Bug with `doi` field when using `.get_timeseries`

[khider]

When using get_timeseries to expand properties into a dictionary for each column, there is a duplicate 'doi' for publication. One is called pub1_doi and seems to contain the actual doi. The other is called pub1_DOI and seems the be stuck at "hello".

To reproduce:

from pylipd.lipd import LiPD
url1 = 'https://lipdverse.org/data/iso2k100_CO04LIRA/1_0_2//CO04LIRA.lpd'
url2 = 'https://lipdverse.org/data/ch2kTA18TAS01/1_0_2//TA18TAS01.lpd'

data_url = [url1,url2]

D1 = LiPD()
D1.load(data_url)

ts = D1.get_timeseries(D1.get_all_dataset_names())

[varunratnakar]

Was an error in the legacy code. Fixed

[khider]

I still het two DOI's field. And now the DOI is just an empty list instead of Hello.

This is the first example in this notebook