Store pre-aggregations in a directory within an export bucket with the BigQuery driver

**Is your feature request related to a problem? Please describe.**
I want to store pre-aggregations for different tenants in separate folders within the same export bucket.

**Describe the solution you'd like**
Maybe to have `CUBEJS_DB_EXPORT_BUCKET` changed to `CUBEJS_DB_EXPORT_PATH`, or at least have the semantic changes to support both plain buckets and `s3://...` or `gs://...` paths like boto/gsutil does.

**Describe alternatives you've considered**
There is no alternative solution except for using different buckets altogether.

**Additional context**
The BigQuery driver uses a bucket eg `CUBEJS_DB_EXPORT_BUCKET=cube-cloud-staging-export-bucket`. The Athena driver supports S3 paths eg `CUBEJS_DB_EXPORT_BUCKET=s3://cubejs-benchmarks/home/app/.../`.