Wrong date specified for data directory
Summary
Wrong date specified for data directory.
This was first reported on 2021-11-02: https://git.ecdf.ed.ac.uk/swain-lab/python-pipeline/aliby_wpass/-/issues/44
This otherwise does not affect the pipeline or data analysis in any way, as long as the user specifies the name of the directory accordingly. My only concern is that it may cause confusion.
Steps to reproduce
Run pipeline on an experiment based on Omero ID, e.g. via https://git.ecdf.ed.ac.uk/swain-lab/aliby/skeletons/-/blob/master/notebooks/2.%20Run%20experiment.ipynb
What is the current bug behavior?
Name of directory includes the date the Omero object is created, rather than the date of the experiment, e.g. 27993_2022_06_26_flavin_rim11_cenpkkoetter_10mgpL_02_00
rather than 27993_2022_06_24_flavin_rim11_cenpkkoetter_10mgpL_02_00
. There usually is a difference between the two dates because (a) experiments running overnight, (b) uploads taking time, (c) (un)scheduled Windows updates rebooting the Omero upload PC -- therefore stopping uploads and requiring manual intervention to restart them.
However, the metadata of the HDF5 file returns the date of the experiment, as it should. Specifically, the date
attribute returns a Unix time stamp (float) that refers to 00:00 of the local time zone of the date of the experiment -- in this case, 1656025200.0
What is the expected correct behavior?
Name of directory should include the date of the experiment.
Possible fixes
I suspect it is because the routine that determines the name of the directory takes the value from the time the Omero object was created. As the pipeline does not 'know' the date of the experiment at this point -- not having gone through the Acq file that has this information -- it's probably the best it can do. Perhaps after this metadata has been accessed, the pipeline can rename the directory.
(TODO: Find the line of code that is responsible.)