Skip to content
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

Jump to bottom

Improve precision of float multiple_of validation #1373

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
+77 −23
Open

Improve precision of float multiple_of validation #1373

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3432e82
Improve precision of float multiple_of validation
K-dash Jul 22, 2024
8e9b338
Add equivalent decimal tests for consistency
K-dash Jul 22, 2024
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
20 changes: 17 additions & 3 deletions src/validators/float.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -108,9 +108,23 @@ impl Validator for ConstrainedFloatValidator {
return Err(ValError::new(ErrorTypeDefaults::FiniteNumber, input));
}
if let Some(multiple_of) = self.multiple_of {
let rem = float % multiple_of;
let threshold = float.abs() / 1e9;
if rem.abs() > threshold && (rem - multiple_of).abs() > threshold {
const EPSILON_FACTOR: f64 = 100.0;
let epsilon_threshold = f64::EPSILON * EPSILON_FACTOR;

// Round the result of dividing the input value by the multiple
let rounded = (float / multiple_of).round();

// Calculate the difference between the rounded value and the original value
let diff = (float - rounded * multiple_of).abs();
Comment on lines +117 to +118
Copy link
Contributor

Choose a reason for hiding this comment

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

It feels to me like given we evaluate threshold in units of multiple_of, we might want to do similarly here to end up with same scale of multiple_of (by dividing through by rounded)?

Suggested change
// Calculate the difference between the rounded value and the original value
let diff = (float - rounded * multiple_of).abs();
// Calculate how close the rounded division is to multiple_of
let diff = (float / rounded - multiple_of).abs();


// Calculate the relative error (avoid division by zero)
let relative_error = if float != 0.0 { diff / float.abs() } else { 0.0 };

// Threshold (considering both relative and absolute error)
let threshold = epsilon_threshold.max(multiple_of * epsilon_threshold);
Copy link
Contributor

Choose a reason for hiding this comment

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

This can be summarised as: if multiple_of > 1, then threshold will be increased by the same corresponding factor.


// Check if the difference exceeds the threshold and the relative error is significant
if diff > threshold && relative_error > epsilon_threshold {
Copy link
Contributor

Choose a reason for hiding this comment

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

So

  • diff > threshold basically is asking for the absolute error to be greater than epsilon (with a factor of 100), possibly scaled up by multiple_of.
  • relative_error > epsilon_threshold is basically asking if diff / float is greater than epsilon.

