gitea_admin/sientia-dataops-laborious-legacy
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

rc/SIENTIAPDE-2167- fix: corrige artefato de referência desatualizado no teste e2e de drift

Browse Source
This commit is contained in:
Carla Pinto
2026-10-09 14:55:42 +02:00
parent 7e371fd758
commit 693efbb72b
2 changed files with 7 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
e2e/test_drift.py
View File

@@ -142,7 +142,7 @@ def _recent_minute_timestamps(count: int, offset_minutes: int = 6) -> list[str]:
def _configure_reference_csv(mlflow_repository_stub, reference_rows: pd.DataFrame) -> None:
"""
Wire ``mlflow_repository_stub`` so ``get_reference_data`` returns
``reference_rows`` by writing them to ``dst_path/retrain_input.csv``.
``reference_rows`` by writing them to ``dst_path/evaluation_data.csv``.
Args:
- mlflow_repository_stub: External MLflow repository fixture.
@@ -158,7 +158,8 @@ def _configure_reference_csv(mlflow_repository_stub, reference_rows: pd.DataFram
run_id='fake-reference-run'
)
file_info = MagicMock()
file_info.path = 'retrain_input.csv'
# Matches `MLFlow._REFERENCE_ARTIFACT_CANDIDATES` (SIENTIAPDE-1646).
file_info.path = 'evaluation_data.csv'
mlflow_repository_stub._client.list_artifacts.return_value = [file_info]
mlflow_repository_stub.download_artifacts.side_effect = _download

8
laborious/activities/mlflow.py
View File

@@ -664,7 +664,7 @@ class MLFlow(SientiaMonitoring):
"""
Pick the first available reference CSV artifact path from the MLflow run.
Candidates are checked in priority order: ``retrain_input.csv``, then ``train_data.csv``.
Candidates are checked in priority order: ``evaluation_data.csv``, then ``test_data.csv``.
A path matches when it equals the candidate or ends with ``/<candidate>`` for nested layouts.
Args:
@@ -703,9 +703,9 @@ class MLFlow(SientiaMonitoring):
"""
Download reference training CSV from the MLflow run linked to the production alias.
Resolves ``retrain_input.csv`` or ``train_data.csv`` via artifact listing before download.
``retrain_input.csv`` is preferred when both exist (most recent retrain snapshot). Used by
drift workflows to compare live data against the reference distribution logged with the model.
Resolves ``evaluation_data.csv`` or ``test_data.csv`` via artifact listing before download.
``evaluation_data.csv`` is preferred when both exist. Used by drift workflows to compare
live data against the reference distribution logged with the model.
Timestamps are normalized to ``DATETIME_FORMAT`` string columns before returning records.
Args: