请求体 - 嵌套模型

使用 FastAPI,你可以定义、校验、记录文档并使用任意深度嵌套的模型(归功于Pydantic)。

List 字段

你可以将一个属性定义为拥有子元素的类型。例如 Python list

Python 3.10+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel
  3. app = FastAPI()
  4. class Item(BaseModel):
  5. name: str
  6. description: str | None = None
  7. price: float
  8. tax: float | None = None
  9. tags: list = []
  10. @app.put("/items/{item_id}")
  11. async def update_item(item_id: int, item: Item):
  12. results = {"item_id": item_id, "item": item}
  13. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: list = []
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results

这将使 tags 成为一个由元素组成的列表。不过它没有声明每个元素的类型。

具有子类型的 List 字段

但是 Python 有一种特定的方法来声明具有子类型的列表:

从 typing 导入 List

首先,从 Python 的标准库 typing 模块中导入 List

  1. from typing import List, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: List[str] = []
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results

声明具有子类型的 List

要声明具有子类型的类型,例如 listdicttuple

  • typing 模块导入它们
  • 使用方括号 [] 将子类型作为「类型参数」传入
  1. from typing import List
  2. my_list: List[str]

这完全是用于类型声明的标准 Python 语法。

对具有子类型的模型属性也使用相同的标准语法。

因此,在我们的示例中,我们可以将 tags 明确地指定为一个「字符串列表」:

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel
  3. app = FastAPI()
  4. class Item(BaseModel):
  5. name: str
  6. description: str | None = None
  7. price: float
  8. tax: float | None = None
  9. tags: list[str] = []
  10. @app.put("/items/{item_id}")
  11. async def update_item(item_id: int, item: Item):
  12. results = {"item_id": item_id, "item": item}
  13. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: list[str] = []
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results
  1. from typing import List, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: List[str] = []
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results

Set 类型

但是随后我们考虑了一下,意识到标签不应该重复,它们很大可能会是唯一的字符串。

Python 具有一种特殊的数据类型来保存一组唯一的元素,即 set

然后我们可以导入 Set 并将 tag 声明为一个由 str 组成的 set

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel
  3. app = FastAPI()
  4. class Item(BaseModel):
  5. name: str
  6. description: str | None = None
  7. price: float
  8. tax: float | None = None
  9. tags: set[str] = set()
  10. @app.put("/items/{item_id}")
  11. async def update_item(item_id: int, item: Item):
  12. results = {"item_id": item_id, "item": item}
  13. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: set[str] = set()
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results
  1. from typing import Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Item(BaseModel):
  6. name: str
  7. description: Union[str, None] = None
  8. price: float
  9. tax: Union[float, None] = None
  10. tags: Set[str] = set()
  11. @app.put("/items/{item_id}")
  12. async def update_item(item_id: int, item: Item):
  13. results = {"item_id": item_id, "item": item}
  14. return results

这样,即使你收到带有重复数据的请求,这些数据也会被转换为一组唯一项。

而且,每当你输出该数据时,即使源数据有重复,它们也将作为一组唯一项输出。

并且还会被相应地标注 / 记录文档。

嵌套模型

Pydantic 模型的每个属性都具有类型。

但是这个类型本身可以是另一个 Pydantic 模型。

因此,你可以声明拥有特定属性名称、类型和校验的深度嵌套的 JSON 对象。

上述这些都可以任意的嵌套。

定义子模型

例如,我们可以定义一个 Image 模型:

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: str
  6. name: str
  7. class Item(BaseModel):
  8. name: str
  9. description: str | None = None
  10. price: float
  11. tax: float | None = None
  12. tags: set[str] = set()
  13. image: Image | None = None
  14. @app.put("/items/{item_id}")
  15. async def update_item(item_id: int, item: Item):
  16. results = {"item_id": item_id, "item": item}
  17. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: str
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results
  1. from typing import Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: str
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: Set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results

将子模型用作类型

然后我们可以将其用作一个属性的类型:

