Skip to content

Threshold must be a real number #322

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 11 commits into from
Nov 2, 2021
Merged

Threshold must be a real number #322

Show file tree
Hide file tree
Changes from 8 commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
8fc975c
Add venv to gitignore
Sep 13, 2021
40bbce8
Check if threshold is a real value
Sep 13, 2021
60dc0d4
Simplified threshold type-check
mvargas33 Sep 17, 2021
a92d2c9
Follow linter rules
mvargas33 Sep 17, 2021
ddce7db
Fix last linter error
Sep 21, 2021
ac4584e
Add test to check correct behaviour. Sacrified simplicity for the boo…
Sep 21, 2021
0d80fdf
Merge remote-tracking branch 'upstream/master' into feat-threshold
Oct 28, 2021
166228f
Update test. Stick to custom message. It's bool permissive
Oct 28, 2021
47c21ac
Explicit boolean permissive case in test
Oct 28, 2021
81c7f73
Changed isinstance for custom ValueError message
Oct 28, 2021
796ba93
TypeError for most input. ValueError for String case.
Nov 2, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions metric_learn/base_metric.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -569,8 +569,10 @@ def set_threshold(self, threshold):
The pairs classifier with the new threshold set.
"""
check_is_fitted(self, 'preprocessor_')

self.threshold_ = threshold
if not isinstance(threshold, (int, float)):
raise ValueError('Parameter threshold must be a real number. '
'Got {} instead.'.format(type(threshold)))
self.threshold_ = float(threshold) # int -> float
Copy link
Member

@bellet bellet Oct 28, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sorry to nitpick but can't you just call self.threshold_ = float(threshold) at the beginning and throw the custom warning if a ValueError is raised ? This would avoid calling isinstance; and would just let bool go through (converting to 0/1), which is fine.

Copy link
Contributor Author

@mvargas33 mvargas33 Oct 28, 2021 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I changed it for this

try:
  self.threshold_ = float(threshold)
except Exception:
  raise ValueError('Parameter threshold must be a real number. '
                   'Got {} instead.'.format(type(threshold)))

I guess if a warning is shown instead, the code can fail somewhere else, because it won't stop the execution.

Copy link
Contributor Author

@mvargas33 mvargas33 Oct 28, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And forgot to mention that isinstance(threshold, (int, float)) is permissive for bool. float(threshold) is also permissive.

return self

def calibrate_threshold(self, pairs_valid, y_valid, strategy='accuracy',
Expand Down
25 changes: 25 additions & 0 deletions test/test_pairs_classifiers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -180,6 +180,31 @@ def test_set_threshold():
assert identity_pairs_classifier.threshold_ == 0.5


@pytest.mark.parametrize('value', ["ABC", None, [1, 2, 3], {'key': None},
(1, 2), set(),
np.array([[[0.], [1.]], [[1.], [3.]]]),
np.array([0.5]), np.array([[[0.5]]])])
def test_set_wrong_type_threshold(value):
"""
Test that `set_threshold` indeed sets the threshold
and cannot accept nothing but float or integers, but
being permissive with boolean True=1.0 and False=0.0
"""
model = IdentityPairsClassifier()
model.fit(np.array([[[0.], [1.]]]), np.array([1]))
msg = ('Parameter threshold must be a real number. '
'Got {} instead.'.format(type(value)))

with pytest.raises(ValueError) as e: # String
model.set_threshold(value)
assert str(e.value).startswith(msg)

model.set_threshold(1) # Integer
assert model.threshold_ == 1.0
model.set_threshold(0.1) # Float
assert model.threshold_ == 0.1


def test_f_beta_1_is_f_1():
# test that putting beta to 1 indeed finds the best threshold to optimize
# the f1_score
Expand Down