Closed aje08106 closed 3 years ago
In PR #664, I changed the behaviour. You will not find the uid of the CUDS prepended.
However, if you create two file CUDS objects that refer to different files (meaning that their contents are not identical) that have the same name, then you will find a uid in the name of the second file, like this: file (7050a16f-ab03-4d55-a933-1ab1468ec345).extension
. As I said, this will only happen when there is both a name and content conflict.
If you want to restore the old behaviour, instantiate the TransportSessionClient
with the keyword argument file_uid=True
.
Suppose the file
tensile_test_data_output.txt
belongs tomt.text_report_file
which is a sub-class ofcuba.file
, the line of codewrapper.session.load_by_oclass(mt.text_report_file)
loads the cuds object and downloads the result file to the dowload directory at once. (Connected via DataspaceSession)But the file name gets changed to something like
5e6aed06692b4d308b8e523633f06bd0-tensile_test_data_output.txt
(prepended with a hash - maybe hash of file or uuid) causing difficulty in identifying the files at client-side. Just because of that, an extra piece of code has to be written at the client side just to remove those hashes. Would be more convenient if we get file names without prepended hash in the results folder.