zensvi.download.gs.GSDownloader

class zensvi.download.gs.GSDownloader(log_path=None)

Bases: zensvi.download.base.BaseDownloader

Global Streetscapes class to download the original NUS Global Streetscapes dataset.

To download the models used in the dataset, please see cv/classification/utils/global_streetscapes.py

download_svi()

Required abstract methods from parent class.

property repo_id

Property for Huggingface download.

Returns:

repo_id

Return type:

str

property repo_type

Property for Huggingface download.

Returns:

repo_type

Return type:

str

download_all_data(local_dir='data/')

Download all folders and files, recursively, from data/ This folder contains all metadata (csv) for all images of the Global Streetscapes dataset.

Parameters:

local_dir (str) – Local folder to download the data

download_manual_labels(local_dir='manual_labels/')

Download all folders and files, recursively, from manual_labels/ This folder contains all the manual labels (csv) for train and test images as well as the raw images compressed in in tar.gz.

Parameters:

local_dir (str) – Local folder to download the data

download_train(local_dir='manual_labels/train/')

Download all folders and files, recursively, from manual_labels/train/ This folder contains all the manual labels (csv) for only the train images.

Parameters:

local_dir (str) – Local folder to download the data

download_test(local_dir='manual_labels/test/')

Download all folders and files, recursively, from manual_labels/test/ This folder contains all the manual labels (csv) for only the test images.

Parameters:

local_dir (str) – Local folder to download the data

download_img_tar(local_dir='manual_labels/img/')

Download all folders and files, recursively, from manual_labels/img/ This folder contains all the raw images, compressed in tar.gz.

Parameters:

local_dir (str) – Local folder to download the data

property log_path

Property for log_path.

Returns:

Path to the log file

Return type:

str