Skip to content

Commit 1fe13ed

Browse files
rtavenarrtavenar
committed
Fixed doctests
1 parent b05d315 commit 1fe13ed

File tree

4 files changed

+35
-49
lines changed

4 files changed

+35
-49
lines changed

‎.travis.yml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -32,5 +32,5 @@ install:
3232
script:
3333
- python setup.py develop
3434
- flake8 examples/ ot/ test/
35-
- python -m pytest -v test/ ot/ --doctest-modules --cov=ot
35+
- python -m pytest -v test/ ot/ --doctest-modules --ignore ot/gpu/ --cov=ot
3636
# - py.test ot test

‎ot/bregman.py‎

Lines changed: 7 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -1360,10 +1360,9 @@ def empirical_sinkhorn(X_s, X_t, reg, a=None, b=None, metric='sqeuclidean', numI
13601360
>>> reg = 0.1
13611361
>>> X_s = np.reshape(np.arange(n_s), (n_s, 1))
13621362
>>> X_t = np.reshape(np.arange(0, n_t), (n_t, 1))
1363-
>>> emp_sinkhorn = empirical_sinkhorn(X_s, X_t, reg, verbose=False)
1364-
>>> print(emp_sinkhorn)
1365-
>>> [[4.99977301e-01 2.26989344e-05]
1366-
[2.26989344e-05 4.99977301e-01]]
1363+
>>> empirical_sinkhorn(X_s, X_t, reg, verbose=False) # doctest: +NORMALIZE_WHITESPACE
1364+
array([[4.99977301e-01, 2.26989344e-05],
1365+
[2.26989344e-05, 4.99977301e-01]])
13671366
13681367
13691368
References
@@ -1451,9 +1450,8 @@ def empirical_sinkhorn2(X_s, X_t, reg, a=None, b=None, metric='sqeuclidean', num
14511450
>>> reg = 0.1
14521451
>>> X_s = np.reshape(np.arange(n_s), (n_s, 1))
14531452
>>> X_t = np.reshape(np.arange(0, n_t), (n_t, 1))
1454-
>>> loss_sinkhorn = empirical_sinkhorn2(X_s, X_t, reg, verbose=False)
1455-
>>> print(loss_sinkhorn)
1456-
>>> [4.53978687e-05]
1453+
>>> empirical_sinkhorn2(X_s, X_t, reg, verbose=False)
1454+
array([4.53978687e-05])
14571455
14581456
14591457
References
@@ -1560,9 +1558,8 @@ def empirical_sinkhorn_divergence(X_s, X_t, reg, a=None, b=None, metric='sqeucli
15601558
>>> reg = 0.1
15611559
>>> X_s = np.reshape(np.arange(n_s), (n_s, 1))
15621560
>>> X_t = np.reshape(np.arange(0, n_t), (n_t, 1))
1563-
>>> emp_sinkhorn_div = empirical_sinkhorn_divergence(X_s, X_t, reg)
1564-
>>> print(emp_sinkhorn_div)
1565-
>>> [2.99977435]
1561+
>>> empirical_sinkhorn_divergence(X_s, X_t, reg)
1562+
array([2.99977435])
15661563
15671564
15681565
References

‎ot/stochastic.py‎

Lines changed: 24 additions & 35 deletions
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111

1212

1313
def coordinate_grad_semi_dual(b, M, reg, beta, i):
14-
'''
14+
r'''
1515
Compute the coordinate gradient update for regularized discrete distributions for (i, :)
1616
1717
The function computes the gradient of the semi dual problem:
@@ -51,7 +51,7 @@ def coordinate_grad_semi_dual(b, M, reg, beta, i):
5151
5252
Examples
5353
--------
54-
54+
>>> import ot
5555
>>> n_source = 7
5656
>>> n_target = 4
5757
>>> reg = 1
@@ -63,8 +63,7 @@ def coordinate_grad_semi_dual(b, M, reg, beta, i):
6363
>>> Y_target = rng.randn(n_target, 2)
6464
>>> M = ot.dist(X_source, Y_target)
6565
>>> method = "ASGD"
66-
>>> asgd_pi = stochastic.solve_semi_dual_entropic(a, b, M, reg,
67-
method, numItermax)
66+
>>> asgd_pi = ot.stochastic.solve_semi_dual_entropic(a, b, M, reg, method, numItermax)
6867
>>> print(asgd_pi)
6968
7069
References
@@ -84,7 +83,7 @@ def coordinate_grad_semi_dual(b, M, reg, beta, i):
8483

