4in1

.gitignore

@@ -50,3 +50,6 @@ XCode/
 # Build directory for C++ tests
 cpp_test_build/
 lib/cmake-build-debug
+
+# benchmark data
+tools/benchmark/data

.travis.yml

@@ -15,28 +15,25 @@ matrix:
   - os: linux
     services: docker
     env:
-      - DOCKER_IMAGE=xdatainitiative/tick_ubuntu:1.4
-      - PYVER=3.5.3
+      - DOCKER_IMAGE=xdatainitiative/tick_debian:0.5.1
+      - PYVER=3.6.8
   - os: linux
     services: docker
     env:
-      - DOCKER_IMAGE=xdatainitiative/tick_ubuntu:1.4
-      - PYVER=3.6.1
+      - DOCKER_IMAGE=xdatainitiative/tick_debian:0.5.1
+      - PYVER=3.7.3
   - os: osx
     language: generic
     sudo: required
     env:
-      - PYVER=3.5.3
-    osx_image: xcode9.1
+      - PYVER=3.6.8
+    osx_image: xcode10.2
   - os: osx
     language: generic
     sudo: required
     env:
-      - PYVER=3.6.1
-    osx_image: xcode9.1
-
-before_install:
-  - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then rvm get stable; fi
+      - PYVER=3.7.3
+    osx_image: xcode10.2
 
 install:
   - if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then source tools/travis/osx_install.sh; fi

INSTALL.md

@@ -4,7 +4,7 @@
 
 ### Requirements
 
-_tick_ currently works on Linux/OSX (Windows is experimental) systems and requires Python 3.5 or newer. Please have the required Python dependencies in your Python environment:
+_tick_ currently works on Linux/OSX (Windows is experimental) systems and requires Python 3.6 or newer. Please have the required Python dependencies in your Python environment:
 
 - numpy
 - scipy
@@ -17,11 +17,21 @@ If you build and install _tick_ via _pip_ these dependencies will automatically
 
     pip install -r requirements.txt
 
-[Swig](http://www.swig.org/Doc3.0/SWIGDocumentation.html) might also be necessary if precompiled binaries are not available for your distribution.
+[Swig](http://www.swig.org/Doc4.0/SWIGDocumentation.html) might also be necessary if precompiled binaries are not available for your distribution.
 
 ### Source installations
 
-For source installations, a C++ compiler capable of compiling C++11 source code (such as recent versions of [gcc](https://gcc.gnu.org/) or [clang](https://clang.llvm.org/)) must be available. In addition, SWIG version 3.07 or newer must also be available.
+For source installations, a C++ compiler capable of compiling C++11 source code (such as recent versions of [gcc](https://gcc.gnu.org/) or [clang](https://clang.llvm.org/)) must be available. In addition, SWIG version 4.0 or newer must also be available. It is also recommended to build swig if necessary without PCRE
+
+The following script is provided as is with not assurance for working, and may require "sudo" permissions:
+
+    git clone https://github.com/swig/swig -b rel-4.0.0 swig && \
+    cd swig && ./autogen.sh && ./configure --without-pcre && \
+    make && make install
+
+It also may require the following (styled for a Debian type system)
+
+    apt-get install -y autotools-dev automake gawk bison flex
 
 ### Install using _pip_
 
@@ -56,7 +66,7 @@ This will build all required C++ extensions, and install the extensions in your
   Download python and swig, and add their respective installation directories to the PATH environment variable.
 
     python  3.6   - https://www.python.org/downloads/windows/
-    swigwin 3.0.* - http://www.swig.org/download.html
+    swigwin 4.0.* - http://www.swig.org/download.html
 
   Download wheel files from http://www.lfd.uci.edu/~gohlke/pythonlibs - this is not required for Anaconda
 
@@ -113,6 +123,19 @@ This will build all extensions, and install them directly in the source director
 Note also that special scripts, intended for developers, are available. `./clean_build_test.sh` removes all compiled binaries and runs the full compilation process, with all `C++` and `Python` unit-tests. This can take some time.
 Similarly, `./build_test.sh` runs the full compilation process, without removing first binaries, while `full_clean.sh` only removes them.
 