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: str
  6. name: str
  7. class Item(BaseModel):
  8. name: str
  9. description: str | None = None
  10. price: float
  11. tax: float | None = None
  12. tags: set[str] = set()
  13. image: Image | None = None
  14. @app.put("/items/{item_id}")
  15. async def update_item(item_id: int, item: Item):
  16. results = {"item_id": item_id, "item": item}
  17. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: str
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results
  1. from typing import Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: str
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: Set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results

这意味着 FastAPI 将期望类似于以下内容的请求体:

  1. {
  2. "name": "Foo",
  3. "description": "The pretender",
  4. "price": 42.0,
  5. "tax": 3.2,
  6. "tags": ["rock", "metal", "bar"],
  7. "image": {
  8. "url": "http://example.com/baz.jpg",
  9. "name": "The Foo live"
  10. }
  11. }

再一次,仅仅进行这样的声明,你将通过 FastAPI 获得:

  • 对被嵌入的模型也适用的编辑器支持(自动补全等)
  • 数据转换
  • 数据校验
  • 自动生成文档

特殊的类型和校验

除了普通的单一值类型(如 strintfloat 等)外,你还可以使用从 str 继承的更复杂的单一值类型。

要了解所有的可用选项,请查看关于 来自 Pydantic 的外部类型 的文档。你将在下一章节中看到一些示例。

例如,在 Image 模型中我们有一个 url 字段,我们可以把它声明为 Pydantic 的 HttpUrl,而不是 str

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel, HttpUrl
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: HttpUrl
  6. name: str
  7. class Item(BaseModel):
  8. name: str
  9. description: str | None = None
  10. price: float
  11. tax: float | None = None
  12. tags: set[str] = set()
  13. image: Image | None = None
  14. @app.put("/items/{item_id}")
  15. async def update_item(item_id: int, item: Item):
  16. results = {"item_id": item_id, "item": item}
  17. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results
  1. from typing import Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: Set[str] = set()
  14. image: Union[Image, None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results

该字符串将被检查是否为有效的 URL,并在 JSON Schema / OpenAPI 文档中进行记录。

带有一组子模型的属性

你还可以将 Pydantic 模型用作 listset 等的子类型:

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel, HttpUrl
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: HttpUrl
  6. name: str
  7. class Item(BaseModel):
  8. name: str
  9. description: str | None = None
  10. price: float
  11. tax: float | None = None
  12. tags: set[str] = set()
  13. images: list[Image] | None = None
  14. @app.put("/items/{item_id}")
  15. async def update_item(item_id: int, item: Item):
  16. results = {"item_id": item_id, "item": item}
  17. return results
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: set[str] = set()
  14. images: Union[list[Image], None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results
  1. from typing import List, Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: Set[str] = set()
  14. images: Union[List[Image], None] = None
  15. @app.put("/items/{item_id}")
  16. async def update_item(item_id: int, item: Item):
  17. results = {"item_id": item_id, "item": item}
  18. return results

这将期望(转换,校验,记录文档等)下面这样的 JSON 请求体:

  1. {
  2. "name": "Foo",
  3. "description": "The pretender",
  4. "price": 42.0,
  5. "tax": 3.2,
  6. "tags": [
  7. "rock",
  8. "metal",
  9. "bar"
  10. ],
  11. "images": [
  12. {
  13. "url": "http://example.com/baz.jpg",
  14. "name": "The Foo live"
  15. },
  16. {
  17. "url": "http://example.com/dave.jpg",
  18. "name": "The Baz"
  19. }
  20. ]
  21. }

Info

请注意 images 键现在具有一组 image 对象是如何发生的。

深度嵌套模型

你可以定义任意深度的嵌套模型:

