ENH: Adding option to skip randomization of seed. + other fixes

[EmmaRenauld]

• Adding an option to the _dev tracking to have fixed seeds. This allows to create a figure like this:

• Changing the --discard_last_point option to make it the default choice (as discussed with Gab and other; this makes it more like Dipy)

• Fixing the way the numpy generator and seed are used, as per the numpy doc. If you think it's not the best way to do it, I can just delete the last commit.

[pep8speaks]

Hello @EmmaRenauld, Thank you for updating !

• In the file scilpy/tracking/propagator.py:

Line 70:80: E501 line too long (80 > 79 characters)
Line 496:80: E501 line too long (80 > 79 characters)

• In the file scilpy/tracking/tests/test_seed.py:

Line 45:1: W391 blank line at end of file

Comment last updated at 2023-11-07 18:25:28 UTC

[arnaudbore]

Build passed ! Good Job 🍻 !

[CHrlS98]

I tested and it seems to work properly.

However, I think this would be more useful if, instead of disabling jittering the seeding positions, we added a --nb_repeats arguments telling how many times we want to track from each seed. The behaviour you want to show could still happen, but it would generalize better and might find other applications than only for teaching purposes?

[CHrlS98]

any way we can shorten this param name?

[EmmaRenauld]

Changed for no_seed_displacement. Ok?

[EmmaRenauld]

However, I think this would be more useful if, instead of disabling jittering the seeding positions, we added a --nb_repeats arguments telling how many times we want to track from each seed. The behaviour you want to show could still happen, but it would generalize better and might find other applications than only for teaching purposes?

That is nice, but right now it seems complicated for me. We would have to remember how many times we called get_next_pos.

[arnaudbore]

Build passed ! Good Job 🍻 !

[CHrlS98]

However, I think this would be more useful if, instead of disabling jittering the seeding positions, we added a --nb_repeats arguments telling how many times we want to track from each seed. The behaviour you want to show could still happen, but it would generalize better and might find other applications than only for teaching purposes?

That is nice, but right now it seems complicated for me. We would have to remember how many times we called get_next_pos.

Wouldn't we just need to return the same seed n_repeat times ? This just changes the start position of a streamline, no?

[arnaudbore]

Build passed ! Good Job 🍻 !

[EmmaRenauld]

Should be good!

[arnaudbore]

Build passed ! Good Job 🍻 !

[CHrlS98]

I would like to validate that the note.txt file in tests is at the right place, but otherwise, everything looks good to me.

[arnaudbore]

Build passed ! Good Job 🍻 !