mlrun 1.4.0rc25__py3-none-any.whl → 1.5.0rc2__py3-none-any.whl

mlrun/api/crud/model_monitoring/grafana.py

@@ -20,14 +20,12 @@ import pandas as pd
 from fastapi.concurrency import run_in_threadpool
 from sqlalchemy.orm import Session
 
-import mlrun.api.crud
 import mlrun.api.utils.auth.verifier
-import mlrun.common.model_monitoring
 import mlrun.common.schemas
 from mlrun.api.utils.singletons.project_member import get_project_member
+from mlrun.common.model_monitoring.helpers import parse_model_endpoint_store_prefix
 from mlrun.errors import MLRunBadRequestError
 from mlrun.utils import config, logger
-from mlrun.utils.model_monitoring import parse_model_endpoint_store_prefix
 from mlrun.utils.v3io_clients import get_frames_client
 
 
@@ -62,7 +60,7 @@ async def grafana_list_endpoints(
     body: Dict[str, Any],
     query_parameters: Dict[str, str],
     auth_info: mlrun.common.schemas.AuthInfo,
-) -> List[mlrun.common.schemas.GrafanaTable]:
+) -> List[mlrun.common.schemas.model_monitoring.grafana.GrafanaTable]:
     project = query_parameters.get("project")
 
     # Filters
@@ -111,27 +109,47 @@ async def grafana_list_endpoints(
     endpoint_list.endpoints = allowed_endpoints
 
     columns = [
-        mlrun.common.schemas.GrafanaColumn(text="endpoint_id", type="string"),
-        mlrun.common.schemas.GrafanaColumn(text="endpoint_function", type="string"),
-        mlrun.common.schemas.GrafanaColumn(text="endpoint_model", type="string"),
-        mlrun.common.schemas.GrafanaColumn(text="endpoint_model_class", type="string"),
-        mlrun.common.schemas.GrafanaColumn(text="first_request", type="time"),
-        mlrun.common.schemas.GrafanaColumn(text="last_request", type="time"),
-        mlrun.common.schemas.GrafanaColumn(text="accuracy", type="number"),
-        mlrun.common.schemas.GrafanaColumn(text="error_count", type="number"),
-        mlrun.common.schemas.GrafanaColumn(text="drift_status", type="number"),
-        mlrun.common.schemas.GrafanaColumn(
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="endpoint_id", type="string"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="endpoint_function", type="string"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="endpoint_model", type="string"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="endpoint_model_class", type="string"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="first_request", type="time"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="last_request", type="time"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="accuracy", type="number"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="error_count", type="number"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="drift_status", type="number"
+        ),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
             text="predictions_per_second", type="number"
         ),
-        mlrun.common.schemas.GrafanaColumn(text="latency_avg_1h", type="number"),
+        mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+            text="latency_avg_1h", type="number"
+        ),
     ]
 
-    table = mlrun.common.schemas.GrafanaTable(columns=columns)
+    table = mlrun.common.schemas.model_monitoring.grafana.GrafanaTable(columns=columns)
     for endpoint in endpoint_list.endpoints:
         if (
             filter_router
             and endpoint.status.endpoint_type
-            == mlrun.common.model_monitoring.EndpointType.ROUTER
+            == mlrun.common.schemas.model_monitoring.EndpointType.ROUTER
         ):
             continue
         row = [
@@ -148,19 +166,21 @@ async def grafana_list_endpoints(
 
         if (
             endpoint.status.metrics
-            and mlrun.common.model_monitoring.EventKeyMetrics.GENERIC
+            and mlrun.common.schemas.model_monitoring.EventKeyMetrics.GENERIC
             in endpoint.status.metrics
         ):
             row.extend(
                 [
                     endpoint.status.metrics[
-                        mlrun.common.model_monitoring.EventKeyMetrics.GENERIC
+                        mlrun.common.schemas.model_monitoring.EventKeyMetrics.GENERIC
                     ][
-                        mlrun.common.model_monitoring.EventLiveStats.PREDICTIONS_PER_SECOND
+                        mlrun.common.schemas.model_monitoring.EventLiveStats.PREDICTIONS_PER_SECOND
                     ],
                     endpoint.status.metrics[
-                        mlrun.common.model_monitoring.EventKeyMetrics.GENERIC
-                    ][mlrun.common.model_monitoring.EventLiveStats.LATENCY_AVG_1H],
+                        mlrun.common.schemas.model_monitoring.EventKeyMetrics.GENERIC
+                    ][
+                        mlrun.common.schemas.model_monitoring.EventLiveStats.LATENCY_AVG_1H
+                    ],
                 ]
             )
 