8584

8685
def sag_entropic_transport(a, b, M, reg, numItermax=10000, lr=None):
87-
'''
86+
r'''
8887
Compute the SAG algorithm to solve the regularized discrete measures
8988
optimal transport max problem
9089
@@ -133,7 +132,7 @@ def sag_entropic_transport(a, b, M, reg, numItermax=10000, lr=None):
133132
134133
Examples
135134
--------
136-
135+
>>> import ot
137136
>>> n_source = 7
138137
>>> n_target = 4
139138
>>> reg = 1
@@ -145,8 +144,7 @@ def sag_entropic_transport(a, b, M, reg, numItermax=10000, lr=None):
145144
>>> Y_target = rng.randn(n_target, 2)
146145
>>> M = ot.dist(X_source, Y_target)
147146
>>> method = "ASGD"
148-
>>> asgd_pi = stochastic.solve_semi_dual_entropic(a, b, M, reg,
149-
method, numItermax)
147+
>>> asgd_pi = ot.stochastic.solve_semi_dual_entropic(a, b, M, reg, method, numItermax)
150148
>>> print(asgd_pi)
151149
152150
References
@@ -176,7 +174,7 @@ def sag_entropic_transport(a, b, M, reg, numItermax=10000, lr=None):
176174

177175

178176
def averaged_sgd_entropic_transport(a, b, M, reg, numItermax=300000, lr=None):
179-
'''
177+
r'''
180178
Compute the ASGD algorithm to solve the regularized semi continous measures optimal transport max problem
181179
182180
The function solves the following optimization problem:
@@ -223,7 +221,7 @@ def averaged_sgd_entropic_transport(a, b, M, reg, numItermax=300000, lr=None):
223221
224222
Examples
225223
--------
226-
224+
>>> import ot
227225
>>> n_source = 7
228226
>>> n_target = 4
229227
>>> reg = 1
@@ -235,8 +233,7 @@ def averaged_sgd_entropic_transport(a, b, M, reg, numItermax=300000, lr=None):
235233
>>> Y_target = rng.randn(n_target, 2)
236234
>>> M = ot.dist(X_source, Y_target)
237235
>>> method = "ASGD"
238-
>>> asgd_pi = stochastic.solve_semi_dual_entropic(a, b, M, reg,
239-
method, numItermax)
236+
>>> asgd_pi = ot.stochastic.solve_semi_dual_entropic(a, b, M, reg, method, numItermax)
240237
>>> print(asgd_pi)
241238
242239
References
@@ -264,7 +261,7 @@ def averaged_sgd_entropic_transport(a, b, M, reg, numItermax=300000, lr=None):
264261

265262

266263
def c_transform_entropic(b, M, reg, beta):
267-
'''
264+
r'''
268265
The goal is to recover u from the c-transform.
269266
270267
The function computes the c_transform of a dual variable from the other
@@ -303,7 +300,7 @@ def c_transform_entropic(b, M, reg, beta):
303300
304301
Examples
305302
--------
306-
303+
>>> import ot
307304
>>> n_source = 7
308305
>>> n_target = 4
309306
>>> reg = 1
@@ -315,8 +312,7 @@ def c_transform_entropic(b, M, reg, beta):
315312
>>> Y_target = rng.randn(n_target, 2)
316313
>>> M = ot.dist(X_source, Y_target)
317314
>>> method = "ASGD"
318-
>>> asgd_pi = stochastic.solve_semi_dual_entropic(a, b, M, reg,
319-
method, numItermax)
315+
>>> asgd_pi = ot.stochastic.solve_semi_dual_entropic(a, b, M, reg, method, numItermax)
320316
>>> print(asgd_pi)
321317
322318
References
@@ -340,7 +336,7 @@ def c_transform_entropic(b, M, reg, beta):
340336

341337
def solve_semi_dual_entropic(a, b, M, reg, method, numItermax=10000, lr=None,
342338
log=False):
343-
'''
339+
r'''
344340
Compute the transportation matrix to solve the regularized discrete
345341
measures optimal transport max problem
346342
@@ -398,7 +394,7 @@ def solve_semi_dual_entropic(a, b, M, reg, method, numItermax=10000, lr=None,
398394
399395
Examples
400396
--------
401-
397+
>>> import ot
402398
>>> n_source = 7
403399
>>> n_target = 4
404400
>>> reg = 1
@@ -410,8 +406,7 @@ def solve_semi_dual_entropic(a, b, M, reg, method, numItermax=10000, lr=None,
410406
>>> Y_target = rng.randn(n_target, 2)
411407
>>> M = ot.dist(X_source, Y_target)
412408
>>> method = "ASGD"
413-
>>> asgd_pi = stochastic.solve_semi_dual_entropic(a, b, M, reg,
414-
method, numItermax)
409+
>>> asgd_pi = ot.stochastic.solve_semi_dual_entropic(a, b, M, reg, method, numItermax)
415410
>>> print(asgd_pi)
416411
417412
References
@@ -451,7 +446,7 @@ def solve_semi_dual_entropic(a, b, M, reg, method, numItermax=10000, lr=None,
451446

