Performing actions on a related model while saving - django

Question

I need help. For several days now I have been unable to figure out how to get around one problem. I would appreciate any idea or advice.
For simplicity, there is a set of models:
class A(models.Model):
name = models.CharField(max_length=255)
class B(models.Model):
name = models.CharField(max_length=255)
owner = models.ForeignKey(A, on_delete=models.CASCADE, null=False)
def foo():
do something
The number of instances of B belonging to A is not known in advance. The number of instances of B is determined by the user at the time of creating a new instance of A. Moreover, he can create an instance of A without creating a single instance of B (empty set).
Challenge: I need to run foo after saving a new instance of A. This function should handle both the fields of the new instance of A and fields of B (back relation).
Yes, I have set the receiver to post_save signal model A. But here's the problem, at the moment when A is already saved (post_save), instances of B have not yet been saved (pre_save), and the values ​​of the corresponding fields have not been determined. In other words, I cannot get raw B values ​​in receiver A.
Any ideas? Am I doing something wrong?

Answer 1

I found a solution. Maybe it will be useful to someone.
I took advice from our more experienced colleagues to use signals only as a last resort. And the solution turned out to be simple. Although the solution may not seem very elegant.
The solution is based on overriding the save() method of each model
class A(models.Model):
name = models.CharField(max_length=255)
def _call_foo(self, *args, **kwargs):
do something
foo(*args, **kwargs)
def save(self, *args, **kwargs):
changed_fields = []
pk = None
if self.pk:
cls = self.__class__
old = cls.objects.get(pk=self.pk)
new = self
for field in cls._meta.get_fields():
field_name = field.name
try:
if getattr(old, field_name) != getattr(new, field_name):
pk = self.pk
changed_fields.append(field_name)
# Catch field does not exist exception
except Exception as ex:
print(type(ex))
print(ex.arg)
print(ex)
kwargs['update_fields'] = changed_fields
super().save(*args, **kwargs)
else:
super().save(*args, **kwargs)
pk = self.pk
changed_fields.append('__all__')
self._call_foo(self, pk, changed_fields)
class B(models.Model):
name = models.CharField(max_length=255)
owner = models.ForeignKey(A, on_delete=models.CASCADE, null=False)
def save(self, *args, **kwargs):
changed_fields = []
pk = None
if self.pk:
cls = self.__class__
old = cls.objects.get(pk=self.pk)
new = self
for field in cls._meta.get_fields():
field_name = field.name
try:
if getattr(old, field_name) != getattr(new, field_name):
pk = self.pk
changed_fields.append(field_name)
# Catch field does not exist exception
except Exception as ex:
print(type(ex))
print(ex.arg)
print(ex)
kwargs['update_fields'] = changed_fields
super().save(*args, **kwargs)
else:
super().save(*args, **kwargs)
pk = self.pk
changed_fields.append('__all__')
self.owner._call_foo(self, pk, changed_fields)
This solution is better than using signals, at least in that the save method is called if there is a real need for it: either an instance is created, or it is updated.
Of course, need to keep in mind that if the user updates or create several instances of both models (for example, in the inline form, the admin creates several instances of B and one instance of A), then the foo() function will be called several times. But this can already be worked out either at the _call_foo() method level, or by the logic of the foo() function itself.