Skip to content

DOWNLOADER.md

Latest commit

 

History

History
59 lines (49 loc) · 2.96 KB

DOWNLOADER.md

File metadata and controls

59 lines (49 loc) · 2.96 KB
Raw

Currently supported downloaders:

  • MIMIC CXR (MIMIC_CXR in src.datasets.chest_xray.mimic_cxr)
  • VINDR CXR (VINDR_CXR in src.datasets.chest_xray.vindr_cxr)
  • VINDR MAMMO (VINDR in src.datasets.mammo.vindr)
  • BUSI (BUSI in src.datasets.ultrasound.busi)
  • COVID19 (Covid19Dataset in src.datasets.chest_xray.covid19)
  • Brain Tumor (BrainTumorDataset in src.datasets.mri.brain_tumor)
  • Brain Tumor 2 (BinaryBrainTumorDataset in src.datasets.mri.brain_tumor_2)
  • CoronaHack (CoronaHackDataset in src.datasets.chest_xray.coronahack)
  • COVID BLUES (COVIDBLUES in src.datasets.ultrasound.covid_blue)
  • COVID US (COVIDUS in src.datasets.ultrasound.covid_us)

How to download:

Global Downloader (Recommended)

Use the download.py script to download multiple datasets at once:

# Download all supported datasets
python download.py /path/to/datasets

# Download specific datasets
python download.py /path/to/datasets --datasets mimic_cxr vindr_cxr busi

# Continue downloading even if some datasets fail
python download.py /path/to/datasets --continue-on-error

# Show help and list of available datasets
python download.py --help

The global downloader will download datasets to organized subdirectories (e.g., /path/to/datasets/chest_xray/mimic_cxr)

Downloading individual datasets

Create an object of the corresponding dataset class. An example is shown below.

d = BinaryBrainTumorDataset(base_root='data', download=True)

If we run this code from the root directory, the dataset will be downloaded to the 'data' directory. Each dataset will be in its own directory, and the dataset directories are categorized by modality. For example, Brain Tumor 2 will be downloaded to data/mri/brain_tumor_2.

Downloading select files or directories

We support downloading select files from large datasets (MIMIC CXR, VINDR CXR, VINDR MAMMO). In order to download select files, use the file_list parameter to pass in a list of directory/files (strings) to download. Directories should only include the directory name and not the whole path. Files should only include the file name without the extension (also no path). An example is shown below.

d = VINDR(download=True, base_root='data', file_list=['0025a5dc99fd5c742026f0b2b030d3e9',
                                                '16e58fc1d65fa7587247e6224ee96527',
                                                '7fc1f1bb8bb1a7efaf7104e49c4d8b86'])

'0025a5dc99fd5c742026f0b2b030d3e9' specifies everything inside the 0025a5dc99fd5c742026f0b2b030d3e9 directory. '16e58fc1d65fa7587247e6224ee96527' specifies everything inside the 16e58fc1d65fa7587247e6224ee96527 directory. '7fc1f1bb8bb1a7efaf7104e49c4d8b86' specifies the 7fc1f1bb8bb1a7efaf7104e49c4d8b86.dicom file.

The downloader will not download a file multiple times, and will not download a file if it already exists.

Note that certain required files (e.g. metadata and licenses) will be downloaded (if they don't already exist) even if not specified by the file_list parameter.