Addition of the function cleaning_cluster_population

[ChiaraLionello]

I have created the function the works for 2D and 3D labels' arrays. It cleans the clusters with low population.
Solved #136

[matteobecchi]

These are the labels, not the data... aren't they always a 2D array? Each atom at each frame has one integer label, even if the data were multivariate.

[matteobecchi]

For the same reason, I think all the logic in the function for the 3D case is never necessary.

[ChiaraLionello]

I think it depends how we do want to use this function. I've thought different applications:

• 2D labels: cleaning the labels to color the trajectory
• 2D labels: can come also from other clustering methods, not necessarily from onion, and can be used as above
• 3D labels: come from get_onion_analysis and can be use to redo the plot with the number of clusters and the unclassified fraction cleaned. By keeping this function out from get_onion_analysis you are free to change the threshold how many times you want without running the Onion calculation multiple times. In this case, I should also think a way to do that plot again. On the other hand, it's true that's enough to keep only the 2D option and then run it in a for cicle when we need it.

There's no problem for me to delete it in case; since it wasn't difficult to do I've added both the options.

[matteobecchi]

Same as before, labels are always 2D.

[SimoneMartino98]

Just two easy comments. I’ll leave the discussion on the label dimensions to @matteobecchi.

Anyway, nice work!

[SimoneMartino98]

This is not actually a log, rather it should be a warning. It is also true that at the moment we don't have a warning log function. My suggestion here is to change the line with:

logger.warning(f"Excluded value(s) not found in labels: {missing}")

and then add something like the following function in the _internal.logs.py file (after the log function):

  def warning(self, msg: str) -> None:
      """Records an informational warning message to the log.

      Parameters:
          msg:
              The message to record.
      """
      timestamp = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S")
      history_entry = f"[{timestamp}] {msg}"
      console.warning(msg)
      self._log.append(history_entry)

It should work and give a yellow warning msg. If not let's see what happen.

[SimoneMartino98]

I would remove this; we are already in the test folder. Up to you though.

[SimoneMartino98]

i'm not sure; but i think that for this case the arrays should be exactly equal to each other (we are not using float here, they are all int, i'm not seeing any machine dependencies).

Maybe is the same thing, but i think that here we can be more restrictive with:

https://numpy.org/devdocs/reference/generated/numpy.array_equal.html

instead of np.allclose