Annotations

Sometimes multiple columns of a tabular dataset describe a single entity. Annotations provide a way to link such columns together in order to synthesize higher quality data. See the annotations guide for more details.

Address

Python
YAML

from synthesized.config import AddressLabels
from synthesized.metadata.value import Address

address = Address(
    name="annotated_columns",
    labels=AddressLabels(
        postcode="col_postcode",
        county="col_country",
        city="col_city",
        district="col_district",
        street="col_street",
        house_number="col_house"
    )
)

address:
  - name: annotated_columns
    labels:
      postcode: col_postcode
      county: col_country
      city: col_city
      district: col_district
      street: col_street
      house_number: col_house

Properties

name: Name for the annotation. Cannot be the same as any existing column name.
labels (optional): Mapping of labels to column names.

Bank

Python
YAML

from synthesized.config import BankLabels
from synthesized.metadata.value import Bank

address = Bank(
    name="annotated_columns",
    labels=BankLabels(
        sort_code="col_sc",
        account="col_acc"
    )
)

bank:
  - name: annotated_columns
    labels:
      sort_code: col_sc
      account: col_acc

Properties

name: Name for the annotation. Cannot be the same as any existing column name.
labels (optional): Mapping of labels to column names.

Formatted String

Python
YAML

from synthesized.metadata.value import FormattedString

regex = "^(?!666|000|9\\d{2})\\d{3}-(?!00)\\d{2}-(?!0{4})\\d{4}$"
p1 = FormattedString(
  name="colA",
  pattern=regex
)

formatted_string:
  - name: colA
    pattern: "^(?!666|000|9\\d{2})\\d{3}-(?!00)\\d{2}-(?!0{4})\\d{4}$"

Properties

name: Name for the annotation. Cannot be the same as any existing column name.
pattern: A regex string describing the format of the strings in the column.

Person

Python
YAML

from synthesized.config import PersonLabels
from synthesized.metadata.value import Person

p1 = Person(
  name="annotated_columns",
  labels=PersonLabels(
    first_name="col_first",
    last_name="col_last",
    gender="col_gender",
    email="col_email"
  )
)

person:
  - name: annotated_columns
    labels:
      first_name: col_first
      last_name: col_last
      gender: col_gender
      email: col_email

Properties

name: Name for the annotation. Cannot be the same as any existing column name.
labels (optional): Mapping of labels to column names.