Skip to content

Collection

A Collection is a structured set of data such as a database table or CSV file.
All the described entities are available in the igenius_adapters_sdk.entities.collection package.

class Collection(BaseModel):
    uid: str
    attributes_schema: AttributesSchema

Attributes

An Attribute schema provides information on the available data and its type.

class Attribute(BaseModel):
    uid: str
    type: AttributeType

AttributesSchema

An AttributesSchema contains a list of Attributes

class AttributesSchema(BaseModel):
    attributes: List[Attribute] = Field(default=list())

AttributeTypes

Attribute types provide a way to determine if a certain action is available on the attribute. There are 4 main attribute types and UNKNOWN for all that do not fall into the previous 4 categories. Attribute types are unified way between different adapters to describe data source schema.

  • boolean: The boolean attribute has true/ false state. Eg. BOOLEAN type in PostgreSQL but RAW(1) in Oracle databases.
  • categorical: Every attribute that can hold characters. Eg. VARCHAR, TEXT.
  • datetime: Every attribute that can hold information about the date. Eg. DATE, TIMESTAMP.
  • numeric: Every attribute that can hold numeric values. Eg. INT, FLOAT, DECIMAL.
  • unknown: Special attribute type for those whose data and possible operations on them isn't clear. Eg. ROWID, JSONB.
class AttributeType(str, Enum):
    BOOLEAN = 'crystal.topics.data.attribute-types.boolean'
    CATEGORICAL = 'crystal.topics.data.attribute-types.categorical'
    DATETIME = 'crystal.topics.data.attribute-types.datetime'
    NUMERIC = 'crystal.topics.data.attribute-types.numeric'
    UNKNOWN = 'crystal.topics.data.attribute-types.unknown'
Last update: February 2, 2021