Skip to content

Latest commit

 

History

History
454 lines (285 loc) · 13.1 KB

File metadata and controls

454 lines (285 loc) · 13.1 KB

Models

.. rst-class:: emphasize-children

Usage

To get working with models, first you should import them

from tortoise.models import Model

With that you can start describing your own models like that

class Tournament(Model):
    id = fields.IntField(pk=True)
    name = fields.TextField()
    created = fields.DatetimeField(auto_now_add=True)

    def __str__(self):
        return self.name


class Event(Model):
    id = fields.IntField(pk=True)
    name = fields.TextField()
    tournament = fields.ForeignKeyField('models.Tournament', related_name='events')
    participants = fields.ManyToManyField('models.Team', related_name='events', through='event_team')
    modified = fields.DatetimeField(auto_now=True)
    prize = fields.DecimalField(max_digits=10, decimal_places=2, null=True)

    def __str__(self):
        return self.name


class Team(Model):
    id = fields.IntField(pk=True)
    name = fields.TextField()

    def __str__(self):
        return self.name

Let see in details what we accomplished here:

class Tournament(Model):

Every model should be derived from base model. You also can derive from your own model subclasses and you can make abstract models like this

class AbstractTournament(Model):
    id = fields.IntField(pk=True)
    name = fields.TextField()
    created = fields.DatetimeField(auto_now_add=True)

    class Meta:
        abstract = True

    def __str__(self):
        return self.name

This models won't be created in schema generation and won't create relations to other models.

Further we have field fields.DatetimeField(auto_now=True). Options auto_now and auto_now_add work like Django's options.

Use of __models__

If you define the variable __models__ in the module which you load your models from, generate_schema will use that list, rather than automatically finding models for you.

Primary Keys

In Tortoise ORM we require that a model has a primary key.

That primary key will be accessible through a reserved field pk which will be an alias of whichever field has been nominated as a primary key. That alias field can be used as a field name when doing filtering e.g. .filter(pk=...) etc…

Note

We currently support single (non-composite) primary keys of any indexable field type, but only these field types are recommended:

IntField
BigIntField
CharField
UUIDField

One must define a primary key by setting a pk parameter to True. If you don't define a primary key, we will create a primary key of type IntField with name of id for you.

Note

If this is used on an Integer Field, generated will be set to True unless you explicitly pass generated=False as well.

Any of these are valid primary key definitions in a Model:

id = fields.IntField(pk=True)

checksum = fields.CharField(pk=True)

guid = fields.UUIDField(pk=True)

Inheritance

When defining models in Tortoise ORM, you can save a lot of repetitive work by leveraging from inheritance.

You can define fields in more generic classes and they are automatically available in derived classes. Base classes are not limited to Model classes. Any class will work. This way you are able to define your models in a natural and easy to maintain way.

Let's have a look at some examples.

from tortoise import fields
from tortoise.models import Model

class TimestampMixin():
    created_at = fields.DatetimeField(null=True, auto_now_add=True)
    modified_at = fields.DatetimeField(null=True, auto_now=True)

class NameMixin():
    name = fields.CharField(40, unique=True)

class MyAbstractBaseModel(Model):
    id = fields.IntField(pk=True)

    class Meta:
        abstract = True

class UserModel(TimestampMixin, MyAbstractBaseModel):
    # Overriding the id definition
    # from MyAbstractBaseModel
    id = fields.UUIDField(pk=True)

    # Adding additional fields
    first_name = fields.CharField(20, null=True)

    class Meta:
        table = "user"


class RoleModel(TimestampMixin, NameMixin, MyAbstractBaseModel):

    class Meta:
        table = "role"

Using the Meta class is not necessary. But it is a good habit, to give your table an explicit name. This way you can change the model name without breaking the schema. So the following definition is valid.

class RoleModel(TimestampMixin, NameMixin, MyAbstractBaseModel):
    pass

The Meta class

.. autoclass:: tortoise.models.Model.Meta

    .. attribute:: abstract
        :annotation: = False

        Set to ``True`` to indicate this is an abstract class

    .. attribute:: schema
        :annotation: = ""

        Set this to configure a schema name, where table exists

    .. attribute:: table
        :annotation: = ""

        Set this to configure a manual table name, instead of a generated one

    .. attribute:: table_description
        :annotation: = ""

        Set this to generate a comment message for the table being created for the current model

    .. attribute:: unique_together
        :annotation: = None

        Specify ``unique_together`` to set up compound unique indexes for sets of columns.

        It should be a tuple of tuples (lists are fine) in the format of:

        .. code-block:: python3

            unique_together=("field_a", "field_b")
            unique_together=(("field_a", "field_b"), )
            unique_together=(("field_a", "field_b"), ("field_c", "field_d", "field_e"))

    .. attribute:: indexes
        :annotation: = None

        Specify ``indexes`` to set up compound non-unique indexes for sets of columns.

        It should be a tuple of tuples (lists are fine) in the format of:

        .. code-block:: python3

            indexes=("field_a", "field_b")
            indexes=(("field_a", "field_b"), )
            indexes=(("field_a", "field_b"), ("field_c", "field_d", "field_e"))

    .. attribute:: ordering
        :annotation: = None

        Specify ``ordering`` to set up default ordering for given model.
        It should be iterable of strings formatted in same way as ``.order_by(...)`` receives.
        If query is built with ``GROUP_BY`` clause using ``.annotate(...)`` default ordering is not applied.

