allow skipping normalization when compute_cell_density

[mgeplf]

No description provided.

[lecriste]

LGTM.
For reference:
https://bbpteam.epfl.ch/project/issues/browse/BBPP134-1342?focusedId=233502&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-233502

[mgeplf]

@lecriste and @Sebastien-PILUSO can you try this, and see if it solves the problem you were running into?

[drodarie]

I have a counter proposal for this pull request. The issue happened because the outside of the Nissl used is set to 0. Why not solving the issue in the function normalize_intensity in utils.py?
I can provide a simple implementation.

[Sebastien-PILUSO]

@drodarie why would you like to still apply the normalize_intensity function when all the processings applied in it are already applied to the input volume?

[drodarie]

Then the function will not do anything and that's fine. This way, we don't have an extra user option which would make the pipeline even more difficult to use.

[drodarie]

See my proposal here

[mgeplf]

See my proposal here

If that's what you want, that works for me. I very much prefer fewer options that modify the behavior (ie: I agree with what you on we don't have an extra user option...)

[Sebastien-PILUSO]

Ok let's try

[Sebastien-PILUSO]

I ran the code using @drodarie's branch and this gave the same cell density file than on the main branch. So this is okay for me.

[mgeplf]

Done better in #74