@@ -197,18 +217,38 @@ async def grafana_individual_feature_analysis(
     current_stats = endpoint.status.current_stats or {}
     drift_measures = endpoint.status.drift_measures or {}
 
-    table = mlrun.common.schemas.GrafanaTable(
+    table = mlrun.common.schemas.model_monitoring.grafana.GrafanaTable(
         columns=[
-            mlrun.common.schemas.GrafanaColumn(text="feature_name", type="string"),
-            mlrun.common.schemas.GrafanaColumn(text="actual_min", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="actual_mean", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="actual_max", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="expected_min", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="expected_mean", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="expected_max", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="tvd", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="hellinger", type="number"),
-            mlrun.common.schemas.GrafanaColumn(text="kld", type="number"),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="feature_name", type="string"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="actual_min", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="actual_mean", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="actual_max", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="expected_min", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="expected_mean", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="expected_max", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="tvd", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="hellinger", type="number"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaColumn(
+                text="kld", type="number"
+            ),
         ]
     )
 
@@ -254,14 +294,26 @@ async def grafana_overall_feature_analysis(
         feature_analysis=True,
     )
 
-    table = mlrun.common.schemas.GrafanaTable(
+    table = mlrun.common.schemas.model_monitoring.grafana.GrafanaTable(
         columns=[
-            mlrun.common.schemas.GrafanaNumberColumn(text="tvd_sum"),
-            mlrun.common.schemas.GrafanaNumberColumn(text="tvd_mean"),
-            mlrun.common.schemas.GrafanaNumberColumn(text="hellinger_sum"),
-            mlrun.common.schemas.GrafanaNumberColumn(text="hellinger_mean"),
-            mlrun.common.schemas.GrafanaNumberColumn(text="kld_sum"),
-            mlrun.common.schemas.GrafanaNumberColumn(text="kld_mean"),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="tvd_sum"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="tvd_mean"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="hellinger_sum"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="hellinger_mean"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="kld_sum"
+            ),
+            mlrun.common.schemas.model_monitoring.grafana.GrafanaNumberColumn(
+                text="kld_mean"
+            ),
         ]
     )
 
@@ -339,9 +391,11 @@ async def grafana_incoming_features(
     data.index = data.index.astype(np.int64) // 10**6
 
     for feature, indexed_values in data.to_dict().items():
-        target = mlrun.common.schemas.GrafanaTimeSeriesTarget(target=feature)
+        target = mlrun.common.schemas.model_monitoring.grafana.GrafanaTimeSeriesTarget(
+            target=feature
+        )
         for index, value in indexed_values.items():
-            data_point = mlrun.common.schemas.GrafanaDataPoint(
+            data_point = mlrun.common.schemas.model_monitoring.grafana.GrafanaDataPoint(
                 value=float(value), timestamp=index
             )
             target.add_data_point(data_point)

