Schema

Schemas in SP-Lang describe the type and other properties of fields in dynamically types containers such as JSON or Python dictionaries.

It is important to provide information about the type to the SP-Lang because it is used as an input for a type inference and hence optimal performance.

Schema definition

YAML representation of the schema:

---
define:
  type: splang/schema

fields:
  field1:
    type: str
    aliases: ["FieldOne"]

  field2:
    type: ui64

Options

Option type

Defines the data type for the given attribute, such as str, si64 and so on. Refer to a SP-Lang type system for more information.

This option is mandatory.

Option aliases

Defines field aliases for the given attribute, that can be used in the declaration as a synonymic term.

If an field1 has a field alias named FieldOne, the following declarations are equal if the schema is properly defined:

!GET
what: field1
from: !ARG input

!GET
what: FieldOne
from: !ARG input

Option unit

Defines the unit of the attribute, if needed, such as for timestamps. In this case, the unit can be auto for automatical detection, seconds and microseconds.

Function declaration (Python)

The example of the SP-Lang function declaration that uses MYSCHEMA.yaml:

splang.FunctionDeclaration(
    name="main",
    returns="bool",
    arguments={
        'myArgument': 'json<MYSCHEMA>'
    },
)

and MYSCHEMA.yaml itself:

---
define:
  type: splang/schema

fields:
  field1:
    type: str

  field2:
    type: ui64

In-place schemas

SP-Lang allows to specify schema directly in the FunctionDeclaration Python code:

splang.FunctionDeclaration(
    name="main",
    returns="bool",
    arguments={
        'myArgument': 'json<INPLACESCHEMA>'
    },
    schemas=[
        ('INPLACESCHEMA', {
            "field1": "str",
            "field2": "si32",
            "field3": "ui64",
        })
    ]
)

It is done by using tuple, the first item is a schema name, the second is a dictionary with fields.