Dependent GreaterThan constraints doesn't work

[fealho]

As mentioned in #366, two dependent GreaterThan constraints currently doesn't work. Even though this is not exactly supported, it should still work, indicating a possible bug in the code.

Below is an example where the synthetic data for low should be smaller than for n, and n smaller than high, but it doesn`t work:

import pandas as pd
from sdv.tabular import CTGAN
from sdv.constraints import GreaterThan

data = pd.DataFrame({
    'id': str(i) for i in range(300),
    'low': 1.0, 2.0, 3.0100,
    'n': 1.0, 2.0, 3.0100,
    'high': 1.0, 2.0, 3.0*100
})

low_constraint = GreaterThan(low='low',high='n',handling_strategy='transform')
high_constraint = GreaterThan(low='n',high='high',handling_strategy='transform')
constraints = high_constraint, low_constraint

model = CTGAN(epochs=50, primary_key='id', constraints=constraints)
model.fit(data)
model.sample(100)

[fealho]

I think the reason for this error is here: https://github.com/sdv-dev/SDV/blob/master/sdv/metadata/table.py#L554.

Instead, since we are trying to revert the transforms, we should iterate over the constraints in the reverse order.

for constraint in reversed(self._constraints):

[csala]

This was covered by the Between constraint introduced in #367