[damskii9992]

Fixes the unique_name and display_name serialization issue with nested NewBase classes

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.24%. Comparing base (3025900) to head (cbe1e3f).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #325      +/-   ##
===========================================
+ Coverage    83.20%   83.24%   +0.04%     
===========================================
  Files           63       63              
  Lines         5322     5324       +2     
===========================================
+ Hits          4428     4432       +4     
+ Misses         894      892       -2     

Flag	Coverage Δ
unittests	83.24% <100.00%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/easyscience/base_classes/easy_list.py	96.91% <100.00%> (+0.64%)	⬆️
src/easyscience/base_classes/new_base.py	100.00% <100.00%> (ø)
src/easyscience/io/serializer_base.py	90.72% <100.00%> (+0.51%)	⬆️

[rozyczko]

Nice. This also fixes the issue of having the caller's skip list mutated and another issue of __copy__ not keeping nested children's display names.

HOWEVER...
If any model uses EasyList as an attribute, serializing will now raise TypeError, since EasyList.to_dict has no skip.

Check this out. Sample has layers which is an EasyList.

from easyscience.base_classes import EasyList, ModelBase, NewBase

class Sample(ModelBase):
    def __init__(self,
                 layers=None,
                 unique_name=None,
                 display_name=None):
        super().__init__(unique_name=unique_name, display_name=display_name)
        self.layers = layers

base = NewBase(unique_name='base')
layer=EasyList([base])
sample = Sample(layers=layer, unique_name='sample')
print(sample.to_dict())

in develop, this suffers from mangled unique_name serialization (unique_name base -> EasyList_0):

{'@module': '__main__', '@class': 'Sample', '@version': None, 'layers': {'@module': 'easyscience.base_classes.easy_list', '@class': 'EasyList', '@version': '2.5.1+dev4', 'unique_name': 'EasyList_0'}, 'unique_name': 'sample'}

in your branch, the operation fails

TypeError: EasyList.to_dict() got an unexpected keyword argument 'skip'

my 292-remove-objbase branch gives the correct string, with both unique_names correct:

{'@module': '__main__', '@class': 'Sample', '@version': None, 'layers': {'@module': 'easyscience.base_classes.easy_list', '@class': 'EasyList', '@version': '2.5.1+dev4', 'data': [{'@module': 'easyscience.base_classes.new_base', '@class': 'NewBase', '@version': '2.5.1+dev4', 'unique_name': 'base'}]}, 'unique_name': 'sample'}

I think the fix in your branch would be extremely simple:
just add skip to to_dict in EasyList.

-    def to_dict(self) -> dict:
+    def to_dict(self, skip) -> dict:
         """
         Convert the EasyList to a dictionary for serialization.
 
         Returns
         -------
         dict
             Dictionary representation of the EasyList.
         """
-        dict_repr = super().to_dict()
+        dict_repr = super().to_dict(skip=skip)
         if self._protected_types != [NewBase]:
             dict_repr['protected_types'] = [
                 {'@module': cls_.__module__, '@class': cls_.__name__}
                 for cls_ in self._protected_types
             ]  # noqa: E501
-        dict_repr['data'] = [item.to_dict() for item in self._data]
+        dict_repr['data'] = [item.to_dict(skip=skip) for item in self._data]
         return dict_repr

Adding a test for EasyList as an attribute would also be helpful here.

[rozyczko]

serializer already acts on skip, so these checks are redundant, I think

[rozyczko]

The issue seems to have been fixed with the current commit 🔥

[rozyczko]

It looks good now.

Consider removing extra lines as per my comment and adding a unit test for EasyList as an attribute