pandas cat identifiers do not display correctly when referenced using dict-like syntax

[PhilipYip1988]

Some identifiers do not display correctly from a dict-like reference such as a value in a dict or a Series from a DataFrame:

import random
import string
import numpy as np
import pandas as pd
random.seed(0)

df1 = pd.DataFrame({'grades1': random.choices(['a', 'b', 'c', 'd', 'e'], k=30),
                    'grades2': random.choices(['a', 'b', 'c', 'd', 'e'], k=30)})

df1['grades1'] = df1['grades1'].astype('category')
df1['grades2'] = df1['grades2'].astype('category')

For example using:

df1['grades1'].cat.<tab>

Should give consistent behaviour to accessing the identifier via the attribute:

df1.grades1.cat.<tab>

I originally posted this as ipython issue 14819 but was told to post here as it seems to be an issue with jedi and the identifiers display correctly when:

%config Completer.use_jedi = False
import random
import string
import numpy as np
import pandas as pd
random.seed(0)

df1 = pd.DataFrame({'grades1': random.choices(['a', 'b', 'c', 'd', 'e'], k=30),
                    'grades2': random.choices(['a', 'b', 'c', 'd', 'e'], k=30)})

df1['grades1'] = df1['grades1'].astype('category')
df1['grades2'] = df1['grades2'].astype('category')