Pandas Accessing all user-created metadata of a pandas Series

Code Sample, a copy-pastable example if possible

import pandas as pd
import numpy as np

s = pd.Series(np.random.randn(10))
s.foo = 'bar'

print(s.__dict__)

Problem description

I heard on SO that there was some sort of metadata field. I was not able to figure out how to work with this.

On the other hand, I was able to access metadata through the __dict__ attribute. However, I got lots of other properties as well. I could write something like

def get_metadata(series):

    tmp_dict = dict()
    for key in s.__dict__:
        if key.startswith('_'):
            continue
        if key == "is_copy":
            continue
        tmp_dict[key] = s.__dict__[key]

    return tmp_dict

get_metadata(s)

But it seems like this is something that could ship with pandas out of the box.

Expected Output

This is what I'm currently getting

{'foo': 'bar', '_item_cache': {}, '_name': None, '_data': SingleBlockManager
Items: RangeIndex(start=0, stop=10, step=1)
FloatBlock: 10 dtype: float64, 'is_copy': None, '_index': RangeIndex(start=0, stop=10, step=1), '_subtyp': 'series'}

This is what I'd like to see, perhaps through another function. (metadata?)

{'foo': 'bar'}

Output of pd.show_versions()

Comment From: gfyoung

@jreback : I feel like there is an issue where this discussion is already going on with regards to metadata for pandas objects, but I don't quite remember which...

Comment From: jreback

https://github.com/pandas-dev/pandas/issues/8572