Improved error message when attempting to label an unsaved Record such as anArtifact

[Zethson]

Add a description

A user ran into:

Traceback (most recent call last):
 File "/fs/gpfs41/lv03/fileset01/pool/pool-mann-maedler-shared/niklas_workspace/lamin_xenium/xenium_to_sdata.py", line 41, in <module>
  art.ulabels.add(technology_xenium)
 File "/fs/home/schmacke/miniforge3/envs/sdata/lib/python3.10/site-packages/django/db/models/fields/related_descriptors.py", line 656, in __get__
  return self.related_manager_cls(instance)
 File "/fs/home/schmacke/miniforge3/envs/sdata/lib/python3.10/site-packages/django/db/models/fields/related_descriptors.py", line 1076, in __init__
  raise ValueError(
ValueError: "Artifact(uid='6sYRmHuqlMUvllcS0000', is_latest=True, description='Ovarian Cancer Xenium Dataset', suffix='.sdata', kind='dataset', size=28462729153, hash='XKMmMiOv9dhZD_v8Q1yp1A', n_files=3403, space_id=1, storage_id=1, run_id=13, created_by_id=5, created_at=<django.db.models.expressions.DatabaseDefault object at 0x14f6481820b0>)" needs to have a value for field "id" before this many-to-many relationship can be used.

and they did not understand this error at all. It's cryptic. We need a better error message that suggest that the Record that is being labeled (an Artifact here but could be anything) needs to be saved first.

[Zethson]

from django.db.models.fields.related_descriptors import ManyToManyDescriptor
from django.db.models.query import QuerySet

class EnhancedM2MDescriptor(ManyToManyDescriptor):
    def __get__(self, instance, cls=None):
        if instance is not None and not instance.pk:
            raise ValueError(
                f"Cannot access {self.field.name} on unsaved {instance.__class__.__name__}. "
                f"Save {instance.__class__.__name__} first."
            )
        
        manager = super().__get__(instance, cls)
        
        original_bulk_create = manager.bulk_create
        def enhanced_bulk_create(objs, **kwargs):
            if isinstance(objs, (list, QuerySet)):
                unsaved = [obj for obj in objs if not obj.pk]
                if unsaved:
                    raise ValueError(
                        f"Cannot bulk create relationship - these objects need to be saved first: "
                        f"{', '.join(str(obj) for obj in unsaved[:3])}{'...' if len(unsaved) > 3 else ''}"
                    )
            return original_bulk_create(objs, **kwargs)
            
        manager.bulk_create = enhanced_bulk_create
        return manager

class EnhancedManyToManyField(models.ManyToManyField):
    def contribute_to_class(self, cls, name):
        super().contribute_to_class(cls, name)
        setattr(cls, name, EnhancedM2MDescriptor(self.remote_field, self.remote_field.get_related_field()))

A claude starting point for me to reason when I revisit this issue later. This is the complex version with bulk support. I need to look at it properly

[Zethson]

A user also recently tried things like:

reference = ln.Reference(bla)

artifact.references.add(reference)
artifact.save()

which errored because the Reference was not saved. We need user friendly errors here.