ENH Compile & install on Python 3

Author: luispedro

.gitignore

@@ -9,3 +9,4 @@ build
 dist/
 milk.egg-info
 docs/milk/
+*.cpython*.so

milk/__init__.py

@@ -1,5 +1,5 @@
 # -*- coding: utf-8 -*-
-# Copyright (C) 2008-2013, Luis Pedro Coelho <[email protected]>
+# Copyright (C) 2008-2015, Luis Pedro Coelho <[email protected]>
 # vim: set ts=4 sts=4 sw=4 expandtab smartindent:
 # 
 # Permission is hereby granted, free of charge, to any person obtaining a copy
@@ -59,10 +59,10 @@
     from .supervised.defaultlearner import defaultlearner
     from .unsupervised.kmeans import kmeans
     from .unsupervised import pdist, zscore, pca
-    from milk_version import __version__
-except ImportError, e:
+    from .milk_version import __version__
+except ImportError as e:
     import sys
-    print >>sys.stderr, '''\
+    print('''\
 Could not import submodules (exact error was: %s).
 
 There are many reasons for this error the most common one is that you have
@@ -71,7 +71,7 @@
 milk **without changing the current directory**.
 
 Try installing and then changing to another directory before importing milk.
-''' % e
+''' % e, file=sys.stderr)
 
 __all__ = [
     '__version__',

milk/active/eimpact.py

@@ -44,7 +44,7 @@ def expected_impacts(D,labels,U):
     u_labels = (u_probs > .5)
     impacts = []
     for u,p in zip(U,u_probs):
-        print len(impacts)
+        print(len(impacts))
         label_classifier.train(numpy.vstack((D,u)),numpy.hstack((labels,[0])))
         u_labels_0 = label_classifier(U)
 

milk/demos/rf_wine_2d.py

@@ -16,7 +16,7 @@
 # cross validate with this learner and return predictions on left-out elements
 cmat,names, preds = milk.nfoldcrossvalidation(features, labels, classifier=learner, return_predictions=1)
 
-print 'cross-validation accuracy:', cmat.trace()/float(cmat.sum())
+print('cross-validation accuracy:', cmat.trace()/float(cmat.sum()))
 
 # dimensionality reduction for display
 x,v = milk.unsupervised.pca(features)

milk/ext/jugparallel.py

@@ -64,10 +64,10 @@ def nfoldcrossvalidation(features, labels, **kwargs):
     jug.CompoundTask : This function can be used as argument to CompoundTask
     '''
     nfolds = kwargs.get('nfolds', 10)
-    features,labels = map(identity, (features,labels))
-    kwargs = dict( (k,identity(v)) for k,v in kwargs.iteritems())
+    features,labels = list(map(identity, (features,labels)))
+    kwargs = dict( (k,identity(v)) for k,v in kwargs.items())
     nfold_one = TaskGenerator(milk.nfoldcrossvalidation)
-    mapped = [nfold_one(features, labels, folds=[i], **kwargs) for i in xrange(nfolds)]
+    mapped = [nfold_one(features, labels, folds=[i], **kwargs) for i in range(nfolds)]
     return jug_reduce(_nfold_reduce, mapped)
 
 
@@ -137,7 +137,7 @@ def kmeans_select_best(features, ks, repeats=1, method='AIC', R=None, **kwargs):
         start = 7
     results = []
     for ki,k in enumerate(ks):
-        for i in xrange(repeats):
+        for i in range(repeats):
             results.append(kmeans(features, k, R=(start+7*repeats*ki+i), **kwargs))
     return _select_best(features, results, method)[1]
 

milk/measures/__init__.py

@@ -1 +1 @@
-from measures import accuracy, waccuracy, zero_one_loss, confusion_matrix, bayesian_significance
+from .measures import accuracy, waccuracy, zero_one_loss, confusion_matrix, bayesian_significance

milk/measures/measures.py

@@ -119,7 +119,7 @@ def confusion_matrix(real, predicted, normalisedlabels=False, names=None):
     if not normalisedlabels:
         from ..supervised.normalise import normaliselabels
         real, names = normaliselabels(real)
-        predicted = map(names.index, predicted)
+        predicted = list(map(names.index, predicted))
     n = np.max(real)+1
     cmat = np.zeros((n,n))
     for r,p in zip(real, predicted):

milk/measures/nfoldcrossvalidation.py

@@ -6,6 +6,7 @@
 from __future__ import division
 from ..supervised.classifier import normaliselabels
 import numpy as np
+from functools import reduce
 
 __all__ = ['foldgenerator', 'getfold', 'nfoldcrossvalidation']
 def foldgenerator(labels, nfolds=None, origins=None, folds=None, multi_label=False):
@@ -45,7 +46,7 @@ def foldgenerator(labels, nfolds=None, origins=None, folds=None, multi_label=Fal
              'milk.nfoldcrossvalidation.foldgenerator: origins must be of same size as labels')
         origins = np.asanyarray(origins)
     fmin = len(labels)
-    for ell in xrange(len(names)):
+    for ell in range(len(names)):
         if multi_label:
             matching = (orig for i,orig in enumerate(origins) if labels[i,ell])
         else:
@@ -92,7 +93,7 @@ def foldgenerator(labels, nfolds=None, origins=None, folds=None, multi_label=Fal
                 fold[origins == orig] = f
                 foldweight[f] += w
 
-    for f in xrange(nfolds):
+    for f in range(nfolds):
         if folds is not None and f not in folds: continue
         yield (fold != f), (fold == f)
 

milk/supervised/_lasso.cpp

@@ -8,6 +8,7 @@
 #include <cassert>
 #include <queue>
 #include "eigen3/Eigen/Dense"
+#include "../utils/utils.h"
 
 using namespace Eigen;
 
@@ -198,10 +199,5 @@ const char  * module_doc =
 
 } // namespace
 
-extern "C"
-void init_lasso()
-  {
-    import_array();
-    (void)Py_InitModule3("_lasso", methods, module_doc);
-  }
+DECLARE_MODULE(_lasso)
 

milk/supervised/_perceptron.cpp

@@ -1,10 +1,11 @@
-// Copyright (C) 2011, Luis Pedro Coelho <[email protected]>
+// Copyright (C) 2011-2015, Luis Pedro Coelho <[email protected]>
 // License: MIT
 
 #include <iostream>
 #include <memory>
 #include <cmath>
 #include <cassert>
+#include "../utils/utils.h"
 extern "C" {
     #include <Python.h>
     #include <numpy/ndarrayobject.h>
@@ -82,10 +83,5 @@ const char  * module_doc =
 
 } // namespace
 
-extern "C"
-void init_perceptron()
-  {
-    import_array();
-    (void)Py_InitModule3("_perceptron", methods, module_doc);
-  }
+DECLARE_MODULE(_perceptron)