Skip to content

nets fit-scaler

Fits and applies scalers to myco sample archives.


usage: nets fit-scaler [-h] [-o basename] [--dstnodata value]
                       [--dtype data_type] [--max-samples count]
                       [--transform-chunksize count] [-n]
                       myco_sample_archive scaler_type

positional arguments:
  myco_sample_archive   Path to the myco sample archive to scale
  scaler_type           Scaler to fit/transform to the sample data. Choices:
                        box_cox, normal, pca, pca_normalized, robust,
                        robust_normalized, standard, uniform, yeo_johnson

optional arguments:
  -h, --help            show this help message and exit
  -o basename           Base path to output files. Filename extensions for the
                        scaler and for transformed data will be applied.
                        (default: None)
  --dstnodata value     Output nodata value (default: -99)
  --dtype data_type     The numpy data type of the scaled data (default:
                        float32)
  --max-samples count   The maximum number of samples to fit scalers with.
                        Decrease this number to use less memory. Set to 0 to
                        ignore. (default: 10000000)
  --transform-chunksize count
                        Apply transformations in chunks this size. (default:
                        10000)
  -n, --no-transform    Do not apply the fitted scaler transform to the input
                        data. (default: False)