Investigate fall-back to Toshiba extractor if no RDSR
Issue #700 closed
With reference to pull request #248
Need to look at whether we can make use of the legacy Toshiba extractor automatically. Need to check how this might work if Orthanc receives RDSR, images and dose images - does this create duplicate entries, for example.
A related issue I've come across whilst improving the logging for QR is that if there is no RDSR for a CT study, if there is a series with a series description
_get_toshiba_dose_imageswill never be called, effectively meaning you couldn't use the routine with a Philips scanner.
This might not be an issue, as the Philips dose images have most of the information we want, but would have to be changed if you did want to use it with a Philips system, or if there was a system without RDSR but with a dose image with that series description that wasn't Philips!
One of the reasons that the legacy Toshiba CT extractor is configured to only operate on specific systems is that it is not known to provide correct results for a wide-range of systems. I know it works OK for the systems that I use it for, but I don't know that it will work correctly for other systems. I therefore think that it is a bad idea to have the Orthanc Lua script automatically fall-back to legacy Toshiba extractor if the DICOM object is not an RDSR. I think that the use of the legacy Toshiba extractor should require that the end user specifically configures it to be enabled for their system, and that the user validates the results (as noted in the Orthanc configuration document: https://docs.openrem.org/en/latest/netdicom-orthanc-config.html?highlight=orthanc#orthanc-store-configuration).
Closing due to the lack of response to @David Platten's question. Please feel free to reopen if required!