feature/SIENTIAPDE-2162 - Inject mongo_repository into gates and add best-effort prediction metadata write

This commit is contained in:
Carla Pinto
2026-09-28 19:33:24 +02:00
parent 9920cd61c2
commit 8b69b2cb29
5 changed files with 266 additions and 1 deletions

View File

@@ -7,6 +7,8 @@ with workflow.unsafe.imports_passed_through():
from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
from sientia_do.observability.logger import Logger
from sientia_do.repository.minio_repository_sync import MinioRepository
from sientia_do.repository.mongodb_repository_sync import MongoDBRepository
from sientia_do.utils.connectors_config import build_mongodb_config
from sientia_model.model_repository.mlflow_repository import SientiaMLflowRepository
from sientia_model.model_repository.plugin_store import PluginStore
@@ -48,6 +50,7 @@ class Activities(Storage, MLFlow, Gates, OPC, ModelMetrics, API):
notification_handler: NotificationHandler,
metrics_controller: MetricsController | None = None,
mlflow_repository: SientiaMLflowRepository | None = None,
mongo_repository: MongoDBRepository | None = None,
):
"""
Wire Postgres, MinIO, MLflow, OPC, gates, metrics, and PI Web API into a single object.
@@ -67,6 +70,9 @@ class Activities(Storage, MLFlow, Gates, OPC, ModelMetrics, API):
- metrics_controller: Optional shared controller; if ``None``, a new one is created.
- mlflow_repository: Optional ``SientiaMLflowRepository`` for unit/e2e tests; in production
leave unset so the repository is built from environment via ``build_mlflow_config()``.
- mongo_repository: Optional ``MongoDBRepository`` (gate metadata, best-effort) for
unit/e2e tests; in production leave unset so the repository is built from environment
via ``build_mongodb_config()``, same pattern as ``mlflow_repository`` above.
Raises:
Exception: If any parent ``__init__`` fails (e.g. invalid config keys).
@@ -89,6 +95,16 @@ class Activities(Storage, MLFlow, Gates, OPC, ModelMetrics, API):
metrics_controller=mc,
)
if mongo_repository is None:
mongo_cfg = build_mongodb_config()
mongo_repository = MongoDBRepository(
connection_string=mongo_cfg['connection_string'],
database_name=mongo_cfg['database_name'],
logger=logger,
notification_handler=notification_handler,
metrics_controller=mc,
)
minio_repository = MinioRepository(
endpoint=minio_config['endpoint_url'],
access_key=minio_config['access_key'],
@@ -129,6 +145,7 @@ class Activities(Storage, MLFlow, Gates, OPC, ModelMetrics, API):
Gates.__init__(
self,
minio_repository=minio_repository,
mongo_repository=mongo_repository,
logger=logger,
notification_handler=notification_handler,
metrics_controller=mc,

View File

@@ -5,6 +5,7 @@ with workflow.unsafe.imports_passed_through():
from collections.abc import Callable, Mapping
from typing import Any
from bson import ObjectId
from pandas import DataFrame
from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
from sientia_do.notifications.models import NotificationLevel
@@ -12,6 +13,7 @@ with workflow.unsafe.imports_passed_through():
from sientia_do.observability.metrics_controller import MetricsController
from sientia_do.observability.sientia_monitoring import SientiaMonitoring
from sientia_do.repository.minio_repository_sync import MinioRepository
from sientia_do.repository.mongodb_repository_sync import MongoDBRepository
from sientia_do.utils.formatters import create_sample_dict
from laborious import metrics
@@ -86,11 +88,14 @@ class Gates(SientiaMonitoring):
"""
minio_repository: MinioRepository | None = None
mongo_repository: MongoDBRepository | None = None
_MAX_DEBUG_DATAFRAME_ROWS = 100
_PREDICTION_METADATA_COLLECTION = 'prediction_metadata'
def __init__(
self,
minio_repository: MinioRepository | None = None,
mongo_repository: MongoDBRepository | None = None,
logger: Logger | None = None,
notification_handler: NotificationHandler | None = None,
metrics_controller: MetricsController | None = None,
@@ -101,11 +106,14 @@ class Gates(SientiaMonitoring):
Args:
logger: Logger instance for observability and debugging
notification_handler: Notification handler for alerts and monitoring
mongo_repository: MongoDB, where gate metadata (triggered filter, policy, counter) is
written best-effort. `None` is a valid state (see `_write_prediction_metadata`).
Raises:
Exception: If BaseActivity initialization fails
"""
self.minio_repository = minio_repository
self.mongo_repository = mongo_repository
SientiaMonitoring.__init__(
self,
logger=logger,
@@ -123,6 +131,11 @@ class Gates(SientiaMonitoring):
self.minio_repository.close()
finally:
self.minio_repository = None
if self.mongo_repository is not None:
try:
self.mongo_repository.close()
finally:
self.mongo_repository = None
SientiaMonitoring.shutdown(self)
def __del__(self):
@@ -146,6 +159,56 @@ class Gates(SientiaMonitoring):
metadata,
)
def _write_prediction_metadata(
self, document: dict[str, Any] | None, metadata: dict[str, Any]
) -> str | None:
"""
Best-effort write of gate metadata (triggered filter, policy, counter) to Mongo.
Implements both business rules for this feature in one place:
- The prediction write must never be blocked or failed by a metadata write. A failed
`insert(...)` is swallowed, never re-raised, logged here with gate-level context (which
`mongo_repository.insert(...)` itself can't know) — matching how `storage.py`'s
`cleanup_minio_objects_expired` also logs again at the call site even though the
repository already logs internally and emits its own Prometheus metric
(`MONGODB_WRITE_ERROR_COUNT`); the generic library log is for the metric/trace, this
one is for whoever is reading logs for this specific gate.
- Nothing is written when there's nothing relevant to report. A falsy `document` (e.g.
no filter triggered) skips Mongo entirely, so callers can call this unconditionally
and get `None` back for "no metadata this time" instead of deciding upfront whether
there's something worth saving.
`insert(...)` doesn't hand back the inserted document's `_id`, so it is generated here,
client-side, before the insert, as a plain string (not a BSON ObjectId) — matching this
codebase's convention of string `_id`s (e.g. the `models` collection) and the 24-hex-char
size of the `predictions.metadata_id` Postgres column that will reference it.
Args:
document: Metadata payload (filter, policy, counter, etc.); `None`/empty skips the
write entirely. No `_id` key expected.
metadata: Workflow metadata for contextual logging.
Return:
The generated id, or `None` when there's nothing to write, Mongo isn't configured, or
the insert failed.
"""
if not document or self.mongo_repository is None:
return None
generated_id = str(ObjectId())
try:
self.mongo_repository.insert(
self._PREDICTION_METADATA_COLLECTION,
{**document, '_id': generated_id},
metadata,
)
except Exception as e:
self.error(f'Failed to write prediction metadata: {e}', metadata)
return None
return generated_id
@staticmethod
def _read_filter_entry(config: dict[str, Any]) -> tuple[str, dict[str, Any]]:
"""