# `Spark.Dsl.Entity`
[🔗](https://github.com/ash-project/spark/blob/v2.7.0/lib/spark/dsl/entity.ex#L5)

Declares a DSL entity.

A dsl entity represents a dsl constructor who's resulting value is a struct.
This lets the user create complex objects with arbitrary(mostly) validation rules.

The lifecycle of creating entities is complex, happening as Elixir is compiling
the modules in question. Some of the patterns around validating/transforming entities
have not yet solidified. If you aren't careful and don't follow the guidelines listed
here, you can have subtle and strange bugs during compilation. Anything not isolated to
simple value validations should be done in `transformers`. See `Spark.Dsl.Transformer`.

An entity has a `target` indicating which struct will ultimately be built. An entity
also has a `schema`. This schema is used for documentation, and the options are validated
against it before continuing on with the DSL.

To create positional arguments to the builder, use `args`. The values provided to
`args` need to be in the provided schema as well. They will be positional arguments
in the same order that they are provided in the `args` key.

`auto_set_fields` will set the provided values into the produced struct (they do not need
to be included in the schema).

`transform` is a function that takes a created struct and can alter it. This happens immediately
after handling the DSL options, and can be useful for setting field values on a struct based on
other values in that struct. If you need things that aren't contained in that struct, use an
`Spark.Dsl.Transformer`. This function returns `{:ok, new_entity}` or `{:error, error}`, so this can
also be used to validate the entity.

`entities` allows you to specify a keyword list of nested entities. Nested entities are stored
on the struct in the corresponding key, and are used in the same way entities are otherwise.

`singleton_entity_keys` specifies a set of entity keys (specified above) that should only have a
single value. This will be validated and unwrapped into `nil` | `single_value` on success.

`identifier` expresses that a given entity is unique by that field, validated by the DSL.

## Example

```elixir
@my_entity %Spark.Dsl.Entity{
  name: :my_entity,
  target: MyStruct,
  schema: [my_field: [type: :atom, required: false]]
}
```

Once compiled by Spark, entities can be invoked with a keyword list:

```elixir
my_entity my_field: :value
```

Or with a do block:

```elixir
my_entity do
  my_field :value
end
```

For a full example, see `Spark.Dsl.Extension`.

# `args`

```elixir
@type args() :: [atom() | {:optional, atom()} | {:optional, atom(), any()}]
```

Specifies positional arguments for an Entity.

An entity declared like this:

```elixir
@entity %Spark.Dsl.Entity{
  name: :entity,
  target: Entity,
  schema: [
    positional: [type: :atom, required: true],
    other: [type: :atom, required: false],
  ],
  args: [:positional]
}
```

Can be instantiated like this:

```elixir
entity :positional_argument do
  other :other_argument
end
```

# `auto_set_fields`

```elixir
@type auto_set_fields() :: keyword(any())
```

Set the provided key value pairs in the produced struct. These fields do not need to be included in the Entity's schema.

# `deprecations`

```elixir
@type deprecations() :: keyword(String.t())
```

# `describe`

```elixir
@type describe() :: String.t()
```

User provided documentation.

Documentation provided in a `Entity`'s `describe` field will be included by `Spark` in any generated documentation that includes the `Entity`.

# `docs`

```elixir
@type docs() :: String.t()
```

Internal field. Not set by user.

# `entities`

```elixir
@type entities() :: keyword(t())
```

A keyword list of nested entities.

# `entity`

```elixir
@type entity() :: %{
  :__struct__ => module(),
  :__spark_metadata__ => spark_meta() | nil,
  optional(:__identifier__) => term(),
  optional(atom()) => term()
}
```

# `examples`

```elixir
@type examples() :: [String.t()]
```

# `hide`

```elixir
@type hide() :: [atom()]
```

# `id`

```elixir
@type id() :: term()
```

# `imports`

```elixir
@type imports() :: [module()]
```

# `links`

```elixir
@type links() :: keyword([String.t()]) | nil
```

# `modules`

```elixir
@type modules() :: [atom()]
```

# `name`

```elixir
@type name() :: atom() | nil
```

# `no_depend_modules`

```elixir
@type no_depend_modules() :: [atom()]
```

# `recursive_as`

```elixir
@type recursive_as() :: atom() | nil
```

# `singleton_entity_keys`

```elixir
@type singleton_entity_keys() :: [atom()]
```

# `snippet`

```elixir
@type snippet() :: String.t()
```

# `spark_meta`

```elixir
@opaque spark_meta()
```

# `t`

```elixir
@type t() :: %Spark.Dsl.Entity{
  args: args(),
  auto_set_fields: auto_set_fields(),
  deprecations: deprecations(),
  describe: describe(),
  docs: docs(),
  entities: entities(),
  examples: examples(),
  hide: hide(),
  identifier: id(),
  imports: imports(),
  links: links(),
  modules: modules(),
  name: name(),
  no_depend_modules: no_depend_modules(),
  recursive_as: recursive_as(),
  schema: Spark.Options.schema(),
  singleton_entity_keys: singleton_entity_keys(),
  snippet: snippet(),
  target: target(),
  transform: transform()
}
```

# `target`

```elixir
@type target() :: module() | nil
```

Defines the struct that will be built from this entity definition.

The struct will need to have fields for all [`entities`](#t:entities/0), `t:schema/0` fields, and `t:auto_set_fields/0`.

# `transform`

```elixir
@type transform() :: {module(), function :: atom(), args :: [any()]} | nil
```

Specifies a function that will run on the target struct after building.

```elixir
@my_entity %Spark.Dsl.Entity{
  name: :my_entity,
  target: MyEntity,
  schema: [
    my_field: [type: :list, required: true]
  ],
  transform: {MyModule, :max_three_items, []}
}

def max_three_items(my_entity) do
  if length(my_entity.my_field) > 3 do
    {:error, "Can't have more than three items"}
  else
    {:ok, my_entity}
  end
end
```

# `anno`

```elixir
@spec anno(entity()) :: :erl_anno.anno() | nil
```

Get the annotation for an entity.

Returns the annotation that was captured when the entity was defined,
or `nil` if no annotation is available or the entity doesn't have the
`:__spark_metadata__` field.

# `property_anno`

```elixir
@spec property_anno(entity(), atom()) :: :erl_anno.anno() | nil
```

Get the annotation for a specific property of an entity.

Returns the annotation that was captured when the property was set on the entity,
or `nil` if no annotation is available for that property or the entity doesn't have
the `:__spark_metadata__` field.

---

*Consult [api-reference.md](api-reference.md) for complete listing*