Add StudyInstanceUID to Exam schema + re-ingest data with repeated instanceNumbers

[aiavenak]

This fixes https://github.com/IUSCA/rady-qc/issues/39

Some data appears as having two images with repeated InstanceNumbers. This will be solved by using the StudyInstanceUID as an exam's identifier, instead of using the StudyTimestamp, as we have been doing so far.

If we want this change to be applied to the entire dataset, we will have to re-ingest all data. I am not sure that this is necessary, but, it is not trivial to single out all the cases where there are multiple StudyInstanceUID with identical StudyTimestamp.

Regardless, we need to first standardize the header files (i.e. tarballs vs .json files) before we can start re-ingesting any data at all

[agopu]

Somewhat dependent on https://github.com/IUSCA/rady-qc/issues/51 (last sentence's in Andrea's ticket summary above)

[agopu]

Discussed in video call on 2019-04-02

• Sundar thinks John needs to clarify this with Karmen.

• Per Andrea, it seems using StudyInstanceUID will cause fracturing of an exam for a single subject on the PET/CT side.

• Sundar thinks that different reconstruction schemes are separate enough that this is probably OK.

• @jdwest0711 offers to confer with Karmen but this may not be necessary (see below).

• @aiavenak states

• [x] We have to make a code change to handle this change

• [ ] We have to then reingest everything from our metadata store (i.e tarballs).

@jdwest0711 points out that he and @mparamas will likely resend PET/CT data anyway? That is a separate ticket as is the ability to look at [Orthanc reception status on the SCA side] (https://github.com/IUSCA/rady-qc/issues/68).

[agopu]

• [ ] @youngmd states we may not have to reingest everything, he will talk with @aiavenak about this.