mayafox
commented
1 year ago possible fix
def multipleOf(validator, dB, instance, schema):
if not validator.is_type(instance, "number"):
return
if isinstance(dB, float):
precision = len(str(dB).split('.')[1])
quotient = instance / dB
try:
failed = int(round(quotient, precision)) != quotient
except OverflowError:
# When `instance` is large and `dB` is less than one,
# quotient can overflow to infinity; and then casting to int
# raises an error.
#
# In this case we fall back to Fraction logic, which is
# exact and cannot overflow. The performance is also
# acceptable: we try the fast all-float option first, and
# we know that fraction(dB) can have at most a few hundred
# digits in each part. The worst-case slowdown is therefore
# for already-slow enormous integers or Decimals.
failed = (Fraction(instance) / Fraction(dB)).denominator != 1
else:
failed = instance % dB
if failed:
yield ValidationError(f"{instance!r} is not a multiple of {dB}")
cwegener
Julian
danner26
jsenecal
Encountered the following error when fixing the schema validation errors for a new device I was adding into netbox/devicetype-library
Failed validating 'multipleOf' in schema['properties']['weight']['items']['properties']['value']: {'minimum': 0, 'multipleOf': 0.01, 'type': 'number'}
On instance['weight'][0]['value']: 10.11
Did a bit of poking around, changing the value to 1011 <<< integer, and not a float, caused the check to pass. This seems to potentially be an issue in jsonschema,
https://github.com/python-jsonschema/jsonschema/blob/main/jsonschema/_validators.py line 181
If I am reading the code right, jsonschema is running into a rounding error when running this validation.
db = 0.01 instance = 10.11 isinstance(db, float) True quotient = instance / db quotient 1010.9999999999999 int(quotient) 1010 int(round(quotient, 2)) 1011
Any Thoughts, additional information I should collect?