Investigate limited access items that appear to have CC-BY license

[fedorov]

My expectation is that there should be no Limited access items that are distributed under CC-BY license, the below is not empty, and seems to be wrong.

[s-paquette]

SELECT collection_id, license_short_name
FROM `idc-dev-etl.idc_v8_pub.dicom_all`
WHERE access='Limited'
GROUP BY collection_id, license_short_name

[bcli4d]

These instances are from the GBM-MR-NER-Outcomes analysis results collection, which has a CC BY 3.0 license. However, the instance data is mixed in with TCGA-GBM data and is therefore not publicly available.

[bcli4d]

There are 147 series in TCGA-GBM with GBM-MR-NER-Outcomes' source_doi: SELECT distinct tcia_api_collection_id, SeriesInstanceUID FROMidc-dev-etl.idc_v8_pub.auxiliary_metadatawhere source_doi = '10.7937/K9/TCIA.2014.FAB7YRPZ'

[bcli4d]

@fedorov As requested, ETL Workflow, v9 describes how DOIs are assigned to series, specifically, in sections 3.2.3.2 and 3.2.3.5. The corresponding code is here and here respectively.

[fedorov]

ticket submitted to TCIA https://help.cancerimagingarchive.net/servicedesk/customer/portal/1/TH-49543

[bcli4d]

@andrey, I seem to recall that you wanted to investigate this issue further before we proceed to restore the DOI of these 147 series to that of the analysis result? (I previously changed the DOI to that of TCGA-GBM)