nuclick transforms: min_area and do_reconstruction are documented but never used

Describe the bug

Two documented parameters in monai/apps/nuclick/transforms.py are accepted and stored but never used, so setting them has no effect.

1. SplitLabeld.min_area

Documented as "min_area: The smallest allowable object size.", assigned in __init__, and then never referenced again — __call__ uses mask_value, others_value and to_binary_mask, but nothing filters by area.

$ grep -rn "\.min_area\b" monai/
monai/apps/nuclick/transforms.py:157:        self.min_area = min_area

That single line is the only occurrence in the package.

2. PostFilterLabeld.do_reconstruction

Documented as "do_reconstruction: Boolean Flag, Perform a morphological reconstruction of an image...", assigned in __init__, never read. __call__ runs

masks = self.post_processing(label, self.thresh, self.min_size, self.min_hole)

and post_processing does not accept the flag at all:

def post_processing(self, preds, thresh=0.33, min_size=10, min_hole=30):
    masks = preds > thresh
    for i in range(preds.shape[0]):
        masks[i] = morphology.remove_small_objects(masks[i], min_size=min_size)
        masks[i] = morphology.remove_small_holes(masks[i], area_threshold=min_hole)
    return masks

so there is no morphological-reconstruction step anywhere in the class. nuc_points is likewise stored and unused, which is consistent with the reconstruction path (which is what would need the click points) never having been wired up.

To Reproduce

grep -rn "\.min_area\b" monai/          # 1 hit: the assignment
grep -rn "\.do_reconstruction\b" monai/ # 1 hit: the assignment

Expected behavior

Either the parameters take effect, or they are not offered.

Why I am filing this rather than sending a PR

Both plausible fixes involve a judgement call I would rather not make for you:

• implement them — filtering by min_area, and adding a reconstruction branch using nuc_points — is writing new behaviour, and I do not know what the NuClick reference implementation expects here;
• remove them — accurate, but they are public keyword arguments on two transforms, so it is a breaking change that probably wants a deprecation cycle.

Happy to send a PR for whichever direction you prefer.

Environment

Verified against the current dev branch; the parameters are present and unused there.