Lets say that we are getting POSTed a form like this in Django:

rate=10
items= [23,12,31,52,83,34]

The items are primary keys of an Item model. I have a bunch of business logic that will run and create more items based on this data, the results of some db lookups, and some business logic. I want to put that logic into a save signal or an overridden Model.save() method of another model (let's call it Inventory). The business logic will run when I create a new Inventory object using this form data. Inventory will look like this:

class Inventory(models.Model):
    picked_items = models.ManyToManyField(Item, related_name="items_picked_set")
    calculated_items = models.ManyToManyField(Item, related_name="items_calculated_set")
    rate = models.DecimalField()
    ... other fields here ... 

New calculated_items will be created based on the passed in items which will be stored as picked_items.

My question is this: is it better for the save() method on this model to accept:

• the request object (I don't really like this coupling)
• the form data as arguments or kwargs (a list of primary keys and the other form fields)
• a list of Items (The caller form or view will lookup the list of Items and create a list as well as pass in the other form fields)
• some other approach?

I know this is a bit subjective, but I was wondering what the general idea is. I've looked through a lot of code but I'm having a hard time finding a pattern I like.