Django元数据数据库的最佳实践是什么？

Question

I have to develop a webservice for storing data and I want to do it on Python/Django. It has only 2 elements at frontend - a table and a diagram. So basically the output in template is totally standard for all the views in all apps. I have 2 classes that compile data for output for these 2 elements in JSON, that goes to ReactJS at frontend.

Now, I have several apps and the only difference between them are models. Even CRUD-views and URLS are almost the same for all them. So i'm thinking of using some metadatabase to store all data in a unified way. Like storing all values in 3 tables - for strings, int and float.

The thing that i'm afraid of is performance. If all my metadata like types of objects, categories and so on will be stored in a single table, I guess it will be hell to get even a simple value out of 4 or 5 requests using Django ORM and a common model for all things in the system.

On the other hand, there's an option that I can't create new models programmatically on the fly. Once it's in production - there should be no changes in the code, because of the checksum of the project files.

I need an advice how to handle this situation. I thought of something like Drupal metadatabase, but Drupal creates new tables on the fly, and it doesn't need models in the code to work with them. Is there a way to do it like that in Django?

Answer 1

Judging from your requirements you would probably need to use Generic relations: https://docs.djangoproject.com/en/1.8/ref/contrib/contenttypes/ but it wont be the end of the world if all metadata are saved in a single table. Lets say you have a 3 different models strings, ints, floats. You need to create a MetaData Model that acts as a key value store:

class MetaData(models.Model):
    key = models.CharField(max_length=256, db_index=True)
    value = models.CharField(max_length=256)

    # Generic relationship to an arbitrary object
    content_type = models.ForeignKey(ContentType)
    object_id = models.PositiveIntegerField()
    content_object = generic.GenericForeignKey('content_type', 'object_id')

Then in a simple case you can use save values to this table like that: f=Float.objects.get(pk=1) v = MetaData(content_object=f, key="float_calculation_precision", value="6") v.save()

If you know before hand that you are going to be using only those 3 models you can add a reverse GenericRelation field on each one of those models.

class Floats(models.Model):
    metadata = generic.GenericRelation(MetaData)

class Ints(models.Model):
    metadata = generic.GenericRelation(MetaData)

class Strings(models.Model):
    metadata = generic.GenericRelation(MetaData)

and then you can access their metadata like that:

Floats.metadata.all()

to get all metadata objects or:

f = Floats.objects.get(pk=1)

f.metadata.all()

to get all metadata objects for a specific object:

You can also use https://github.com/rafaelsierra/django-metadata package where you can do something like that:

mymodel.metadata['something'] = 'some value'
mymodel.metadata['something']