Allow using math.isclose for Python 3.5+ #110

Activity

1. seperman commented on Mar 20, 2019

   

   seperman

   on Mar 20, 2019

   Member

   More actions

   @stevenjackson121 This is a very cool idea. I have been thinking about how to use the math.isclose. The main problem is how to do that during the "hashing" phase where numbers are trimmed to their significant digits before getting hashed and at that point we don't know which number value in the future during the diffing part we will be comparing to. The hashing is only used when "ignore_order=True". So if we want to use math.isclose, we will hit some unexpected results when ignore_order=True.
2. seperman commented on Jun 23, 2020

   

   seperman

   on Jun 23, 2020

   Member

   More actions

   Hi @stevenjackson121
   Have you checked the combination of significant_digits and scientific format notation? https://zepworks.com/deepdiff/5.0.0/numbers.html#number-format-notation
3. martinResearch commented on Jan 20, 2022

   

   martinResearch

   on Jan 20, 2022

   Last edited by martinResearch

   More actions

   It looks like it is now using is_close with abs_tol whenever math_epsilon is not none https://github.com/seperman/deepdiff/blob/7e5e9954d635423d691f37194f3d1d93d363a128/deepdiff/diff.py#L1116

   However It would be great to add another parameter that would allow to use is_close with rel_tol instead of abs_tol
   As a workaround I am currently inheriting from DeepDiff and adding a parameter in the derived class:

   class DeepDiff(deepdiff.DeepDiff):
       """Class inherited from deepdiff.DeepDiff that adds the parameter rel_tol"""
   
       def __init__(self, t1,  t2, rel_tol=None, **kwargs):
           # adding rel_tol attribute before calling the baseclasse init as the diffeference gets 
           # computed within the constructor
           self.rel_tol = rel_tol
           super().__init__(t1,  t2, **kwargs)
   
       def _diff_numbers(self, level):
           if self.rel_tol is not None:
               if not is_close(level.t1, level.t2, rel_tol=self.rel_tol):
                   self._report_result("values_changed", level)
           else:
               super()._diff_numbers(level)
   

   It would be good to get the change integrated in the original class by adding this argument to the constructor.

   With this change DeepDiff(10050001,10049999, rel_tol=1e-6) gives no difference
   while DeepDiff(10050001,10049999, math_epsilon=0.1) and DeepDiff(1005001,1004999, significant_digits=2, number_format_notation="e") both detect a difference
4. martinResearch commented on Jan 21, 2022

   

   martinResearch

   on Jan 21, 2022

   More actions

   @seperman I can prepare a pull request to add an argument rel_tol to the DeepDiff constructor if that can help.
5. seperman commented on Jan 21, 2022

   

   seperman

   on Jan 21, 2022 via email

   Member

   More actions

   Hi Martin, That would be great! Thank you, Sep Dehpour

   …

   On Jan 21, 2022, at 8:04 AM, Martin de La Gorce ***@***.***> wrote:  @seperman I can prepare a pull request to add an argument rel_tol to the DeepDiff constructor if that can help. — Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android. You are receiving this because you were mentioned.
6. dpinol commented on Oct 22, 2026

   

   dpinol

   on Oct 22, 2026

   More actions

   @seperman please let me know what you think about #644