+It is possible to build with support for Intel MKL in the tick C++ libraries, to do so requires having the intel MKL libraries installed and running the build command like so (example provided is for linux with MKL installed to /opt/intel)
+
+    MKLROOT=/opt/intel/mkl python setup.py build_ext --inplace
+
+On MacOS it's possible an error may occur when trying to find a "intel_thread.dyld"
+To fix this try:
+
+export DYLD_LIBRARY_PATH="/opt/intel/lib:$DYLD_LIBRARY_PATH"
+
+Replacing "/opt/intel/lib" to "$MKLROOT/../lib" if necessary.
+
+Upgrading to the latest version of MKL may also resolve this.
+
 ### Alternative installation method
 
 Also supported is the build tool [maiken](https://github.com/dekken/maiken) which works with various compilers on major platforms.
@@ -138,3 +161,10 @@ Logging from mkn can be enabled with the "KLOG" environment variable
     KLOG=3 # DEBUG
 
 If there is a compile error it should be displayed without KLOG being set.
+
+
+### Python 3.5
+
+Python 3.5 is not officially supported any more, but it should still work.
+To build with it, edit "python_min_ver = (3, 6, 0)" in setup.py.
+

LICENSE.txt

@@ -1,4 +1,4 @@
-Copyright (c) 2015-2017, the tick developers
+Copyright (c) 2015-2018, the tick developers
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

appveyor.yml

@@ -29,50 +29,20 @@ cache:
 
 install:
   - git submodule update --init
-  - curl -o mkn.exe -L https://ci.appveyor.com/api/projects/dekken/maiken/artifacts/bin/build/mkn.exe
-  - IF NOT EXIST C:\ProgramData\chocolatey\bin\swig.exe choco install tools\windows\swig.3.0.12.nupkg --yes --limit-output #> $null
+  - curl -o mkn.exe -L https://github.com/Dekken/maiken/raw/binaries/win10_x64/mkn.exe
+  - IF NOT EXIST C:\ProgramData\chocolatey\bin\swig.exe choco install tools\windows\swig.4.0.0.nupkg --yes --limit-output #> $null
+  - cat C:\ProgramData\chocolatey\logs\chocolatey.log
   - git clone https://github.com/X-DataInitiative/tick_appveyor -b master --depth 1 appveyor
 
 build_script:
   - cp "C:\Program Files (x86)\Windows Kits\10\bin\x64\rc.exe" %CD%
   - cp %CD%\rc.exe %CD%\lib
-  - SET MKN_CL_PREFERRED=1 # forces mkn to use cl even if gcc/clang are found
-  - SET MKN_COMPILE_THREADS=1 # mkn use 1 thread heap space issue
-  - SET SWIG=0 # disables swig for mkn
   - call "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build\vcvars64.bat"
-
+  - cl # print compiler version
   # doesnt exist for somereason?
   - mkdir -p "C:\Program Files (x86)\Windows Kits\10\include\10.0.16299.0\cppwinrt"
 
