refactor: Rename internal functions used externally in _utils.py (#1959)

Author: jialuoo

bigframes/functions/_function_client.py

@@ -245,7 +245,7 @@ def provision_bq_managed_function(
 
         # Augment user package requirements with any internal package
         # requirements.
-        packages = _utils._get_updated_package_requirements(
+        packages = _utils.get_updated_package_requirements(
             packages, is_row_processor, capture_references, ignore_package_version=True
         )
         if packages:
@@ -258,7 +258,7 @@ def provision_bq_managed_function(
         bq_function_name = name
         if not bq_function_name:
             # Compute a unique hash representing the user code.
-            function_hash = _utils._get_hash(func, packages)
+            function_hash = _utils.get_hash(func, packages)
             bq_function_name = _utils.get_bigframes_function_name(
                 function_hash,
                 session_id,
@@ -539,12 +539,12 @@ def provision_bq_remote_function(
         """Provision a BigQuery remote function."""
         # Augment user package requirements with any internal package
         # requirements
-        package_requirements = _utils._get_updated_package_requirements(
+        package_requirements = _utils.get_updated_package_requirements(
             package_requirements, is_row_processor
         )
 
         # Compute a unique hash representing the user code
-        function_hash = _utils._get_hash(def_, package_requirements)
+        function_hash = _utils.get_hash(def_, package_requirements)
 
         # If reuse of any existing function with the same name (indicated by the
         # same hash of its source code) is not intended, then attach a unique

bigframes/functions/_function_session.py

@@ -597,7 +597,7 @@ def wrapper(func):
                 bqrf_metadata = _utils.get_bigframes_metadata(
                     python_output_type=py_sig.return_annotation
                 )
-                post_process_routine = _utils._build_unnest_post_routine(
+                post_process_routine = _utils.build_unnest_post_routine(
                     py_sig.return_annotation
                 )
                 py_sig = py_sig.replace(return_annotation=str)

bigframes/functions/_utils.py

@@ -63,7 +63,7 @@ def get_remote_function_locations(bq_location):
     return bq_location, cloud_function_region
 
 
-def _get_updated_package_requirements(
+def get_updated_package_requirements(
     package_requirements=None,
     is_row_processor=False,
     capture_references=True,
@@ -105,7 +105,7 @@ def _get_updated_package_requirements(
     return requirements
 
 
-def _clean_up_by_session_id(
+def clean_up_by_session_id(
     bqclient: bigquery.Client,
     gcfclient: functions_v2.FunctionServiceClient,
     dataset: bigquery.DatasetReference,
@@ -169,7 +169,7 @@ def _clean_up_by_session_id(
             pass
 
 
-def _get_hash(def_, package_requirements=None):
+def get_hash(def_, package_requirements=None):
     "Get hash (32 digits alphanumeric) of a function."
     # There is a known cell-id sensitivity of the cloudpickle serialization in
     # notebooks https://github.com/cloudpipe/cloudpickle/issues/538. Because of
@@ -279,7 +279,7 @@ def get_python_version(is_compat: bool = False) -> str:
     return f"python{major}{minor}" if is_compat else f"python-{major}.{minor}"
 
 
-def _build_unnest_post_routine(py_list_type: type[list]):
+def build_unnest_post_routine(py_list_type: type[list]):
     sdk_type = function_typing.sdk_array_output_type_from_python_type(py_list_type)
     assert sdk_type.array_element_type is not None
     inner_sdk_type = sdk_type.array_element_type

bigframes/functions/function.py

@@ -90,7 +90,7 @@ def _try_import_routine(
         return BigqueryCallableRoutine(
             udf_def,
             session,
-            post_routine=_utils._build_unnest_post_routine(override_type),
+            post_routine=_utils.build_unnest_post_routine(override_type),
         )
     return BigqueryCallableRoutine(udf_def, session, is_managed=not is_remote)
 
@@ -107,7 +107,7 @@ def _try_import_row_routine(
         return BigqueryCallableRowRoutine(
             udf_def,
             session,
-            post_routine=_utils._build_unnest_post_routine(override_type),
+            post_routine=_utils.build_unnest_post_routine(override_type),
         )
     return BigqueryCallableRowRoutine(udf_def, session, is_managed=not is_remote)
 

bigframes/pandas/__init__.py

@@ -293,7 +293,7 @@ def clean_up_by_session_id(
         session.bqclient, dataset, session_id
     )
 
-    bff_utils._clean_up_by_session_id(
+    bff_utils.clean_up_by_session_id(
         session.bqclient, session.cloudfunctionsclient, dataset, session_id
     )
 

bigframes/testing/utils.py

@@ -440,11 +440,11 @@ def get_function_name(func, package_requirements=None, is_row_processor=False):
     """Get a bigframes function name for testing given a udf."""
     # Augment user package requirements with any internal package
     # requirements.
-    package_requirements = bff_utils._get_updated_package_requirements(
+    package_requirements = bff_utils.get_updated_package_requirements(
         package_requirements, is_row_processor
     )
 
     # Compute a unique hash representing the user code.
-    function_hash = bff_utils._get_hash(func, package_requirements)
+    function_hash = bff_utils.get_hash(func, package_requirements)
 
     return f"bigframes_{function_hash}"

tests/system/large/functions/test_remote_function.py

@@ -527,8 +527,8 @@ def add_one(x):
         add_one_uniq, add_one_uniq_dir = make_uniq_udf(add_one)
 
         # Expected cloud function name for the unique udf
-        package_requirements = bff_utils._get_updated_package_requirements()
-        add_one_uniq_hash = bff_utils._get_hash(add_one_uniq, package_requirements)
+        package_requirements = bff_utils.get_updated_package_requirements()
+        add_one_uniq_hash = bff_utils.get_hash(add_one_uniq, package_requirements)
         add_one_uniq_cf_name = bff_utils.get_cloud_function_name(
             add_one_uniq_hash, session.session_id
         )