HistorySchema Documentation

HistorySchema Documentation#

The HistorySchema class defines the structure of a dataset, including its index and column names, each mapped to their respective data types. This schema is often used in data storage, retrieval, and serialization processes.

Class Overview#

class HistorySchema(Serializable, metaclass=RegistryBase):
    """
    A schema is the structure of a data set.
    It contains the index names mapped to their respective types and levels,
    as well as the column names mapped to their types.
    """

Usage#

Creating a Schema#

You can create a HistorySchema object by specifying the index and columns:

schema = HistorySchema(
    index={"security": str, "date": pd.Timestamp},
    columns={"open": float, "close": float}
)

Accessing Schema Properties#

  • index_names: A list of the index names.

    schema.index_names  # ['security', 'date']

  • column_names: A list of the column names.

    schema.column_names  # ['open', 'close']

Copying a Schema#

You can create a copy of the schema:

schema_copy = schema.copy()

Checking Index and Column Membership#

You can check if an index or column name exists in the schema:

schema.in_index('security')  # True
schema.in_column('open')     # True

Attributes#

  • index (Dict[str, Type]): A dictionary where keys are index names (e.g., security, date) and values are the respective types (e.g., str, Timestamp).

  • columns (Dict[str, Type]): A dictionary where keys are column names (e.g., open, close) and values are the respective types (e.g., float).

Serialization#

HistorySchema supports serialization to and from dictionaries and JSON.

Converting to a Dictionary#

You can convert the schema to a dictionary representation:

schema_dict = schema.to_dict()
# Output: {'index': {'security': 'str', 'date': 'Timestamp'}, 'columns': {'open': 'float', 'close': 'float'}}

Deserializing from a Dictionary#

You can reconstruct the schema from a dictionary:

schema_from_dict = HistorySchema.from_dict(schema_dict)

Converting to JSON#

You can serialize the schema to a JSON string:

schema_json = schema.__json__()
# Output: '{"index": {"security": "str", "date": "Timestamp"}, "columns": {"open": "float", "close": "float"}}'

Saving and Loading from File#

The schema can be saved to a file and loaded back.

  • Save to File:

schema.store('/path/to/save', 'schema_key')

  • Load from File:

loaded_schema = HistorySchema.load('/path/to/save', 'schema_key')

Class Autodoc#

Contents