Support for writing to partitioned and clustered BigQuery tables using Python

According to the BigQuery sink docs:
```

BigQuery sink currently does not fully support writing to BigQuery time partitioned tables. But writing
to a *single* partition may work if that does not involve creating a new table (for example, when writing
to an existing table with `create_disposition=CREATE_NEVER` and `write_disposition=WRITE_APPEND`).

```

The sink should support partitioning (and clustering).

Imported from Jira [BEAM-7108](https://issues.apache.org/jira/browse/BEAM-7108). Original Jira may contain additional context.
Reported by: oliviervg1.