Python/sqlalchemy/get column names after running SQL

Question

I have an already existing oracle DB, and there are many tables. Each table has hundreds of the columns as such I don't want to define class. I use below code to create the table object and it seems works as I can print the columns:

def get_engine(username, password, db):
    try:
        connection = f"oracle+cx_oracle://{username}:{password}@{db}"
        
        engine = sqlalchemy.create_engine(connection, arraysize=1000)
        return engine
    except SQLAlchemyError as e:
        print(e)

engine = get_engine(username = username, password = password, db = db)
Base = declarative_base()
meta = Base.metadata
meta.bind = engine
schema = "beta"

activity = Table('ACTIVITY',
                        meta,
                        autoload=True, 
                         autoload_with=engine,
                        schema=schema 
                        )

stmt = select([
    activity.columns.activity_date, 
    activity.columns.activity_number,
    activity.columns.activity_user
]
).where(and_(
    activity.columns.activity_date== date(2021, 9, 1)
#     ,
#    activity_date.columns.activity_user <> 'Mary George'
))

Then - I read the result into a pandas data frame

p = pd.DataFrame(results)
connection.close()

But i couldn't get the column names from the stmt, or the results.

Can you please shed some lights?

Thanks