-  # Python 3.7 TODO
-
-  # Python 3.6
-  - cat appveyor/pip/numpy/numpy-1.14.0+mkl-cp36* > appveyor/pip/numpy-1.14.0+mkl-cp36-cp36m-win_amd64.whl
-  - C:\Python36-x64\Scripts\pip install appveyor\pip\numpy-1.14.0+mkl-cp36-cp36m-win_amd64.whl
-  - C:\Python36-x64\Scripts\pip install appveyor\pip\numpydoc-0.7.0-py2.py3-none-any.whl
-  - C:\Python36-x64\Scripts\pip install appveyor\pip\scipy-1.0.0-cp36-cp36m-win_amd64.whl
-  - C:\Python36-x64\Scripts\pip install -r requirements.txt
-  - C:\Python36-x64\Scripts\pip install tensorflow
-  - C:\Python36-x64\python setup.py build_ext --inplace
-  - C:\Python36-x64\python -m unittest discover -v . "*_test.py"
-  - bash -c "PATH=\"$PWD:/c/Python36-x64/:/c/Program Files/Git/usr/bin:$PATH\"; sh/mkn.sh"
-  - bash -c "PATH=\"$PWD:/c/Python36-x64/:/c/Program Files/Git/usr/bin:$PATH\"; sh/gtest.sh"
-  - C:\Python36-x64\python setup.py bdist_wheel
-  - rm -rf build
-  - rm -rf lib\bin
-
-  # # Python 3.5 # with addition of gtests appveyor is taking more than an hour and timing out
-  # - cat appveyor/pip/numpy/numpy-1.14.0+mkl-cp35* > appveyor/pip/numpy-1.14.0+mkl-cp35-cp35m-win_amd64.whl
-  # - C:\Python35-x64\Scripts\pip install appveyor\pip\numpy-1.14.0+mkl-cp35-cp35m-win_amd64.whl
-  # - C:\Python35-x64\Scripts\pip install appveyor\pip\numpydoc-0.7.0-py2.py3-none-any.whl
-  # - C:\Python35-x64\Scripts\pip install appveyor\pip\scipy-1.0.0-cp35-cp35m-win_amd64.whl
-  # - C:\Python35-x64\Scripts\pip install -r requirements.txt
-  # - C:\Python35-x64\Scripts\pip install tensorflow
-  # - C:\Python35-x64\python setup.py build_ext --inplace
-  # - C:\Python35-x64\python -m unittest discover -v . "*_test.py"
-  # - C:\Python35-x64\python setup.py bdist_wheel
-  # - rm -rf build
-  # - rm -rf lib\bin
+  - bash -c "PATH=\"$PWD:/c/Python37-x64/:/c/Program Files/Git/usr/bin:$PATH\"; tools/appveyor/run.sh"
 
 artifacts:
   - path: "dist\\*.whl"

doc/conf.py

@@ -14,7 +14,7 @@
 
 import sys
 import os
-
+import sphinxext.google_analytics
 try:
   import sphinx_bootstrap_theme
   use_bootstrap_theme = True
@@ -92,9 +92,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = '0.4'
+version = '0.5'
 # The full version, including alpha/beta/rc tags.
-release = '0.4.0'
+release = '0.5.0'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

