How to do it...

In the following steps, we will validate a set of data that's been loaded from CSV using the Cerberus package:

1. First, we need to build a schema that describes the data we expect. To do this, we must define a simple schema for floating-point numbers:

float_schema = {"type": "float", "coerce": float, "min": -1.0,
   "max": 1.0}

2. Next, we build the schema for individual items. These will be the rows of our data:

item_schema = {
    "type": "dict",
    "schema": {
        "id": {"type": "string"},
        "number": {"type": "integer", "coerce": int},
        "lower": float_schema,
        "upper": float_schema,
    }
}

3. Now, we can define the schema for the whole document, which will contain a list of items:
...