Allow kagglehub.dataset_download to download entire dataset to a specified directory

[Montekkundan]

When using kagglehub.dataset_download, the path parameter appears to be designed for downloading a specific file from the dataset rather than defining a destination directory on the user's machine. This creates confusion and limits functionality, especially when a user want to download the entire dataset directly to their current working directory or a specified folder.

For example, the following code:

import kagglehub

path = kagglehub.dataset_download("preritbhagat/stress-non-stress-images", path="./data")

print("Path to dataset files:", path)

this results in a 404 error

i tried with

import kagglehub

# Download latest version
path = kagglehub.dataset_download("preritbhagat/stress-non-stress-images" , path="FINAL_TFEID/FINALTFEID_NONSTRESS/f01_dfh_hx.jpg")

print("Path to dataset files:", path)

this downloads the single image, as documented in the docs of this function; where FINAL_TFEID/FINALTFEID_NONSTRESS/f01_dfh_hx.jpg is the path of the file in that dataset.

• Is there currently a way to download an entire dataset to a specific directory (e.g., ./data or the current working directory) using kagglehub.dataset_download?
• If not, would it be possible to enhance the function to include a parameter (e.g., destination) for specifying a custom download location for the entire dataset?

[andreapisa9]

I agree with @Montekkundan for what concerns the confusing naming choice. @Montekkundan , you can define the download destination by changing the KAGGLEHUB_CACHE environment variable:

export KAGGLEHUB_CACHE=/path/to/your/preferred/directory

[calderjo]

duplicate: #175