Use str type for driver and name in HiveDialect (#450)

PyHive's HiveDialect usage of bytes for the name and driver fields is not the norm is causing issues upstream: apache/superset#22316
Even other dialects within PyHive use strings. SQLAlchemy does not strictly require a string, but all the stock dialects return a string, so I figure it is heavily implied.

I think the risk of breaking something upstream with this change is low (but it is there ofc). I figure in most cases we just make someone's `str(dialect.driver)` expression redundant.

Examples for some of the other stock sqlalchemy dialects (name and driver fields using str):
https://github.com/zzzeek/sqlalchemy/blob/main/lib/sqlalchemy/dialects/sqlite/pysqlite.py#L501
https://github.com/zzzeek/sqlalchemy/blob/main/lib/sqlalchemy/dialects/sqlite/base.py#L1891
https://github.com/zzzeek/sqlalchemy/blob/main/lib/sqlalchemy/dialects/mysql/base.py#L2383
https://github.com/zzzeek/sqlalchemy/blob/main/lib/sqlalchemy/dialects/mysql/mysqldb.py#L113
https://github.com/zzzeek/sqlalchemy/blob/main/lib/sqlalchemy/dialects/mysql/pymysql.py#L59

pyhive/sqlalchemy_hive.py

@@ -228,8 +228,8 @@ def _translate_colname(self, colname):
 
 
 class HiveDialect(default.DefaultDialect):
-    name = b'hive'
-    driver = b'thrift'
+    name = 'hive'
+    driver = 'thrift'
     execution_ctx_cls = HiveExecutionContext
     preparer = HiveIdentifierPreparer
     statement_compiler = HiveCompiler