Reference utils

This file describes methods associated with dataset analysis and loading.

Analysis

compute_span(df, col_label='identity')

Compute the time span of the dataset.

The span is defined as the latest time minus the earliest time of image taken. The times are computed separately for each individual.

Parameters:

Name	Type	Description	Default
df	DataFrame	A full dataframe of the data.	required
col_label	str	Column name containing individual animal names (labels).	'identity'

Returns:

Type	Description
float	The span of the dataset in seconds.

display_statistics(df, unknown_name='', col_label='identity')

Prints statistics about the dataframe.

Parameters:

Name	Type	Description	Default
df	DataFrame	A full dataframe of the data.	required
unknown_name	str	Name of the unknown class.	''
col_label	str	Column name containing individual animal names (labels).	'identity'

Loading

get_dataframe_path(root_dataframe, class_dataset)

Creates path to the pickled dataframe.

Parameters:

Name	Type	Description	Default
root_dataframe	str	Path where all dataframes are stored.	required
class_dataset	type	Type of WildlifeDataset.	required

Returns:

Type	Description
str	Path to the dataframe.

get_dataset_folder(root_dataset, class_dataset)

Creates path to the dataset data.

Parameters:

Name	Type	Description	Default
root_dataset	str	Path where all datasets are stored.	required
class_dataset	type	Type of WildlifeDataset.	required

Returns:

Type	Description
str	Path to the stored data.

load_dataset(class_dataset, root_dataset, root_dataframe, overwrite=False, **kwargs)

Loads dataset from a pickled dataframe or creates it.

If the dataframe is already saved in a pkl file, it loads it. Otherwise, it creates the dataframe and saves it in a pkl file.

Parameters:

Name	Type	Description	Default
class_dataset	type	Type of WildlifeDataset to load.	required
root_dataset	str	Path where all datasets are stored.	required
root_dataframe	str	Path where all dataframes are stored.	required
overwrite	bool	Whether the pickled dataframe should be overwritten.	False

Returns:

Type	Description
WildlifeDataset	The loaded dataset.

load_datasets(class_datasets, root_dataset, root_dataframe, **kwargs)

Loads multiple datasets as described in load_dataset.

Parameters:

Name	Type	Description	Default
class_datasets	List[type]	List of types of WildlifeDataset to download.	required
root_dataset	str	Path where all datasets are stored.	required
root_dataframe	str	Path where all dataframes are stored.	required

Returns:

Type	Description
list[WildlifeDataset]	The list of loaded datasets.