Support direct output to BigQuery

[bashir2]

Currently, if users want to export their FHIR resources to BigQuery, they should either export to a GCP FHIR-store first OR create the Parquet files and manually import them through BigQuery tools. We should support direct output to BigQuery in our pipelines.

A related issue is #454

[jjtswan]

Just to follow up - direct output to BigQuery is a lower priority given two reasons:

1. If you're outputing to BigQuery, then likely you can use GCP FHIR Store, and leverage HC API to facilitate the export.
2. It seems like generating parquet from the FHIR-data-pipes can be loaded into GCP, so there is a workaround to getting this into GCP without adding a direct integration.