bootstrap_hindcast probabilistic bug

[aaronspring]

You are encouraged to add information about all relevant sections, but not required to submit an issue.

Code Sample, a copy-pastable example if possible

A "Minimal, Complete and Verifiable Example" will make it much easier for maintainers to help you:
http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports

hind = load_dataset('CESM-DP-SST')
hist = load_dataset('CESM-LE')
obs = load_dataset('ERSST')
kw=dict()
kw['metric']='crpss'
kw['comparison']='m2o'
s = compute_hindcast(hind,obs,**kw)  # works
bs = bootstrap_hindcast(hind,hist,obs,bootstrap=3,**kw)  # fails

Problem description

0
---------------------------------------------------------------------------
InvalidIndexError                         Traceback (most recent call last)
<ipython-input-26-71ef6f1607c5> in <module>
----> 1 bs = bootstrap_hindcast(hind,hist,obs,bootstrap=3,**kw)

~/Coding/climpred/climpred/bootstrap.py in bootstrap_hindcast(hind, hist, verif, metric, comparison, dim, sig, bootstrap, pers_sig, **metric_kwargs)
    665         compute=compute_hindcast,
    666         resample_uninit=bootstrap_uninitialized_ensemble,
--> 667         **metric_kwargs,
    668     )
    669 

~/Coding/climpred/climpred/bootstrap.py in bootstrap_compute(hind, verif, hist, metric, comparison, dim, sig, bootstrap, pers_sig, compute, resample_uninit, **metric_kwargs)
    437             add_attrs=False,
    438             dim=dim,
--> 439             **metric_kwargs,
    440         )
    441         print('shuffle_dim', shuffle_dim, '\n')

~/Coding/climpred/climpred/checks.py in wrapper(*args, **kwargs)
    144         # to the actual function call rather than showing a simple Exception
    145         # (probably IndexError from trying to subselect an empty dec_args list)
--> 146         return func(*args, **kwargs)
    147 
    148     return wrapper

~/Coding/climpred/climpred/prediction.py in compute_hindcast(hind, verif, metric, comparison, dim, max_dof, add_attrs, **metric_kwargs)
    258             )
    259         )
--> 260     skill = xr.concat(plag, 'lead')
    261     skill['lead'] = forecast.lead.values
    262     # rename back to init

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/xarray/core/concat.py in concat(objs, dim, data_vars, coords, compat, positions, fill_value, join)
    133             "objects, got %s" % type(first_obj)
    134         )
--> 135     return f(objs, dim, data_vars, coords, compat, positions, fill_value, join)
    136 
    137 

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/xarray/core/concat.py in _dataset_concat(datasets, dim, data_vars, coords, compat, positions, fill_value, join)
    301     datasets = [ds.copy() for ds in datasets]
    302     datasets = align(
--> 303         *datasets, join=join, copy=False, exclude=[dim], fill_value=fill_value
    304     )
    305 

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/xarray/core/alignment.py in align(join, copy, indexes, exclude, fill_value, *objects)
    296                 if join == "exact":
    297                     raise ValueError(f"indexes along dimension {dim!r} are not equal")
--> 298                 index = joiner(matching_indexes)
    299                 joined_indexes[dim] = index
    300             else:

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/pandas/core/indexes/base.py in __or__(self, other)
   2383 
   2384     def __or__(self, other):
-> 2385         return self.union(other)
   2386 
   2387     def __xor__(self, other):

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/pandas/core/indexes/base.py in union(self, other, sort)
   2517             return self._union_incompatible_dtypes(other, sort=sort)
   2518 
-> 2519         return self._union(other, sort=sort)
   2520 
   2521     def _union(self, other, sort):

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/pandas/core/indexes/base.py in _union(self, other, sort)
   2566                 result.extend([x for x in rvals if x not in value_set])
   2567         else:
-> 2568             indexer = self.get_indexer(other)
   2569             indexer, = (indexer == -1).nonzero()
   2570 

~/anaconda3/envs/climpred-dev/lib/python3.6/site-packages/pandas/core/indexes/base.py in get_indexer(self, target, method, limit, tolerance)
   2983         if not self.is_unique:
   2984             raise InvalidIndexError(
-> 2985                 "Reindexing only valid with uniquely" " valued Index objects"
   2986             )
   2987 

InvalidIndexError: Reindexing only valid with uniquely valued Index objects

Expected Output

no error

Output of climpred.show_versions()

# Paste the output here xr.show_versions() here

[aaronspring]

this error isn't present for dim='member'

[aaronspring]

unsure whether dim='init' should even be allowed to a probabilistic metric... maybe I need to reimplement the checks for dim from compute_ also in bootstrap_. I hoped just handling this in compute_ would be enough

[aaronspring]

I wouldnt say its a bug. users have to be careful about the resampling_dim, with PR #320 more user control.