452447
def batch_grad_dual(a, b, M, reg, alpha, beta, batch_size, batch_alpha,
453448
batch_beta):
454-
'''
449+
r'''
455450
Computes the partial gradient of the dual optimal transport problem.
456451
457452
For each (i,j) in a batch of coordinates, the partial gradients are :
@@ -506,7 +501,7 @@ def batch_grad_dual(a, b, M, reg, alpha, beta, batch_size, batch_alpha,
506501
507502
Examples
508503
--------
509-
504+
>>> import ot
510505
>>> n_source = 7
511506
>>> n_target = 4
512507
>>> reg = 1
@@ -520,9 +515,7 @@ def batch_grad_dual(a, b, M, reg, alpha, beta, batch_size, batch_alpha,
520515
>>> X_source = rng.randn(n_source, 2)
521516
>>> Y_target = rng.randn(n_target, 2)
522517
>>> M = ot.dist(X_source, Y_target)
523-
>>> sgd_dual_pi, log = stochastic.solve_dual_entropic(a, b, M, reg,
524-
batch_size,
525-
numItermax, lr, log)
518+
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg, batch_size, numItermax, lr, log)
526519
>>> print(log['alpha'], log['beta'])
527520
>>> print(sgd_dual_pi)
528521
@@ -548,7 +541,7 @@ def batch_grad_dual(a, b, M, reg, alpha, beta, batch_size, batch_alpha,
548541

549542

550543
def sgd_entropic_regularization(a, b, M, reg, batch_size, numItermax, lr):
551-
'''
544+
r'''
552545
Compute the sgd algorithm to solve the regularized discrete measures
553546
optimal transport dual problem
554547
@@ -597,7 +590,7 @@ def sgd_entropic_regularization(a, b, M, reg, batch_size, numItermax, lr):
597590
598591
Examples
599592
--------
600-
593+
>>> import ot
601594
>>> n_source = 7
602595
>>> n_target = 4
603596
>>> reg = 1
@@ -611,9 +604,7 @@ def sgd_entropic_regularization(a, b, M, reg, batch_size, numItermax, lr):
611604
>>> X_source = rng.randn(n_source, 2)
612605
>>> Y_target = rng.randn(n_target, 2)
613606
>>> M = ot.dist(X_source, Y_target)
614-
>>> sgd_dual_pi, log = stochastic.solve_dual_entropic(a, b, M, reg,
615-
batch_size,
616-
numItermax, lr, log)
607+
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg, batch_size, numItermax, lr, log)
617608
>>> print(log['alpha'], log['beta'])
618609
>>> print(sgd_dual_pi)
619610
@@ -644,7 +635,7 @@ def sgd_entropic_regularization(a, b, M, reg, batch_size, numItermax, lr):
644635

645636
def solve_dual_entropic(a, b, M, reg, batch_size, numItermax=10000, lr=1,
646637
log=False):
647-
'''
638+
r'''
648639
Compute the transportation matrix to solve the regularized discrete measures
649640
optimal transport dual problem
650641
@@ -695,7 +686,7 @@ def solve_dual_entropic(a, b, M, reg, batch_size, numItermax=10000, lr=1,
695686
696687
Examples
697688
--------
698-
689+
>>> import ot
699690
>>> n_source = 7
700691
>>> n_target = 4
701692
>>> reg = 1
@@ -709,9 +700,7 @@ def solve_dual_entropic(a, b, M, reg, batch_size, numItermax=10000, lr=1,
709700
>>> X_source = rng.randn(n_source, 2)
710701
>>> Y_target = rng.randn(n_target, 2)
711702
>>> M = ot.dist(X_source, Y_target)
712-
>>> sgd_dual_pi, log = stochastic.solve_dual_entropic(a, b, M, reg,
713-
batch_size,
714-
numItermax, lr, log)
703+
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg, batch_size, numItermax, lr, log)
715704
>>> print(log['alpha'], log['beta'])
716705
>>> print(sgd_dual_pi)
717706

‎ot/utils.py‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -285,9 +285,9 @@ class deprecated(object):
285285
The optional extra argument will be appended to the deprecation message
286286
and the docstring. Note: to use this with the default value for extra, put
287287
in an empty of parentheses:
288-
>>> from ot.deprecation import deprecated
289-
>>> @deprecated()
290-
... def some_function(): pass
288+
>>> from ot.deprecation import deprecated # doctest: +SKIP
289+
>>> @deprecated() # doctest: +SKIP
290+
... def some_function(): pass # doctest: +SKIP
291291
292292
Parameters
293293
----------

0 commit comments

Comments
 (0)