Strange compatibility issue with pandas

Hello, I am finding a strange issue where pandas breaks after I import some features from econml. For instance, if I want to inspect the label column:
df['label' works fine with no econml imports. But it throws an exception after importing:
from econml.dml import LinearDML, CausalForestDML

Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/lib/python3.8/site-packages/sympy/interactive/printing.py", line 27, in _displayhook print(stringify_func(arg, **settings)) File "/usr/local/lib/python3.8/site-packages/sympy/printing/printer.py", line 372, in __call__ return self.__wrapped__(*args, **kwargs) File "/usr/local/lib/python3.8/site-packages/sympy/printing/pretty/pretty.py", line 2871, in pretty return pp.doprint(expr) File "/usr/local/lib/python3.8/site-packages/sympy/printing/pretty/pretty.py", line 66, in doprint return self._print(expr).render(**self._settings) File "/usr/local/lib/python3.8/site-packages/sympy/printing/printer.py", line 331, in _print return printmethod(expr, **kwargs) File "/usr/local/lib/python3.8/site-packages/sympy/printing/pretty/pretty.py", line 999, in _print_Series args = list(expr.args) File "/usr/local/lib/python3.8/site-packages/pandas/core/generic.py", line 5487, in __getattr__ return object.__getattribute__(self, name) AttributeError: 'Series' object has no attribute 'args'

What am I missing here?

Reproduce the report with Python 3.8 by importing pandas, then LinearDML and CausalForestDML from econml.dml, and inspecting the label column. Start with the supplied traceback, particularly the interaction with SymPy's display hook and pandas Series. Done means the reported column inspection no longer raises this exception after the econml imports, with the behavior covered by a regression test.