DB-API depends on pyarrow when decimal query parameters are used

[tswast]

It seems we introduced a hard dependency on pyarrow to detect data types of query parameters in the DB-API module. While pyarrow is recommended as it's needed to use the BigQuery Storage API, I don't think we want to depend on it for basic functionality.

I believe we added this hard dependency here: https://github.com/googleapis/python-bigquery/pull/527/files#diff-2221cd9b6b85f2ed1f943c474e8a8628e6ca28ec9a96c0aa5873cd6a66826bdeR193

We use pyarrow to disambiguate between NUMERIC and BIGNUMERIC. I'm not seeing a good way to check the precision / scale in the decimal type, so in the case that pyarrow is not available, we should guess the type is NUMERIC.

Stack trace

__________________________________________________________________________________ test_dml ___________________________________________________________________________________
test/test_sqlalchemy_bigquery.py:430: in test_dml
    engine.execute(table_dml.insert(ONE_ROW_CONTENTS_DML))
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:2235: in execute
    return connection.execute(statement, *multiparams, **params)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:1011: in execute
    return meth(self, multiparams, params)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/sql/elements.py:298: in _execute_on_connection
    return connection._execute_clauseelement(self, multiparams, params)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:1124: in _execute_clauseelement
    ret = self._execute_context(
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:1316: in _execute_context
    self._handle_dbapi_exception(
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:1514: in _handle_dbapi_exception
    util.raise_(exc_info[1], with_traceback=exc_info[2])
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/util/compat.py:182: in raise_
    raise exception
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/base.py:1276: in _execute_context
    self.dialect.do_execute(
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/sqlalchemy/engine/default.py:608: in do_execute
    cursor.execute(statement, parameters)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/_helpers.py:263: in with_closed_check
    return method(self, *args, **kwargs)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/cursor.py:165: in execute
    query_parameters = _helpers.to_query_parameters(parameters)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/_helpers.py:168: in to_query_parameters
    return to_query_parameters_dict(parameters)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/_helpers.py:147: in to_query_parameters_dict
    param = scalar_to_query_parameter(value, name=name)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/_helpers.py:51: in scalar_to_query_parameter
    parameter_type = bigquery_scalar_type(value)
../../miniconda3/envs/pybigquery/lib/python3.9/site-packages/google/cloud/bigquery/dbapi/_helpers.py:193: in bigquery_scalar_type
    scalar_object = pyarrow.scalar(value)
E   AttributeError: 'NoneType' object has no attribute 'scalar'