feat: adds the capability to include custom user agent string (#819)

* adds the capability to include customer user agent string

* Adds deprecation warning

* 🦉 Updates from OwlBot post-processor

See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md

* clarifies purpose and asert for two tests regarding old verbose arg

* 🦉 Updates from OwlBot post-processor

See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md

* updates comment regarding verbosity

* Update pandas_gbq/gbq.py

---------

Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>

pandas_gbq/gbq.py

@@ -267,6 +267,8 @@ def __init__(
         auth_redirect_uri=None,
         client_id=None,
         client_secret=None,
+        user_agent=None,
+        rfc9110_delimiter=False,
     ):
         global context
         from google.api_core.exceptions import ClientError, GoogleAPIError
@@ -284,6 +286,8 @@ def __init__(
         self.auth_redirect_uri = auth_redirect_uri
         self.client_id = client_id
         self.client_secret = client_secret
+        self.user_agent = user_agent
+        self.rfc9110_delimiter = rfc9110_delimiter
 
         default_project = None
 
@@ -337,11 +341,15 @@ def log_elapsed_seconds(self, prefix="Elapsed", postfix="s.", overlong=6):
 
     def get_client(self):
         import google.api_core.client_info
-        import pandas
 
         bigquery = FEATURES.bigquery_try_import()
+
+        user_agent = create_user_agent(
+            user_agent=self.user_agent, rfc9110_delimiter=self.rfc9110_delimiter
+        )
+
         client_info = google.api_core.client_info.ClientInfo(
-            user_agent="pandas-{}".format(pandas.__version__)
+            user_agent=user_agent,
         )
         return bigquery.Client(
             project=self.project_id,
@@ -961,6 +969,8 @@ def to_gbq(
     auth_redirect_uri=None,
     client_id=None,
     client_secret=None,
+    user_agent=None,
+    rfc9110_delimiter=False,
 ):
     """Write a DataFrame to a Google BigQuery table.
 
@@ -1072,6 +1082,13 @@ def to_gbq(
     client_secret : str
         The Client Secret associated with the Client ID for the Google Cloud Project
         the user is attempting to connect to.
+    user_agent : str
+        Custom user agent string used as a prefix to the pandas version.
+    rfc9110_delimiter : bool
+        Sets user agent delimiter to a hyphen or a slash.
+        Default is False, meaning a hyphen will be used.
+
+        .. versionadded:: 0.23.3
     """
 
     _test_google_api_imports()
@@ -1130,6 +1147,8 @@ def to_gbq(
         auth_redirect_uri=auth_redirect_uri,
         client_id=client_id,
         client_secret=client_secret,
+        user_agent=user_agent,
+        rfc9110_delimiter=rfc9110_delimiter,
     )
     bqclient = connector.client
 
@@ -1409,3 +1428,59 @@ def create(self, dataset_id):
             self.client.create_dataset(dataset)
         except self.http_error as ex:
             self.process_http_error(ex)
+
+
+def create_user_agent(
+    user_agent: Optional[str] = None, rfc9110_delimiter: bool = False
+) -> str:
+    """Creates a user agent string.
+
+    The legacy format of our the user agent string was: `product-x.y.z` (where x,
+    y, and z are the major, minor, and micro version numbers).
+
+    Users are able to prepend this string with their own user agent identifier
+    to render something similar to `<my_user_agent> pandas-x.y.z`.
+
+    The legacy format used a hyphen to separate the product from the product
+    version which differs slightly from the format recommended by RFC9110, which is:
+    `product/x.y.z`. To produce a user agent more in line with the RFC, set
+    rfc9110_delimiter to True. This setting does not depend on whether a
+    user_agent is also supplied.
+
+    Reference:
+        https://www.rfc-editor.org/info/rfc9110
+
+    Args:
+        user_agent (Optional[str]): User agent string.
+
+        rfc9110_delimiter (Optional[bool]): Sets delimiter to a hyphen or a slash.
+        Default is False, meaning a hyphen will be used.
+
+    Returns (str):
+        Customized user agent string.
+
+    Deprecation Warning:
+        In a future major release, the default delimiter will be changed to
+        a `/` in accordance with RFC9110.
+    """
+    import pandas as pd
+
+    if rfc9110_delimiter:
+        delimiter = "/"
+    else:
+        warnings.warn(
+            "In a future major release, the default delimiter will be "
+            "changed to a `/` in accordance with RFC9110.",
+            PendingDeprecationWarning,
+            stacklevel=2,
+        )
+        delimiter = "-"
+
+    identity = f"pandas{delimiter}{pd.__version__}"
+
+    if user_agent is None:
+        user_agent = identity
+    else:
+        user_agent = f"{user_agent} {identity}"
+
+    return user_agent

tests/unit/test_gbq.py

@@ -635,7 +635,15 @@ def test_read_gbq_wo_verbose_w_new_pandas_no_warnings(monkeypatch, recwarn):
         mock.PropertyMock(return_value=False),
     )
     gbq.read_gbq("SELECT 1", project_id="my-project", dialect="standard")
-    assert len(recwarn) == 0
+    # This test was intended to check for warnings about the deprecation of
+    # the argument `verbose` (which was removed from gbq (~v0.4.0) and
+    # pandas (~v0.23.0). (See https://github.com/googleapis/python-bigquery-pandas/pull/158/files)
+    # This test should not fail upon seeing a warning in regards to a pending
+    # deprecation related to rfc9110 delimiters.
+    # TODO this and related tests have likely outlived their usefulness,
+    # consider removing.
+    for warning in recwarn.list:
+        assert "delimiter" in str(warning.message)
 
 
 def test_read_gbq_with_old_bq_raises_importerror(monkeypatch):
@@ -660,7 +668,15 @@ def test_read_gbq_with_verbose_old_pandas_no_warnings(monkeypatch, recwarn):
         dialect="standard",
         verbose=True,
     )
-    assert len(recwarn) == 0
+    # This test was intended to check for warnings about the deprecation of
+    # the argument `verbose` (which was removed from gbq (~v0.4.0) and
+    # pandas (~v0.23.0). (See https://github.com/googleapis/python-bigquery-pandas/pull/158/files)
+    # This test should not fail upon seeing a warning in regards to a pending
+    # deprecation related to rfc9110 delimiters.
+    # TODO this and related tests have likely outlived their usefulness,
+    # consider removing.
+    for warning in recwarn.list:
+        assert "delimiter" in str(warning.message)
 
 
 def test_read_gbq_with_private_raises_notimplmentederror():

tests/unit/test_to_gbq.py

@@ -4,6 +4,7 @@
 
 import google.api_core.exceptions
 import google.cloud.bigquery
+import pandas as pd
 from pandas import DataFrame
 import pytest
 
@@ -158,3 +159,27 @@ def test_to_gbq_with_if_exists_unknown():
             project_id="myproj",
             if_exists="unknown",
         )
+
+
+@pytest.mark.parametrize(
+    "user_agent,rfc9110_delimiter,expected",
+    [
+        (
+            "test_user_agent/2.0.42",
+            False,
+            f"test_user_agent/2.0.42 pandas-{pd.__version__}",
+        ),
+        (None, False, f"pandas-{pd.__version__}"),
+        (
+            "test_user_agent/2.0.42",
+            True,
+            f"test_user_agent/2.0.42 pandas/{pd.__version__}",
+        ),
+        (None, True, f"pandas/{pd.__version__}"),
+    ],
+)
+def test_create_user_agent(user_agent, rfc9110_delimiter, expected):
+    from pandas_gbq.gbq import create_user_agent
+
+    result = create_user_agent(user_agent, rfc9110_delimiter)
+    assert result == expected