HRA stressor comparison bug on sample data

[dcdenu4]

A user on the forum noted they were getting an error when running HRA with sample data in 3.14.0.

https://community.naturalcapitalproject.org/t/error-in-hra-with-dawnloaded-sample-data/4339

It looks like when comparing stressors one table is lowercasing while the other is not. For the habitat_stressor_info table validation.get_validated_dataframe is being used which does return a table with lowercased values. For the criteria table, utils.read_csv_to_dataframe is returning a table where values are NOT lowercased.

Note: validation.get_validated_dataframe calls utils.read_csv_to_dataframe, so this is likely an unique case handling the HRA criteria table.

I can reproduce the issue with the latest invest:main

[dcdenu4]

Interestingly, all tests pass so we must not be running a regression test on sample data. And looking at the HRA tests, all the unit tests use lowercase values for table creations.

[davemfish]

Interestingly, all tests pass so we must not be running a regression test on sample data. And looking at the HRA tests, all the unit tests use lowercase values for table creations.

see also #1468

[dcdenu4]

So, I think that reading the table with validation.get_validated_dataframe is the issue here. That function does some additional mutation to the dataframe that helps when validating. But it causes consequences when using in the model directly. In this case, the function lowercases the values of columns which utils.read_csv_to_dataframe does not. So let's use the same utils function for reading when we're comparing.