Apply Models: Advanced

In this section, we discuss the flexibility of input format, the defaults of required information, and the use of filter information for applying models.

Input format

Like building models, applying models is also compatible with .xlsx, .csv, or DataFrame image set information input. Please refer to build models input format, and use vampire.model.apply_models()

Input file structure

The input file for applying models consists of required information in the first 4 columns and optional filter information in additional columns, if needed.

Defaults of required information

The input DataFrame img_info_df must contain, in order, the 4 required columns of

img_set_pathstr

Path to the directory containing the image set(s) used to apply model.

model_pathstr

Path to the pickle file that stores model information.

output_pathstr

Path of the directory used to output model and figures. Defaults to img_set_path.

img_set_namestr, default

Name of the image set being applied to. Defaults to time of function call.

in the first 4 columns. The default values are used in default columns when

  • the space is left blank in .csv or .xlsx file before converting to DataFrame

  • the space is None/np.NaN in the DataFrame

For examples for defaults, please refer to build models required information.

Use of filter information

Like building models, applying models has the same guidelines for filter information. Please refer to build models filter information.