mlrun/api/crud/model_monitoring/helpers.py

@@ -0,0 +1,159 @@
+# Copyright 2023 Iguazio
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+import json
+import math
+import typing
+
+import sqlalchemy.orm
+
+import mlrun.api
+import mlrun.common
+import mlrun.common.model_monitoring.helpers
+import mlrun.common.schemas.schedule
+import mlrun.errors
+
+Seconds = typing.NewType("Seconds", int)
+Minutes = typing.NewType("Minutes", int)
+
+_SECONDS_IN_MINUTE: Seconds = Seconds(60)
+_MINUTES_IN_HOUR: Minutes = Minutes(60)
+
+
+def seconds2minutes(seconds: Seconds) -> Minutes:
+    return Minutes(math.ceil(seconds / _SECONDS_IN_MINUTE))
+
+
+def get_batching_interval_param(intervals_list: typing.List):
+    """Convert each value in the intervals list into a float number. None
+    Values will be converted into 0.0.
+
+    param intervals_list: A list of values based on the ScheduleCronTrigger expression. Note that at the moment
+                          it supports minutes, hours, and days. e.g. [0, '*/1', None] represents on the hour
+                          every hour.
+
+    :return: A tuple of:
+             [0] = minutes interval as a float
+             [1] = hours interval as a float
+             [2] = days interval as a float
+    """
+    return tuple(
+        [
+            0.0
+            if isinstance(interval, (float, int)) or interval is None
+            else float(f"0{interval.partition('/')[-1]}")
+            for interval in intervals_list
+        ]
+    )
+
+
+def _add_minutes_offset(
+    minute: typing.Optional[typing.Union[int, str]],
+    offset: Minutes,
+) -> typing.Optional[typing.Union[int, str]]:
+    """
+    :param minute: the minute specification in the cron schedule, e.g. "0".
+    :param offset: the offset in minutes to add to the cron minute specification.
+    :return: the minute cron with the offset applied (if supported).
+    """
+    if minute and (
+        (isinstance(minute, str) and str.isdigit(minute)) or isinstance(minute, int)
+    ):
+        minute = (int(minute) + offset) % _MINUTES_IN_HOUR
+    return minute
+
+
+def convert_to_cron_string(
+    cron_trigger: mlrun.common.schemas.schedule.ScheduleCronTrigger,
+    minute_delay: Minutes = Minutes(0),
+) -> str:
+    """Convert the batch interval `ScheduleCronTrigger` into a cron trigger expression"""
+    return "{} {} {} * *".format(
+        _add_minutes_offset(cron_trigger.minute, minute_delay),
+        cron_trigger.hour,
+        cron_trigger.day,
+    ).replace("None", "*")
+
+
+def clean_feature_name(feature_name):
+    return feature_name.replace(" ", "_").replace("(", "").replace(")", "")
+
+
+def json_loads_if_not_none(field: typing.Any) -> typing.Any:
+    return (
+        json.loads(field) if field and field != "null" and field is not None else None
+    )
+
+
+def get_access_key(auth_info: mlrun.common.schemas.AuthInfo):
+    """
+    Get access key from the current data session. This method is usually used to verify that the session
+    is valid and contains an access key.
+
+    param auth_info: The auth info of the request.
+
+    :return: Access key as a string.
+    """
+    access_key = auth_info.data_session
+    if not access_key:
+        raise mlrun.errors.MLRunBadRequestError("Data session is missing")
+    return access_key
+
+
+def get_monitoring_parquet_path(
+    db_session: sqlalchemy.orm.Session, project: str
+) -> str:
+    """Get model monitoring parquet target for the current project. The parquet target path is based on the
+    project artifact path. If project artifact path is not defined, the parquet target path will be based on MLRun
+    artifact path.
+
+    :param db_session: A session that manages the current dialog with the database. Will be used in this function
+                       to get the project record from DB.
+    :param project:    Project name.
+
+    :return:           Monitoring parquet target path.
+    """
+
+    # Get the artifact path from the project record that was stored in the DB
+    project_obj = mlrun.api.crud.projects.Projects().get_project(
+        session=db_session, name=project
+    )
+    artifact_path = project_obj.spec.artifact_path
+    # Generate monitoring parquet path value
+    parquet_path = mlrun.mlconf.get_model_monitoring_file_target_path(
+        project=project,
+        kind=mlrun.common.schemas.model_monitoring.FileTargetKind.PARQUET,
+        target="offline",
+        artifact_path=artifact_path,
+    )
+    return parquet_path
+
+
+def get_stream_path(project: str = None):
+    """Get stream path from the project secret. If wasn't set, take it from the system configurations"""
+
+    stream_uri = mlrun.api.crud.secrets.Secrets().get_project_secret(
+        project=project,
+        provider=mlrun.common.schemas.secret.SecretProviderName.kubernetes,
+        allow_secrets_from_k8s=True,
+        secret_key=mlrun.common.schemas.model_monitoring.ProjectSecretKeys.STREAM_PATH,
+    ) or mlrun.mlconf.get_model_monitoring_file_target_path(
+        project=project,
+        kind=mlrun.common.schemas.model_monitoring.FileTargetKind.STREAM,
+        target="online",
+    )
+
+    return mlrun.common.model_monitoring.helpers.parse_monitoring_stream_path(
+        stream_uri=stream_uri, project=project
+    )