("python2-cython" ,python2-cython)
 900        ("python2-mock" ,python2-mock)
 901        ("python2-nose" ,python2-nose)))
 902     (home-page "http://research.microsoft.com/en-us/um/redmond/projects/mscompbio/fastlmm/")
 903     (synopsis "Perform genome-wide association studies on large data sets")
 904     (description
 905      "FaST-LMM, which stands for Factored Spectrally Transformed Linear Mixed
 906 Models, is a program for performing both single-SNP and SNP-set genome-wide
 907 association studies (GWAS) on extremely large data sets.")
 908     (license license:asl2.0)))