Fix fitting and optimization steps

[drodarie]

This pull request fixes 3 different issues in the fitting and inhibitory optimization steps:

• During the fitting step, the slices indexes provided by the user are supposed correct (global positions as defined by the AIBS). However, the annotation volumes and ISH marker datasets can have an offset. It is now taken into account when loading the marker datasets.
• The fitting should be applied only on the group regions selected. The other regions used to be fitted according to the Rest group. These regions are now ignored during fitting but will obtain and estimate during the optimization step.
• An unnecessary test was preventing leaf regions to have an uncertain first estimate from the fitting when their parent region is given for certain. This should not prevent the optimization as the children region will see their estimates corrected according to the parent region. This test was preventing a case scenario where the user set a non leaf region to a certain density value (with certainty) but not its children.

[drodarie]

For some reason, the tests for the mtype step are now failing: test_mtype_densities.py However, I do not see how any of my changes impacted these tests.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

:exclamation: No coverage uploaded for pull request base (main@2b89cfd). Click here to learn what that means.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #75 +/- ## ======================================= Coverage ? 97.72% ======================================= Files ? 22 Lines ? 1451 Branches ? 0 ======================================= Hits ? 1418 Misses ? 33 Partials ? 0 ``` | [Flag](https://app.codecov.io/gh/BlueBrain/atlas-densities/pull/75/flags?src=pr&el=flags&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=BlueBrain) | Coverage Δ | | |---|---|---| | [pytest](https://app.codecov.io/gh/BlueBrain/atlas-densities/pull/75/flags?src=pr&el=flag&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=BlueBrain) | `97.72% <100.00%> (?)` | | Flags with carried forward coverage won't be shown. [Click here](https://docs.codecov.io/docs/carryforward-flags?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=BlueBrain#carryforward-flags-in-the-pull-request-comment) to find out more.

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[mgeplf]

However, I do not see how any of my changes impacted these tests.

I fixed that in https://github.com/BlueBrain/atlas-densities/pull/76

[mgeplf]

Code looks good, I have to assume it's correct from the scientific standpoint. Reportedly, this has been used for the latest atlas release in staging (https://bbpteam.epfl.ch/project/issues/browse/BBPP134-1342?page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel&focusedCommentId=239469#comment-239469)