geography_as_object=True fails if mode="REPEATED"

[bnaul]

Small oversight I think from #848:

[ins] In [57]: bigquery.Client().query("SELECT [ST_GEOGPOINT(-100, 30)] x").to_dataframe()
Out[57]:
                  x
0  [POINT(-100 30)]

[ins] In [58]: bigquery.Client().query("SELECT [ST_GEOGPOINT(-100, 30)] x").to_dataframe(geography_as_object=T
          ...: rue)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-58-90093b9d3663> in <module>
----> 1 bigquery.Client().query("SELECT [ST_GEOGPOINT(-100, 30)] x").to_dataframe(geography_as_object=True)

~/model/.venv/lib/python3.9/site-packages/google/cloud/bigquery/job/query.py in to_dataframe(self, bqstorage_client, dtypes, progress_bar_type, create_bqstorage_client, max_results, geography_as_object)
   1682         """
   1683         query_result = wait_for_query(self, progress_bar_type, max_results=max_results)
-> 1684         return query_result.to_dataframe(
   1685             bqstorage_client=bqstorage_client,
   1686             dtypes=dtypes,

~/model/.venv/lib/python3.9/site-packages/google/cloud/bigquery/table.py in to_dataframe(self, bqstorage_client, dtypes, progress_bar_type, create_bqstorage_client, geography_as_object)
   1984             for field in self.schema:
   1985                 if field.field_type.upper() == "GEOGRAPHY":
-> 1986                     df[field.name] = df[field.name].dropna().apply(_read_wkt)
   1987
   1988         return df

~/model/.venv/lib/python3.9/site-packages/pandas/core/series.py in apply(self, func, convert_dtype, args, **kwargs)
   4431         dtype: float64
   4432         """
-> 4433         return SeriesApply(self, func, convert_dtype, args, kwargs).apply()
   4434
   4435     def _reduce(

~/model/.venv/lib/python3.9/site-packages/pandas/core/apply.py in apply(self)
   1080             return self.apply_str()
   1081
-> 1082         return self.apply_standard()
   1083
   1084     def agg(self):

~/model/.venv/lib/python3.9/site-packages/pandas/core/apply.py in apply_standard(self)
   1135                 # List[Union[Callable[..., Any], str]]]]]"; expected
   1136                 # "Callable[[Any], Any]"
-> 1137                 mapped = lib.map_infer(
   1138                     values,
   1139                     f,  # type: ignore[arg-type]

~/model/.venv/lib/python3.9/site-packages/pandas/_libs/lib.pyx in pandas._libs.lib.map_infer()

~/model/.venv/lib/python3.9/site-packages/shapely/geos.py in read(self, text)
    288         """Returns geometry from WKT"""
    289         if not isinstance(text, str):
--> 290             raise TypeError("Only str is accepted.")
    291         text = text.encode()
    292         c_string = c_char_p(text)

TypeError: Only str is accepted.

Probably worth doing the iteration over the items and converting each one, but even just skipping on nested instead of failing also seems fine (that would just be a one-liner). @jimfulton @tswast any preference?

EDIT: I updated the title/example here to focus on the geography_as_object flag rather than the to_geodataframe method, since that applies to both methods and is where the actual error occurs.

[tswast]

Skipping the conversion if mode="REPEATED" sounds like a decent option to me.