        .. code-block:: python3

            ordering = ["name", "-score"]

    .. attribute:: manager
        :annotation: = tortoise.manager.Manager

        Specify ``manager`` to override the default manager.
        It should be instance of ``tortoise.manager.Manager`` or subclass.

        .. code-block:: python3

            manager = CustomManager()

ForeignKeyField

tournament = fields.ForeignKeyField('models.Tournament', related_name='events')
participants = fields.ManyToManyField('models.Team', related_name='events')
modified = fields.DatetimeField(auto_now=True)
prize = fields.DecimalField(max_digits=10, decimal_places=2, null=True)

In event model we got some more fields, that could be interesting for us.

fields.ForeignKeyField('models.Tournament', related_name='events')
Here we create foreign key reference to tournament. We create it by referring to model by it's literal, consisting of app name and model name. models is default app name, but you can change it in class Meta with app = 'other'.
related_name
Is keyword argument, that defines field for related query on referenced models, so with that you could fetch all tournaments's events with like this:
await Tournament.first().prefetch_related("events")

The DB-backing field

Note

A ForeignKeyField is a virtual field, meaning it has no direct DB backing. Instead it has a field (by default called :samp:`{FKNAME}_id` (that is, just an _id is appended) that is the actual DB-backing field.

It will just contain the Key value of the related table.

This is an important detail as it would allow one to assign/read the actual value directly, which could be considered an optimization if the entire foreign object isn't needed.

Specifying an FK can be done via either passing the object:

await SomeModel.create(tournament=the_tournament)
# or
somemodel.tournament=the_tournament

or by directly accessing the DB-backing field:

await SomeModel.create(tournament_id=the_tournament.pk)
# or
somemodel.tournament_id=the_tournament.pk

Querying a relationship is typically done by appending a double underscore, and then the foreign object's field. Then a normal query attr can be appended. This can be chained if the next key is also a foreign object:

:samp:`{FKNAME}__{FOREIGNFIELD}__gt=3`

or

:samp:`{FKNAME}__{FOREIGNFK}__{VERYFOREIGNFIELD}__gt=3`

There is however one major limitation. We don't want to restrict foreign column names, or have ambiguity (e.g. a foreign object may have a field called isnull)

Then this would be entirely ambiguous:

:samp:`{FKNAME}__isnull`

To prevent that we require that direct filters be applied to the DB-backing field of the foreign key:

:samp:`{FKNAME}_id__isnull`

Fetching the foreign object

Fetching foreign keys can be done with both async and sync interfaces.

Async fetch:

events = await tournament.events.all()

You can async iterate over it like this:

async for event in tournament.events:
    ...

Sync usage requires that you call fetch_related before the time, and then you can use common functions such as:

await tournament.fetch_related('events')
events = list(tournament.events)
eventlen = len(tournament.events)
if SomeEvent in tournament.events:
    ...
if tournament.events:
    ...
firstevent = tournament.events[0]

To get the Reverse-FK, e.g. an event.tournament we currently only support the sync interface.

await event.fetch_related('tournament')
tournament = event.tournament

ManyToManyField

Next field is fields.ManyToManyField('models.Team', related_name='events'). It describes many to many relation to model Team.

To add to a ManyToManyField both the models need to be saved, else you will get an OperationalError raised.

Resolving many to many fields can be done with both async and sync interfaces.

Async fetch:

participants = await tournament.participants.all()

You can async iterate over it like this:

async for participant in tournament.participants:
    ...

Sync usage requires that you call fetch_related before the time, and then you can use common functions such as:

await tournament.fetch_related('participants')
participants = list(tournament.participants)
participantlen = len(tournament.participants)
if SomeParticipant in tournament.participants:
    ...
if tournament.participants:
    ...
firstparticipant = tournament.participants[0]

The reverse lookup of team.event_team works exactly the same way.

Improving relational type hinting

Since Tortoise ORM is still a young project, it does not have such widespread support by various editors who help you writing code using good autocomplete for models and different relations between them. However, you can get such autocomplete by doing a little work yourself. All you need to do is add a few annotations to your models for fields that are responsible for the relations.

Here is an updated example from :ref:`getting_started`, that will add autocomplete for all models including fields for the relations between models.

from tortoise.models import Model
from tortoise import fields


class Tournament(Model):
    id = fields.IntField(pk=True)
    name = fields.CharField(max_length=255)

    events: fields.ReverseRelation["Event"]

    def __str__(self):
        return self.name


class Event(Model):
    id = fields.IntField(pk=True)
    name = fields.CharField(max_length=255)
    tournament: fields.ForeignKeyRelation[Tournament] = fields.ForeignKeyField(
        "models.Tournament", related_name="events"
    )
    participants: fields.ManyToManyRelation["Team"] = fields.ManyToManyField(
        "models.Team", related_name="events", through="event_team"
    )

    def __str__(self):
        return self.name


class Team(Model):
    id = fields.IntField(pk=True)
    name = fields.CharField(max_length=255)

    events: fields.ManyToManyRelation[Event]

    def __str__(self):
        return self.name

Reference

.. automodule:: tortoise.models
    :members: Model
    :undoc-members: