Skip to content

fix: Only include CandidateResponse if a response is present #6367

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
copybara-service wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: Only include CandidateResponse if a response is present #6367

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
57 changes: 46 additions & 11 deletions tests/unit/vertexai/genai/replays/test_create_evaluation_run.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,9 @@
)
),
)

INFERENCE_CONFIG = types.EvaluationRunInferenceConfig(
model="projects/503583131166/locations/us-central1/publishers/google/models/gemini-2.5-flash"
)

def test_create_eval_run_data_source_evaluation_set(client):
"""Tests that create_evaluation_run() creates a correctly structured EvaluationRun."""
Expand Down Expand Up @@ -189,9 +191,6 @@ def test_create_eval_run_data_source_bigquery_request_set(client):
def test_create_eval_run_with_inference_configs(client):
"""Tests that create_evaluation_run() creates a correctly structured EvaluationRun with inference_configs."""
client._api_client._http_options.api_version = "v1beta1"
inference_config = types.EvaluationRunInferenceConfig(
model="projects/503583131166/locations/us-central1/publishers/google/models/gemini-2.5-flash"
)
evaluation_run = client.evals.create_evaluation_run(
name="test_inference_config",
display_name="test_inference_config",
Expand All @@ -200,7 +199,7 @@ def test_create_eval_run_with_inference_configs(client):
),
dest=GCS_DEST,
metrics=[GENERAL_QUALITY_METRIC],
inference_configs={"model_1": inference_config},
inference_configs={"model_1": INFERENCE_CONFIG},
labels={"label1": "value1"},
)
assert isinstance(evaluation_run, types.EvaluationRun)
Expand All @@ -216,7 +215,7 @@ def test_create_eval_run_with_inference_configs(client):
),
metrics=[GENERAL_QUALITY_METRIC],
)
assert evaluation_run.inference_configs["model_1"] == inference_config
assert evaluation_run.inference_configs["model_1"] == INFERENCE_CONFIG
assert evaluation_run.labels == {
"label1": "value1",
}
Expand Down Expand Up @@ -318,6 +317,45 @@ def test_create_eval_run_with_inference_configs(client):
# )
# assert evaluation_run.error is None

import pandas as pd

def test_create_eval_run_data_source_evaluation_dataset_inference_config(client):
"""Tests that create_evaluation_run() creates a correctly structured EvaluationRun with EvaluationDataset."""
input_df = pd.DataFrame(
{
"prompt": ["prompt1", "prompt2"],
"reference": ["reference1", "reference2"],
}
)
evaluation_run = client.evals.create_evaluation_run(
name="test9",
display_name="test9",
dataset=types.EvaluationDataset(
candidate_name="candidate_1",
eval_dataset_df=input_df,
),
dest=GCS_DEST,
metrics=[GENERAL_QUALITY_METRIC],
inference_configs={"candidate_1": INFERENCE_CONFIG},
)
assert isinstance(evaluation_run, types.EvaluationRun)
assert evaluation_run.display_name == "test9"
assert evaluation_run.state == types.EvaluationRunState.PENDING
assert isinstance(evaluation_run.data_source, types.EvaluationRunDataSource)
# Check evaluation set
assert evaluation_run.data_source.evaluation_set
eval_set = client.evals.get_evaluation_set(
name=evaluation_run.data_source.evaluation_set
)
assert len(eval_set.evaluation_items) == 2
assert evaluation_run.inference_configs["candidate_1"] == INFERENCE_CONFIG
# Check evaluation items
for i, eval_item_name in enumerate(eval_set.evaluation_items):
eval_item = client.evals.get_evaluation_item(name=eval_item_name)
assert eval_item.evaluation_item_type == types.EvaluationItemType.REQUEST
assert eval_item.evaluation_request.prompt.text == input_df.iloc[i]["prompt"]
assert eval_item.evaluation_request.candidate_responses == []
assert evaluation_run.error is None

pytest_plugins = ("pytest_asyncio",)

Expand Down Expand Up @@ -370,9 +408,6 @@ async def test_create_eval_run_async(client):
async def test_create_eval_run_async_with_inference_configs(client):
"""Tests that create_evaluation_run() creates a correctly structured EvaluationRun with inference_configs asynchronously."""
client._api_client._http_options.api_version = "v1beta1"
inference_config = types.EvaluationRunInferenceConfig(
model="projects/503583131166/locations/us-central1/publishers/google/models/gemini-2.5-flash"
)
evaluation_run = await client.aio.evals.create_evaluation_run(
name="test_inference_config_async",
display_name="test_inference_config_async",
Expand All @@ -381,7 +416,7 @@ async def test_create_eval_run_async_with_inference_configs(client):
),
dest=GCS_DEST,
metrics=[GENERAL_QUALITY_METRIC],
inference_configs={"model_1": inference_config},
inference_configs={"model_1": INFERENCE_CONFIG},
labels={"label1": "value1"},
)
assert isinstance(evaluation_run, types.EvaluationRun)
Expand All @@ -397,7 +432,7 @@ async def test_create_eval_run_async_with_inference_configs(client):
),
metrics=[GENERAL_QUALITY_METRIC],
)
assert evaluation_run.inference_configs["model_1"] == inference_config
assert evaluation_run.inference_configs["model_1"] == INFERENCE_CONFIG
assert evaluation_run.labels == {
"label1": "value1",
}
Expand Down
21 changes: 10 additions & 11 deletions vertexai/_genai/_evals_common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1962,6 +1962,15 @@ def _create_evaluation_set_from_dataframe(
for event in row[_evals_constant.INTERMEDIATE_EVENTS]:
if CONTENT in event:
intermediate_events.append(event[CONTENT])
candidate_responses = []
if _evals_constant.RESPONSE in row:
candidate_responses.append(
types.CandidateResponse(
candidate=candidate_name or "Candidate 1",
text=row[_evals_constant.RESPONSE],
events=intermediate_events or None,
)
)
eval_item_requests.append(
types.EvaluationItemRequest(
prompt=(
Expand All @@ -1974,17 +1983,7 @@ def _create_evaluation_set_from_dataframe(
if _evals_constant.REFERENCE in row
else None
),
candidate_responses=[
types.CandidateResponse(
candidate=candidate_name or "Candidate 1",
text=row.get(_evals_constant.RESPONSE, None),
events=(
intermediate_events
if len(intermediate_events) > 0
else None
),
)
],
candidate_responses=candidate_responses,
)
)
logger.info("Writing evaluation item requests to GCS.")
Expand Down
Loading