download()

Downloads the specified benchmark dataset to local storage.

Syntax

def download() -> None

Description

The download() method downloads the specified benchmark dataset from AWS S3 bucket and saves it to the local benchmark/ directory. The download process includes integrity verification.

Returns

No return value. The dataset is saved in the benchmark/ directory after download.

Basic Example

from petsard.loader.benchmarker import BenchmarkerConfig, BenchmarkerRequests

# Create configuration
config = BenchmarkerConfig(
    benchmark_name="adult-income",
    filepath_raw="benchmark://adult-income"
)

# Create downloader and execute download
downloader = BenchmarkerRequests(config.get_benchmarker_config())
downloader.download()

Download Process

1. Check local cache: Check if the dataset already exists in benchmark/ directory
2. Create directory: Create benchmark/ directory if it doesn’t exist
3. Download data: Download dataset from AWS S3
4. Verify integrity: Verify downloaded data using SHA256
5. Save file: Save data to local storage

Error Handling

from petsard.exceptions import BenchmarkDatasetsError

try:
    downloader.download()
except BenchmarkDatasetsError as e:
    print(f"Download failed: {e}")

Possible error causes:

• Network connection failure
• Dataset does not exist
• SHA256 verification failure
• Insufficient disk space

Notes

• First download requires network connection
• Datasets are cached locally, repeated downloads are skipped
• Large dataset downloads may take considerable time
• Recommended to use indirectly through LoaderAdapter rather than direct calls