DateFeaturesConverter
Break a Date column into the calendar numbers hidden inside it.
A date is stored as text and a format, which no model can read. What a
forecast needs from it is calendar position: which month, which weekday,
which week of the year. This converter writes those out as integer
columns named <column>_<feature>, one set per Date column in
scope, and leaves the original column untouched.
The columns it produces are what the exogenous role of
SingleSeriesForecastingTask takes beside the series: variables known
for every period being forecast, since a calendar is known in advance.
The role takes only variables like those, known ahead of time; a price
goes in only once it has been set for the periods being forecast. The
time index itself is protected from the session's preprocessing, so the
calendar comes from a copy of the date in a column of its own.
Cyclical encoding answers the one problem plain calendar numbers have.
Month 12 and month 1 are one step apart in the calendar and eleven apart
as integers, so a model reading the integer treats the year end as the
furthest point from the year start. Encoding a periodic feature of period
p as sin(2*pi*v/p) and cos(2*pi*v/p) puts consecutive values
next to each other on a circle, which is where they belong.
Parameters
- year : boolean, default=
True - Add the calendar year of each date as an integer column.
- month : boolean, default=
True - Add the month of the year (1 to 12) as an integer column.
- day : boolean, default=
False - Add the day of the month (1 to 31) as an integer column.
- weekday : boolean, default=
True - Add the day of the week as an integer column, with Monday as 0 and Sunday as 6.
- quarter : boolean, default=
False - Add the quarter of the year (1 to 4) as an integer column.
- week_of_year : boolean, default=
False - Add the ISO week number (1 to 53) as an integer column.
- day_of_year : boolean, default=
False - Add the day of the year (1 to 366) as an integer column.
- cyclical : boolean, default=
False - Also encode every periodic feature that is on as a sine and cosine pair, so that December sits next to January instead of eleven units away from it.
- drop_source_columns : boolean, default=
False - Remove the columns this converter reads once it has created its new ones.
Methods
fit(self, x: 'DashAIDataset', y: Optional[ForwardRef('DashAIDataset')] = None) -> 'DateFeaturesConverter'
DateFeaturesConverterFind the Date columns in scope and name the columns to be added.
Parameters
- x : DashAIDataset
- The scoped columns.
- y : DashAIDataset, optional
- Ignored. Defaults to None.
Returns
- DateFeaturesConverter
- The fitted converter instance (self).
get_output_type(self, column_name: str = None) -> DashAI.back.types.dashai_data_type.DashAIDataType
DateFeaturesConverterReturn the type of one extracted column.
Parameters
- column_name : str, optional
- Name of the output column, such as
"date_month"or"date_month_sin". Defaults to None.
Returns
- DashAIDataType
Floatfor a sine or cosine column,Integerfor every plain calendar feature.
infer_output_columns(self, inputs: List[Annotated[Union[DashAI.back.preprocessing.structure_types.ColumnItem, DashAI.back.preprocessing.structure_types.BlockItem], FieldInfo(annotation=NoneType, required=True, discriminator='kind')]]) -> DashAI.back.preprocessing.structure_types.StructureDelta
DateFeaturesConverterEstimate the output: the scope kept, plus the calendar features.
Parameters
- inputs : list of ColumnItem | BlockItem
- The dataset state items in this converter's scope.
Returns
- StructureDelta
- The unchanged inputs, plus one column per feature and date column.
transform(self, x: 'DashAIDataset', y: Optional[ForwardRef('DashAIDataset')] = None) -> 'DashAIDataset'
DateFeaturesConverterAppend the selected calendar features of every fitted date column.
Parameters
- x : DashAIDataset
- The dataset to transform.
- y : DashAIDataset, optional
- Ignored. Defaults to None.
Returns
- DashAIDataset
- The dataset with one integer column per selected feature and date column, plus a sine and cosine column per periodic feature when cyclical encoding is on. A row whose date is missing gets a null in every one of them.
build_report(self, data: 'Dict[str, Any]') -> 'Dict[str, Any]'
BaseConverterWrap converter-specific data with the producer converter name.
get_credential(self, name: str)
ConfigObjectResolve a registered credential component by name.
Parameters
- name : str
- Credential component class name (e.g. "HuggingFaceCredential").
Returns
- BaseCredential
- An instance of the requested credential component.
get_metadata(cls) -> 'Dict[str, Any]'
BaseConverterGet metadata for the converter, used by the DashAI frontend.
Parameters
- cls : type
- The converter class (injected automatically by Python for classmethods).
Returns
- Dict[str, Any]
- Dictionary containing display name, short description, image preview path, category, icon, color, and whether the converter is supervised.
get_report(self) -> 'Union[Dict[str, Any], None]'
BaseConverterReturn the converter report produced after execution, if any.
Returns
- Dict[str, Any] or None
- JSON-serializable report dict, or
Nonewhen the converter does not produce one.
get_schema(cls) -> dict
ConfigObjectGenerates the component related Json Schema.
Returns
- dict
- Dictionary representing the Json Schema of the component.
rows_to_keep(self, x: "'DashAIDataset'") -> 'List[int]'
BaseConverterPositions of the rows of x that survive this converter.
Parameters
- x : DashAIDataset
- The converter's scope columns of one split.
Returns
- list of int
- 0-based positions of the kept rows, in order.
session_params(cls, params: 'Optional[Dict[str, Any]]') -> 'Dict[str, Any]'
BaseConverterA session step's params, with the session defaults under them.
Parameters
- params : dict or None
- The params the step stores.
Returns
- dict
SESSION_PARAMSupdated withparams: what the step gives wins.
validate_and_transform(self, raw_data: dict) -> dict
ConfigObjectIt takes the data given by the user to initialize the model and returns it with all the objects that the model needs to work.
Parameters
- raw_data : dict
- A dictionary with the data provided by the user to initialize the model.
Returns
- dict
- A validated dictionary with the necessary objects.