get_input_range

behavenet.plotting.cond_ae_utils.get_input_range(input_type, hparams, sess_ids=None, sess_idx=0, model=None, data_gen=None, version=0, min_p=5, max_p=95, apply_label_masks=False)[source]

Helper function to compute input range for a variety of data types.

Parameters:

  • input_type (str) – ‘latents’ | ‘labels’ | ‘labels_sc’

  • hparams (dict) – needs to contain enough information to specify an autoencoder

  • sess_ids (list, optional) – each entry is a session dict with keys ‘lab’, ‘expt’, ‘animal’, ‘session’; for loading labels and labels_sc

  • sess_idx (int, optional) – session index into data generator

  • model (AE object, optional) – for generating latents if latent file does not exist

  • data_gen (ConcatSessionGenerator object, optional) – for generating latents if latent file does not exist

  • version (int, optional) – specify AE version for loading latents

  • min_p (int, optional) – defines lower end of range; percentile in [0, 100]

  • max_p (int, optional) – defines upper end of range; percentile in [0, 100]

  • apply_label_masks (bool, optional) – True to set masked values to NaN in labels

Returns:

keys are ‘min’ and ‘max’

Return type:

dict