Python 3.10+Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel, HttpUrl
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: HttpUrl
  6. name: str
  7. class Item(BaseModel):
  8. name: str
  9. description: str | None = None
  10. price: float
  11. tax: float | None = None
  12. tags: set[str] = set()
  13. images: list[Image] | None = None
  14. class Offer(BaseModel):
  15. name: str
  16. description: str | None = None
  17. price: float
  18. items: list[Item]
  19. @app.post("/offers/")
  20. async def create_offer(offer: Offer):
  21. return offer
  1. from typing import Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: set[str] = set()
  14. images: Union[list[Image], None] = None
  15. class Offer(BaseModel):
  16. name: str
  17. description: Union[str, None] = None
  18. price: float
  19. items: list[Item]
  20. @app.post("/offers/")
  21. async def create_offer(offer: Offer):
  22. return offer
  1. from typing import List, Set, Union
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. class Item(BaseModel):
  9. name: str
  10. description: Union[str, None] = None
  11. price: float
  12. tax: Union[float, None] = None
  13. tags: Set[str] = set()
  14. images: Union[List[Image], None] = None
  15. class Offer(BaseModel):
  16. name: str
  17. description: Union[str, None] = None
  18. price: float
  19. items: List[Item]
  20. @app.post("/offers/")
  21. async def create_offer(offer: Offer):
  22. return offer

Info

请注意 Offer 拥有一组 Item 而反过来 Item 又是一个可选的 Image 列表是如何发生的。

纯列表请求体

如果你期望的 JSON 请求体的最外层是一个 JSON array(即 Python list),则可以在路径操作函数的参数中声明此类型,就像声明 Pydantic 模型一样:

  1. images: List[Image]

例如:

Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. from pydantic import BaseModel, HttpUrl
  3. app = FastAPI()
  4. class Image(BaseModel):
  5. url: HttpUrl
  6. name: str
  7. @app.post("/images/multiple/")
  8. async def create_multiple_images(images: list[Image]):
  9. return images
  1. from typing import List
  2. from fastapi import FastAPI
  3. from pydantic import BaseModel, HttpUrl
  4. app = FastAPI()
  5. class Image(BaseModel):
  6. url: HttpUrl
  7. name: str
  8. @app.post("/images/multiple/")
  9. async def create_multiple_images(images: List[Image]):
  10. return images

无处不在的编辑器支持

你可以随处获得编辑器支持。

即使是列表中的元素:

请求体 - 嵌套模型 - 图1

如果你直接使用 dict 而不是 Pydantic 模型,那你将无法获得这种编辑器支持。

但是你根本不必担心这两者,传入的字典会自动被转换,你的输出也会自动被转换为 JSON。

任意 dict 构成的请求体

你也可以将请求体声明为使用某类型的键和其他类型值的 dict

无需事先知道有效的字段/属性(在使用 Pydantic 模型的场景)名称是什么。

如果你想接收一些尚且未知的键,这将很有用。


其他有用的场景是当你想要接收其他类型的键时,例如 int

这也是我们在接下来将看到的。

在下面的例子中,你将接受任意键为 int 类型并且值为 float 类型的 dict

Python 3.9+Python 3.8+

  1. from fastapi import FastAPI
  2. app = FastAPI()
  3. @app.post("/index-weights/")
  4. async def create_index_weights(weights: dict[int, float]):
  5. return weights
  1. from typing import Dict
  2. from fastapi import FastAPI
  3. app = FastAPI()
  4. @app.post("/index-weights/")
  5. async def create_index_weights(weights: Dict[int, float]):
  6. return weights

Tip

请记住 JSON 仅支持将 str 作为键。

但是 Pydantic 具有自动转换数据的功能。

这意味着,即使你的 API 客户端只能将字符串作为键发送,只要这些字符串内容仅包含整数,Pydantic 就会对其进行转换并校验。

然后你接收的名为 weightsdict 实际上将具有 int 类型的键和 float 类型的值。

总结

使用 FastAPI 你可以拥有 Pydantic 模型提供的极高灵活性,同时保持代码的简单、简短和优雅。

而且还具有下列好处:

  • 编辑器支持(处处皆可自动补全!)
  • 数据转换(也被称为解析/序列化)
  • 数据校验
  • 模式文档
  • 自动生成的文档