从sql查询迁移到API请求时如何组织Python项目？

Question

我正在开发一个用于从外部数据库获取所有所需数据的项目。很快数据库将不再可访问，并将被端点取代。

假设我有这样的结构

project_root/
├── sql/
│   ├── query1.sql
│   ├── query2.sql
│   └── ...
├── main.py
└── ...

其中main有调用不同查询模块的函数。我应该如何命名新结构？像这样吗？

project_root/
├── requests/
│   ├── request1.py
│   ├── request2.py
│   └── ...
├── main.py
└── ...

Answer 1

假设您的项目中有

users

和

items

。 这里我按实体（

routes

、

repositories

、

controllers

等）对项目进行分组，但也可以按域（

users

、

items

）进行分组，这也很好。

project_root/
├── routes/  # Contains your endpoints, should not contain any business logic.
│   ├── user.py
│   ├── item.py
│   └── ...
├── clients/  # Contains db client, s3 client, any client to interact with other services etc.
│   ├── db_client.py
│   └── ...
├── models/  # Contains db models that should reflect the database structure. Commonly ORM is used e.g. SQLAlchemy. The result of your queries should be converted to the models.
│   ├── user.py
│   ├── item.py
│   └── ...
├── schemas/  # Contains pydantic models: schemas / DTOs
│   ├── user.py
│   ├── item.py
│   └── ...
├── repositories/  # The only place containing database queries. Ideally models should be returned from repository methods.
│   ├── user.py
│   ├── item.py
│   └── ...
├── controllers/  # Contains business logic: take something from client, repository, do something, prepare API answer.
│   ├── user.py
│   ├── item.py
│   └── ...
├── main.py
└── ...

模型和模式之间的差异以及我提到的实体示例，您可以在我之前的答案中找到。