pycontrails.utils.types¶

Convienence types.

Module Attributes

`ArrayLike`	Array like (np.ndarray, xr.DataArray)
`ArrayOrFloat`	Array or Float (np.ndarray, float)
`ArrayScalarLike`	Array like input (np.ndarray, xr.DataArray, float)
`DatetimeLike`	Datetime like input (datetime, pd.Timestamp, np.datetime64)

Functions

`apply_nan_mask_to_arraylike`(arr, nan_mask)	Apply `nan_mask` to `arr` while maintaining the type.
`support_arraylike`(func)	Extend a numpy universal function operating on arrays of floats.
`type_guard`(obj, type_[, error_message])	Shortcut utility to type guard a variable with custom error message.

class pycontrails.utils.types.ArrayLike¶

Array like (np.ndarray, xr.DataArray)

alias of TypeVar(‘ArrayLike’, ~numpy.ndarray, ~xarray.core.dataarray.DataArray)

has_default()¶

class pycontrails.utils.types.ArrayOrFloat¶

Array or Float (np.ndarray, float)

alias of TypeVar(‘ArrayOrFloat’, ~numpy.ndarray, float)

has_default()¶

class pycontrails.utils.types.ArrayScalarLike¶

Array like input (np.ndarray, xr.DataArray, float)

alias of TypeVar(‘ArrayScalarLike’, ~numpy.ndarray, ~xarray.core.dataarray.DataArray, float)

has_default()¶

class pycontrails.utils.types.DatetimeLike¶

Datetime like input (datetime, pd.Timestamp, np.datetime64)

alias of TypeVar(‘DatetimeLike’, ~datetime.datetime, ~pandas._libs.tslibs.timestamps.Timestamp, ~numpy.datetime64, str)

has_default()¶

pycontrails.utils.types.apply_nan_mask_to_arraylike(arr, nan_mask)¶

Apply nan_mask to arr while maintaining the type.

The parameter arr should have a float dtype.

This function is tested against xr.DataArray, pd.Series, and np.ndarray types.

Parameters:

arr (ArrayLike) – Array with np.float64 entries
nan_mask (numpy.ndarray) – Boolean array of the same shape as arr

Returns:

ArrayLike – Array arr with values in nan_mask set to np.nan. The arr is mutated in place if it is a np.ndarray. For xr.DataArray, a copy is returned.

Notes

When arr is a xr.DataArray, this function keeps any attrs from arr in the returned instance.

pycontrails.utils.types.support_arraylike(func)¶

Extend a numpy universal function operating on arrays of floats.

This decorator allows func to support any ArrayScalarLike parameter and keeps the return type consistent with the parameter.

Parameters:: func (Callable[[ArrayScalarLike], np.ndarray]) – A numpy ufunc taking in a single array with float-like dtype. This decorator assumes func returns a numpy array.
Returns:: Callable[[ArrayScalarLike], ArrayScalarLike] – Extended function.