Bit of a "This will probably never happen" issue, but we get the result groupname (in the Run class) by splitting on .py and taking the zeroth element. This obviously returns unexpected results if there is more than one .py in the filename.
Fixing it is "tricky" in that it changes behaviour that has existed for sometime. We should probably (for lyse v3.x) detect and warn if there are multiple .py in the filename and state that this will change in v4. Ideally we would also give a path to utilise the v4 fix sooner (but that seems difficult).
To be honest, we should maybe have a bigger discussion about how the groupname matching just the filename, and ignoring the folder names, is less than ideal as you have to ensure you have unique filenames across all scripts in analysislib, even if they are organised in different folders.
Bit of a "This will probably never happen" issue, but we get the result groupname (in the
Runclass) by splitting on.pyand taking the zeroth element. This obviously returns unexpected results if there is more than one .py in the filename.Fixing it is "tricky" in that it changes behaviour that has existed for sometime. We should probably (for lyse v3.x) detect and warn if there are multiple
.pyin the filename and state that this will change in v4. Ideally we would also give a path to utilise the v4 fix sooner (but that seems difficult).To be honest, we should maybe have a bigger discussion about how the groupname matching just the filename, and ignoring the folder names, is less than ideal as you have to ensure you have unique filenames across all scripts in analysislib, even if they are organised in different folders.