gigl.common.data.load_embeddings_to_bigquery

gigl.common.data.export.load_embeddings_to_bigquery(gcs_folder: GcsUri, project_id: str, dataset_id: str, table_id: str) → None

Loads multiple Avro files containing GNN embeddings from GCS into BigQuery.

Note that this function will upload all Avro files in the GCS folder to BigQuery, recursively. So if we have some nested directories, e.g.:

gs://MY BUCKET/embeddings/shard_0000.avro gs://MY BUCKET/embeddings/nested/shard_0001.avro

Both files will be uploaded to BigQuery.

Args:

gcs_folder (GcsUri): The GCS folder containing the Avro files with embeddings. project_id (str): The GCP project ID. dataset_id (str): The BigQuery dataset ID. table_id (str): The BigQuery table ID.