BUG - MAE method doesn't like pd.Series when using dimension preservation about scores

[tennlee]

Here is a snippet to reproduce this issue:

import pandas as pd
import scores

xda_fcst = scores.sample_data.simple_forecast()
xda_obs = scores.sample_data.simple_observations()
print(scores.continuous.mae(xda_fcst, xda_obs, preserve_dims='all'))

fcst = pd.Series(scores.sample_data.simple_forecast())
obs = pd.Series(scores.sample_data.simple_observations())
print(scores.continuous.mae(fcst, obs))

print(scores.continuous.mae(fcst, obs, preserve_dims='all'))

[tennlee]

Support for this will likely need to be pushed to further down the roadmap for a variety of reasons, mostly to do with complexity of supporting heterogenous data types in an elegant way. There are options which can be considered, but none of them have yet emerged as a clearly appropriate way forward. Pandas support remains on the roadmap for the medium term, but it is more urgent to complete the current tranche of work associated with implementation of required scores and preparation for particular use cases.

[tennlee]

There is now an explicit scores.pandas API which should be used and is supported. This should resolve this use case, please take a look and see if it meets your needs.