feat: add CloudSQLMySQLEngine class

[jackwotherspoon]

Adding CloudSQLMySQLEngine that will be leveraged by the vector store, docloader and memory APIs.

• Code complete
• Lint passing
• Add integration tests (being done in follow-up PR)

Manually tested the following currently:

from langchain_google_cloud_sql_mysql import CloudSQLMySQLEngine
import sqlalchemy

engine = CloudSQLMySQLEngine.from_instance(
    "my-project", "my-region", "my-instance", "my-database"
)

# currently using engine.engine as CloudSQLMySQLEngine does not have a `connect` method
with engine.engine.connect() as db_conn:
    time = db_conn.execute(sqlalchemy.text("SELECT NOW()")).fetchone()
    print(time[0])

[totoleon]

nit: should we have some validation check on input argument? For example, if engine is given, we shouldn't take other arguments like project ID and region etc.

We don't have to do it in one CL, but having a TODO comment would be good.

[jackwotherspoon]

Removed engine from __init__ entirely as I have removed from_engine and from_connection_string which were the methods passing in the engine

[jackwotherspoon]

@totoleon ready for another pass

[jackwotherspoon]

Merging to unblock any future dev, can make any future comments/suggestions in follow-up PR