Flattener can attempt to call replace on a list

[amc-corey-cox]

In a recent PR for linkml-runtime. I encountered the following error in unflatten_from_csv.

        objs = []
        for row in r:
            nu_obj = {}
            for k, v in row.items():
                key_config = config.key_configs.get(k, None)
>               v = v.replace('\\n', '\n').replace('\\t', '\t')
                    ^^^^^^^^^
E               AttributeError: 'list' object has no attribute 'replace'

.venv/lib/python3.12/site-packages/json_flattener/flattener.py:355: AttributeError
----------------------------------------------------------- Captured log call ------------------------------------------------------------
WARNING  linkml_runtime.utils.csvutils:csvutils.py:37 Index slot or schema not specified
======================================================== short test summary info =========================================================
FAILED tests/test_loaders_dumpers/test_csv_tsv_loader_dumper.py::CsvAndTsvGenTestCase::test_table_model - AttributeError: 'list' object has no attribute 'replace'

It appears that our upstream processes can now pass a list object into this function where v.replace() get's called. At minimum, we should guard against the list call here but we may need to do something more intelligent if lists are passed, such as iterating through them.

[amc-corey-cox]

Oh, interesting. I was going to create a PR to solve this but it looks like it is solved in the 2.0 release. However, that 2.0 release doesn't appear to be available on PyPi. Maybe you just need to publish to PyPi and this is done?