Bug fixes in scenicplus + change python version + generate cistopic databases from scratch

[AntoinePassemiers]

• Instead of downloading the huge cistarget databases (~50 Gb in total), now I generate them from the data itself. This is the recommended way to go, as it is more accurate. It takes extra time, but the reduction in memory usage is really worth it.
• Changed python version and base image in the Dockerfile, as the scenicplus pipeline does not work with python 3.11.8.
• Fixed some bugs related to paths, and related to the barcode naming (discrepancy between ATAC and RNA data).
• Changed the output .csv file, now they to contain the expected column names.

I managed to run the whole thing on the test resources.

[janursa]

reviwed