doc/modules/dev.rst

@@ -312,7 +312,7 @@ Create SWIG file
 ~~~~~~~~~~~~~~~~
 
 Now that our proximal operator is defined in C++ we need to make it available
-in Python. We do it thanks to `SWIG <http://www.swig.org/Doc3.0/>`_.
+in Python. We do it thanks to `SWIG <http://www.swig.org/Doc4.0/>`_.
 Hence we have to create a .i file. In tick we store them in the lib/swig folder.
 
 This .i file looks a lot like our .h file.
@@ -515,6 +515,9 @@ serialized copy of the object, and a `__setstate__` method to reconstruct the
 object from a string value (this is where the initialization/constructor
 parameters play in).
 
+Similarly, `TICK_MAKE_TEMPLATED_PICKLABLE` is provided if needed to specify
+a type with templated parameters.
+
 It's important to consider the initialization of the Python object. In some
 cases it might be convenient to add a parameter-less C++ constructor that
 initializes an empty object. Otherwise existing constructors should be used.

doc/modules/solver.rst

@@ -58,7 +58,6 @@ which outputs
 
     Launching the solver SVRG...
       n_iter  |    obj    |  rel_obj
-            0 |    ...    |    ...
             1 |    ...    |    ...
             2 |    ...    |    ...
             3 |    ...    |    ...
@@ -73,7 +72,6 @@ which outputs
 
     Launching the solver SVRG...
       n_iter  |    obj    |  rel_obj
-            0 |  5.29e-01 |  2.37e-01
             1 |  5.01e-01 |  5.15e-02
             2 |  4.97e-01 |  8.44e-03
             3 |  4.97e-01 |  5.00e-04

doc/modules/survival.rst

@@ -26,6 +26,8 @@ Kaplan-Meier estimators.
    nelson_aalen
    kaplan_meier
    ConvSCCS
+   sccs.BatchConvSCCS
+   sccs.StreamConvSCCS
 
 2. Models
 =========

doc/sphinxext/google_analytics.py

@@ -4,7 +4,7 @@
 https://bitbucket.org/birkenfeld/sphinx-contrib/src/e758073384efd1ed5ed1e6286301b7bef71b27cf/googleanalytics/
 """
 
-from sphinx.application import ExtensionError
+from sphinx.errors import ExtensionError
 
 
 def add_ga_javascript(app, pagename, templatename, context, doctree):

examples/plot_2d_linear_regression.py

@@ -0,0 +1,107 @@
+"""
+===================================
+Linear regression in two dimensions
+===================================
+
+In this example, we try to predict the median price of houses in Boston's
+neighbors by looking at two features: the average of the number of rooms per
+dwelling and the pencentage of low status in the population.
+The linear regression is done using the `Boston Housing Dataset`_ which
+contains 13 features used to predict the median price of houses.
+The two features selected are the most efficent on the test set.
+The 3D representation allows a better understanding of the prediction
+mechanism with two features.
+
+This example is inspired by linear regression example from
+`scikit-learn documentation`_.
+
+.. _Boston Housing Dataset: https://www.kaggle.com/c/boston-housing
+.. _scikit-learn documentation: http://scikit-learn.org/stable/auto_examples/linear_model/plot_ols.html#sphx-glr-auto-examples-linear-model-plot-ols-py
+"""
+
+import matplotlib.pyplot as plt
+import numpy as np
+from tick import linear_model
+from sklearn.metrics import mean_squared_error, r2_score
+from sklearn.utils import shuffle
+from sklearn.datasets import load_boston
+from mpl_toolkits.mplot3d import axes3d
+from matplotlib import cm
+
+# Load the Boston Housing Dataset
+features, label = load_boston(return_X_y=True)
+features, label = shuffle(features, label, random_state=0)
+
+# Use two features: the average of the number of rooms per dwelling and
+# the pencentage of low status of the population
+X = features[:, [5, 12]]
+
+# Split the data into training/testing sets
+n_train_data = int(0.8 * X.shape[0])
+
+X_train = X[:n_train_data]
+X_test = X[n_train_data:]
+
+y_train = label[:n_train_data]
+y_test = label[n_train_data:]
+
+# Create linear regression and fit it on the training set
+regr = linear_model.LinearRegression()
+regr.fit(X_train, y_train)
+
+# Make predictions using the testing set
+y_pred = regr.predict(X_test)
+
+print('Coefficients:')
+print(' intercept: {:.2f}'.format(regr.intercept))
+print(' average room per dwelling: {:.2f}'.format(regr.weights[0]))
+print(' percentage of low status in population: {:.2f}'
+      .format(regr.weights[1]))
+
+# The mean squared error
+print('Mean squared error on test set: {:.2f}'.format(
+    mean_squared_error(y_test, y_pred)))
+
+# Explained variance score: 1 is perfect prediction
+print('Variance score on test set: {:.2f}'.format(r2_score(y_test, y_pred)))
+# To work in 3D
+
+# We first generate a mesh grid
+resolution = 10
+x = X_test[:, 0]
+y = X_test[:, 1]
+z = y_test
+
+x_surf = np.linspace(min(x), max(x), resolution)
+y_surf = np.linspace(min(y), max(y), resolution)
+x_surf, y_surf = np.meshgrid(x_surf, y_surf)
+
+# and then predict the label for all values in the grid
+z_surf = np.zeros_like(x_surf)
+mesh_points = np.vstack((x_surf.ravel(), y_surf.ravel())).T
+z_surf.ravel()[:] = regr.predict(mesh_points)
+
+fig = plt.figure(figsize=(20, 5))
+
+# 3D representation under different rotated angles for a better visualazion
+xy_angles = [10, 35, 60, 85]
+z_angle = 20
+
+for i, angle in enumerate(xy_angles):
+    n_columns = len(xy_angles)
+    position = i + 1
+
+    ax = fig.add_subplot(1, n_columns, position, projection='3d')
+
+    ax.view_init(z_angle, angle)
+
+    ax.plot_surface(x_surf, y_surf, z_surf, cmap=cm.hot, rstride=1, cstride=1,
+                    alpha=0.3, linewidth=0.2, edgecolors='black')
+    ax.scatter(x, y, z)
+
+    ax.set_title('angle: {}°'.format(angle))
+    ax.set_zlabel('median house pricing')
+    ax.set_xlabel('avg room per dwelling')
+    ax.set_ylabel('% low status population')
+
+plt.show()