Address pandas warning about deprecated DataFrameGroupBy.apply default (#7235)

* Address pandas warning about deprecated DataFrameGroupBy.apply default

The current default is to pass grouping columns to the apply function,
which in our case does not use them at all.  Here we exclude them with
`include_groups=False` adopting the new behavior.

Adding temporary check that the new output is the same.

* Clean up temporary assertion (all tests passed)

Author: pavoljuhas

cirq-core/cirq/experiments/xeb_fitting.py

@@ -138,7 +138,7 @@ def _try_keep(k):
     else:
         groupby_names = ['cycle_depth']
 
-    return df.groupby(groupby_names).apply(per_cycle_depth).reset_index()
+    return df.groupby(groupby_names).apply(per_cycle_depth, include_groups=False).reset_index()
 
 
 class XEBCharacterizationOptions(ABC):
@@ -702,7 +702,7 @@ def _per_pair(f1):
         groupby = ['layer_i', 'pair_i', 'pair']
     else:
         groupby = ['pair']
-    return fidelities_df.groupby(groupby).apply(_per_pair)
+    return fidelities_df.groupby(groupby).apply(_per_pair, include_groups=False)
 
 
 def before_and_after_characterization(