Gumbel Copula Error

[siddharthfcb]

Environment Details

Please indicate the following details about the environment in which you found the bug:

• Copulas version: 0.5.1
• Python version: 3.8.5
• Operating System: MacOS

Error Description

Is the Gumbel copula meant to be for fitting?The following code gives an error

Steps to reproduce

from copulas.datasets import sample_bivariate_age_income
data = sample_bivariate_age_income()
data.head()
from copulas.bivariate import Gumbel
dist=Gumbel()
dist.fit(data)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-13-a3e24cdf02c2> in <module>
      4 from copulas.bivariate import Gumbel
      5 dist=Gumbel()
----> 6 dist.fit(data)

~/opt/anaconda3/lib/python3.8/site-packages/copulas/bivariate/base.py in fit(self, X)
    174             None
    175         """
--> 176         U, V = split_matrix(X)
    177         self.check_marginal(U)
    178         self.check_marginal(V)

~/opt/anaconda3/lib/python3.8/site-packages/copulas/bivariate/utils.py in split_matrix(X)
     13     """
     14     if len(X):
---> 15         return X[:, 0], X[:, 1]
     16 
     17     return np.array([]), np.array([])

~/opt/anaconda3/lib/python3.8/site-packages/pandas/core/frame.py in __getitem__(self, key)
   2900             if self.columns.nlevels > 1:
   2901                 return self._getitem_multilevel(key)
-> 2902             indexer = self.columns.get_loc(key)
   2903             if is_integer(indexer):
   2904                 indexer = [indexer]

~/opt/anaconda3/lib/python3.8/site-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
   2893             casted_key = self._maybe_cast_indexer(key)
   2894             try:
-> 2895                 return self._engine.get_loc(casted_key)
   2896             except KeyError as err:
   2897                 raise KeyError(key) from err

pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()

pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()

TypeError: '(slice(None, None, None), 0)' is an invalid key

[jmdc-mkojima]

@siddharthfcb
Hello. I've got the same problem and found an ad-hoc approach to solve it. Hope this would be helpful.

The issue seems to arise due to an invalid slice manipulation to pandas objects, i.e., X[:, 0] is not allowed when X is a pandas object. So, a quick fix is to replace the code with X.iloc[:, 0] but this is not enough because X can be a numpy object when using some other functions such as sample().

So, replacing the code in utils.py with the following could be a quick fix:
if len(X): if 'iloc' in dir(X): return X.iloc[:, 0], X.iloc[:, 1] else: return X[:, 0], X[:, 1]

[LSRathore]

@siddharthfcb Hello. I've got the same problem and found an ad-hoc approach to solve it. Hope this would be helpful.

The issue seems to arise due to an invalid slice manipulation to pandas objects, i.e., X[:, 0] is not allowed when X is a pandas object. So, a quick fix is to replace the code with X.iloc[:, 0] but this is not enough because X can be a numpy object when using some other functions such as sample().

So, replacing the code in utils.py with the following could be a quick fix: if len(X): if 'iloc' in dir(X): return X.iloc[:, 0], X.iloc[:, 1] else: return X[:, 0], X[:, 1]

It does not work for me