chore: use bigframes.pandas._read_gbq_colab() in benchmarks (#1942)

Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>

Author: tswast

tests/benchmark/read_gbq_colab/aggregate_output.py

@@ -15,19 +15,15 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas as bpd
 
 PAGE_SIZE = utils.READ_GBQ_COLAB_PAGE_SIZE
 
 
-def aggregate_output(
-    *, project_id, dataset_id, table_id, session: bigframes.session.Session
-):
+def aggregate_output(*, project_id, dataset_id, table_id):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    df = session._read_gbq_colab(
-        f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}"
-    )
+    df = bpd._read_gbq_colab(f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}")
 
     # Simulate getting the first page, since we'll always do that first in the UI.
     df.shape

tests/benchmark/read_gbq_colab/dry_run.py

@@ -15,20 +15,20 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas
 
 
-def dry_run(*, project_id, dataset_id, table_id, session: bigframes.session.Session):
+def dry_run(*, project_id, dataset_id, table_id):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    session._read_gbq_colab(
+    bigframes.pandas._read_gbq_colab(
         f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}",
         dry_run=True,
     )
 
 
 if __name__ == "__main__":
-    config = utils.get_configuration(include_table_id=True)
+    config = utils.get_configuration(include_table_id=True, start_session=False)
     current_path = pathlib.Path(__file__).absolute()
 
     utils.get_execution_time(
@@ -38,5 +38,4 @@ def dry_run(*, project_id, dataset_id, table_id, session: bigframes.session.Sess
         project_id=config.project_id,
         dataset_id=config.dataset_id,
         table_id=config.table_id,
-        session=config.session,
     )

tests/benchmark/read_gbq_colab/filter_output.py

@@ -15,19 +15,20 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas as bpd
 
 PAGE_SIZE = utils.READ_GBQ_COLAB_PAGE_SIZE
 
 
 def filter_output(
-    *, project_id, dataset_id, table_id, session: bigframes.session.Session
+    *,
+    project_id,
+    dataset_id,
+    table_id,
 ):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    df = session._read_gbq_colab(
-        f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}"
-    )
+    df = bpd._read_gbq_colab(f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}")
 
     # Simulate getting the first page, since we'll always do that first in the UI.
     df.shape
@@ -54,5 +55,4 @@ def filter_output(
         project_id=config.project_id,
         dataset_id=config.dataset_id,
         table_id=config.table_id,
-        session=config.session,
     )

tests/benchmark/read_gbq_colab/first_page.py

@@ -15,15 +15,15 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas
 
 PAGE_SIZE = utils.READ_GBQ_COLAB_PAGE_SIZE
 
 
-def first_page(*, project_id, dataset_id, table_id, session: bigframes.session.Session):
+def first_page(*, project_id, dataset_id, table_id):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    df = session._read_gbq_colab(
+    df = bigframes.pandas._read_gbq_colab(
         f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}"
     )
 
@@ -33,7 +33,7 @@ def first_page(*, project_id, dataset_id, table_id, session: bigframes.session.S
 
 
 if __name__ == "__main__":
-    config = utils.get_configuration(include_table_id=True)
+    config = utils.get_configuration(include_table_id=True, start_session=False)
     current_path = pathlib.Path(__file__).absolute()
 
     utils.get_execution_time(
@@ -43,5 +43,4 @@ def first_page(*, project_id, dataset_id, table_id, session: bigframes.session.S
         project_id=config.project_id,
         dataset_id=config.dataset_id,
         table_id=config.table_id,
-        session=config.session,
     )

tests/benchmark/read_gbq_colab/last_page.py

@@ -15,15 +15,15 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas
 
 PAGE_SIZE = utils.READ_GBQ_COLAB_PAGE_SIZE
 
 
-def last_page(*, project_id, dataset_id, table_id, session: bigframes.session.Session):
+def last_page(*, project_id, dataset_id, table_id):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    df = session._read_gbq_colab(
+    df = bigframes.pandas._read_gbq_colab(
         f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}"
     )
 
@@ -34,7 +34,7 @@ def last_page(*, project_id, dataset_id, table_id, session: bigframes.session.Se
 
 
 if __name__ == "__main__":
-    config = utils.get_configuration(include_table_id=True)
+    config = utils.get_configuration(include_table_id=True, start_session=False)
     current_path = pathlib.Path(__file__).absolute()
 
     utils.get_execution_time(
@@ -44,5 +44,4 @@ def last_page(*, project_id, dataset_id, table_id, session: bigframes.session.Se
         project_id=config.project_id,
         dataset_id=config.dataset_id,
         table_id=config.table_id,
-        session=config.session,
     )

tests/benchmark/read_gbq_colab/sort_output.py

@@ -15,17 +15,15 @@
 
 import benchmark.utils as utils
 
-import bigframes.session
+import bigframes.pandas
 
 PAGE_SIZE = utils.READ_GBQ_COLAB_PAGE_SIZE
 
 
-def sort_output(
-    *, project_id, dataset_id, table_id, session: bigframes.session.Session
-):
+def sort_output(*, project_id, dataset_id, table_id):
     # TODO(tswast): Support alternative query if table_id is a local DataFrame,
     # e.g. "{local_inline}" or "{local_large}"
-    df = session._read_gbq_colab(
+    df = bigframes.pandas._read_gbq_colab(
         f"SELECT * FROM `{project_id}`.{dataset_id}.{table_id}"
     )
 
@@ -44,7 +42,7 @@ def sort_output(
 
 
 if __name__ == "__main__":
-    config = utils.get_configuration(include_table_id=True)
+    config = utils.get_configuration(include_table_id=True, start_session=False)
     current_path = pathlib.Path(__file__).absolute()
 
     utils.get_execution_time(
@@ -54,5 +52,4 @@ def sort_output(
         project_id=config.project_id,
         dataset_id=config.dataset_id,
         table_id=config.table_id,
-        session=config.session,
     )

tests/benchmark/utils.py

@@ -25,12 +25,12 @@
 class BenchmarkConfig:
     project_id: str
     dataset_id: str
-    session: bigframes.Session
+    session: bigframes.Session | None
     benchmark_suffix: str | None
     table_id: str | None = None
 
 
-def get_configuration(include_table_id=False) -> BenchmarkConfig:
+def get_configuration(include_table_id=False, start_session=True) -> BenchmarkConfig:
     parser = argparse.ArgumentParser()
     parser.add_argument(
         "--project_id",
@@ -65,7 +65,7 @@ def get_configuration(include_table_id=False) -> BenchmarkConfig:
     )
 
     args = parser.parse_args()
-    session = _initialize_session(_str_to_bool(args.ordered))
+    session = _initialize_session(_str_to_bool(args.ordered)) if start_session else None
 
     return BenchmarkConfig(
         project_id=args.project_id,