braintools.cogtask.identity#
- braintools.cogtask.identity(key, default=None)[source]#
Create an identity encoding that passes through values directly.
Useful when the context already contains properly formatted arrays that should be used directly without transformation.
- Parameters:
- Returns:
Encoder function (ctx, feature) -> np.ndarray
- Return type:
Examples
>>> import numpy as np >>> from braintools.cogtask import Context, Feature, identity >>> ctx = Context() >>> # The context holds a pre-computed encoding >>> ctx['custom_encoding'] = np.array([0.1, 0.5, 0.8, 0.3]) >>> identity('custom_encoding')(ctx, Feature(4, 'stimulus')) Array([0.1, 0.5, 0.8, 0.3], dtype=float32) >>> identity('missing')(ctx, Feature(2, 'stimulus')) # absent key: zeros Array([0., 0.], dtype=float32)