Case of replicate column names differ among years

[rluedde]

>>> from cenpy.moe import replicate_table_utils as crtu
>>> data_14 = crtu.get_replicate_data_api(["B15002"], 2014, "140", "04")
https://www2.census.gov/programs-surveys/acs/replicate_estimates/2014/data/5-year/140/B15002_04.csv.gz
>>> data_18.columns.levels[0][:3]
Index(['ESTIMATE', 'MOE', 'SE'], dtype='object', name='categories')
>>> data_14.columns.levels[0][:3]
Index(['estimate', 'moe', 'SE'], dtype='object', name='categories')

In the 2014 data, the names of these columns are capitalized and all other years (I think) are lowercase. This causes issues in (at least) apply_func.

I don't have good ideas for a solution. I imagine it's a .upper() or .lower() call somewhere in read_replicate_file or get_replicate_data. A comprehension might be needed as well?

[dfolch]

Good idea to always convert to the same case. It is probably not worth the extra code to test for capitalization, just run upper on these three column names for all files read in with read_replicate_file.