Skip to content

Latest commit

 

History

History
141 lines (106 loc) · 5.04 KB

README.md

File metadata and controls

141 lines (106 loc) · 5.04 KB
Raw

idc-index

Actions Status Documentation Status

PyPI version PyPI platforms

Discourse Forum

Warning

This package is in its early development stages. Its functionality and API will change.

Stay tuned for the updates and documentation, and please share your feedback about it by opening issues in this repository, or by starting a discussion in IDC User forum.

About

idc-index is a Python package that enables basic operations for working with NCI Imaging Data Commons (IDC):

  • subsetting of the IDC data using selected metadata attributes
  • download of the files corresponding to selection
  • generation of the viewer URLs for the selected data

Getting started

Install the latest version of the package.

$ pip install --upgrade idc-index

Instantiate IDCClient, which provides the interface for main operations.

from idc_index import IDCClient

client = IDCClient.client()

You can use IDC Portal to browse collections, cases, studies and series, copy their identifiers and download the corresponding files using idc-index helper functions.

You can try this out with the rider_pilot collection, which is just 10.5 GB in size:

client.download_from_selection(collection_id="rider_pilot", downloadDir=".")

... or run queries against the "mini" index of Imaging Data Commons data, and download images that match your selection criteria! The following will select all Magnetic Resonance (MR) series, and will download the first 10.

from idc_index import index

client = index.IDCClient()

query = """
SELECT
  SeriesInstanceUID
FROM
  index
WHERE
  Modality = 'MR'
"""

selection_df = client.sql_query(query)

client.download_from_selection(
    seriesInstanceUID=list(selection_df["SeriesInstanceUID"].values[:10]),
    downloadDir=".",
)

The indices of idc-index

idc-index is named this way because it wraps indices of IDC data: tables containing the most important metadata attributes describing the files available in IDC. The main metadata index is available in the index variable (which is a pandas DataFrame) of IDCClient. Additional index tables such as the clinical_index contain non-DICOM clinical data or slide microscopy specific tables (indicated by the prefix sm) include metadata attributes specific to slide microscopy images. A description of available attributes for all indices can be found here.

Tutorial

Please check out this tutorial notebook for the introduction into using idc-index.

Resources

  • Imaging Data Commons Portal can be used to explore the content of IDC from the web browser
  • s5cmd is a highly efficient, open source, multi-platform S3 client that we use for downloading IDC data, which is hosted in public AWS and GCS buckets. Distributed on PyPI as s5cmd.
  • SlicerIDCBrowser 3D Slicer extension that relies on idc-index for search and download of IDC data

Acknowledgment

This software is maintained by the IDC team, which has been funded in whole or in part with Federal funds from the NCI, NIH, under task order no. HHSN26110071 under contract no. HHSN261201500003l.

If this package helped your research, we would appreciate if you could cite IDC paper below.

Fedorov, A., Longabaugh, W. J. R., Pot, D., Clunie, D. A., Pieper, S. D., Gibbs, D. L., Bridge, C., Herrmann, M. D., Homeyer, A., Lewis, R., Aerts, H. J. W., Krishnaswamy, D., Thiriveedhi, V. K., Ciausu, C., Schacherer, D. P., Bontempi, D., Pihl, T., Wagner, U., Farahani, K., Kim, E. & Kikinis, R. National Cancer Institute Imaging Data Commons: Toward Transparency, Reproducibility, and Scalability in Imaging Artificial Intelligence. RadioGraphics (2023). https://doi.org/10.1148/rg.230180