return Err(ValError::new(
ErrorType::MultipleOf {
multiple_of: multiple_of.into(),
Expand Down
40 changes: 30 additions & 10 deletions tests/validators/test_decimal.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,24 +171,44 @@ def test_decimal_kwargs(py_and_json: PyAndJson, kwargs: dict[str, Any], input_va
@pytest.mark.parametrize(
'multiple_of,input_value,error',
[
(0.5, 0.5, None),
(0.5, 1, None),
# Test cases for multiples of 0.5
*[(0.5, round(i * 0.5, 1), None) for i in range(-4, 5)],
(0.5, 0.49, Err('Input should be a multiple of 0.5')),
(0.5, 0.6, Err('Input should be a multiple of 0.5')),
(0.5, 0.51, Err('Input should be a multiple of 0.5')),
(0.5, -0.75, Err('Input should be a multiple of 0.5')),
(0.5, 0.501, Err('Input should be a multiple of 0.5')),
(0.5, 1_000_000.5, None),
(0.5, 1_000_000.49, Err('Input should be a multiple of 0.5')),
(0.5, int(5e10), None),
# Test cases for multiples of 0.1
*[(0.1, round(i * 0.1, 1), None) for i in range(-10, 11)],
(0.1, 0, None),
(0.1, 0.0, None),
(0.1, 0.2, None),
(0.1, 0.3, None),
(0.1, 0.4, None),
(0.1, 0.5, None),
sydney-runkle marked this conversation as resolved.
Show resolved Hide resolved
(0.1, 0.5001, Err('Input should be a multiple of 0.1')),
(0.1, 0.05, Err('Input should be a multiple of 0.1')),
(0.1, -0.15, Err('Input should be a multiple of 0.1')),
(0.1, 1_000_000.1, None),
(0.1, 1_000_000.05, Err('Input should be a multiple of 0.1')),
(0.1, 1, None),
(0.1, 1.0, None),
(0.1, int(5e10), None),
(2.0, -2.0, None),
# Test cases for multiples of 2.0
*[(2.0, i * 2.0, None) for i in range(-5, 6)],
(2.0, -2.1, Err('Input should be a multiple of 2')),
(2.0, -3.0, Err('Input should be a multiple of 2')),
(2.0, 1_000_002.0, None),
(2.0, 1_000_001.0, Err('Input should be a multiple of 2')),
(2.0, int(5e10), None),
# Test cases for multiples of 0.01
*[(0.01, round(i * 0.01, 2), None) for i in range(-10, 11)],
(0.01, 0.005, Err('Input should be a multiple of 0.01')),
(0.01, -0.015, Err('Input should be a multiple of 0.01')),
(0.01, 1_000_000.01, None),
(0.01, 1_000_000.005, Err('Input should be a multiple of 0.01')),
(0.01, int(5e10), None),
# Test cases for values very close to zero
(0.1, 0.00001, Err('Input should be a multiple of 0.1')),
(0.1, -0.00001, Err('Input should be a multiple of 0.1')),
(0.01, 0.00001, Err('Input should be a multiple of 0.01')),
(0.01, -0.00001, Err('Input should be a multiple of 0.01')),
],
ids=repr,
)
Expand Down
40 changes: 30 additions & 10 deletions tests/validators/test_float.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,24 +105,44 @@ def test_float_kwargs(py_and_json: PyAndJson, kwargs: Dict[str, Any], input_valu
@pytest.mark.parametrize(
'multiple_of,input_value,error',
[
(0.5, 0.5, None),
(0.5, 1, None),
# Test cases for multiples of 0.5
*[(0.5, round(i * 0.5, 1), None) for i in range(-4, 5)],
(0.5, 0.49, Err('Input should be a multiple of 0.5')),
(0.5, 0.6, Err('Input should be a multiple of 0.5')),
(0.5, 0.51, Err('Input should be a multiple of 0.5')),
(0.5, -0.75, Err('Input should be a multiple of 0.5')),
(0.5, 0.501, Err('Input should be a multiple of 0.5')),
(0.5, 1_000_000.5, None),
(0.5, 1_000_000.49, Err('Input should be a multiple of 0.5')),
(0.5, int(5e10), None),
# Test cases for multiples of 0.1
*[(0.1, round(i * 0.1, 1), None) for i in range(-10, 11)],
(0.1, 0, None),
(0.1, 0.0, None),
(0.1, 0.2, None),
(0.1, 0.3, None),
(0.1, 0.4, None),
(0.1, 0.5, None),
sydney-runkle marked this conversation as resolved.
Show resolved Hide resolved
(0.1, 0.5001, Err('Input should be a multiple of 0.1')),
(0.1, 0.05, Err('Input should be a multiple of 0.1')),
(0.1, -0.15, Err('Input should be a multiple of 0.1')),
(0.1, 1_000_000.1, None),
(0.1, 1_000_000.05, Err('Input should be a multiple of 0.1')),
(0.1, 1, None),
(0.1, 1.0, None),
(0.1, int(5e10), None),
(2.0, -2.0, None),
# Test cases for multiples of 2.0
*[(2.0, i * 2.0, None) for i in range(-5, 6)],
(2.0, -2.1, Err('Input should be a multiple of 2')),
(2.0, -3.0, Err('Input should be a multiple of 2')),
(2.0, 1_000_002.0, None),
(2.0, 1_000_001.0, Err('Input should be a multiple of 2')),
(2.0, int(5e10), None),
# Test cases for multiples of 0.01
*[(0.01, round(i * 0.01, 2), None) for i in range(-10, 11)],
(0.01, 0.005, Err('Input should be a multiple of 0.01')),
(0.01, -0.015, Err('Input should be a multiple of 0.01')),
(0.01, 1_000_000.01, None),
(0.01, 1_000_000.005, Err('Input should be a multiple of 0.01')),
(0.01, int(5e10), None),
# Test cases for values very close to zero
(0.1, 0.00001, Err('Input should be a multiple of 0.1')),
(0.1, -0.00001, Err('Input should be a multiple of 0.1')),
(0.01, 0.00001, Err('Input should be a multiple of 0.01')),
(0.01, -0.00001, Err('Input should be a multiple of 0.01')),
],
ids=repr,
)
Expand Down