mlrun 1.10.0rc34__tar.gz → 1.10.0rc35__tar.gz

{mlrun-1.10.0rc34/mlrun.egg-info → mlrun-1.10.0rc35}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlrun
-Version: 1.10.0rc34
+Version: 1.10.0rc35
 Summary: Tracking and config of machine learning runs
 Home-page: https://github.com/mlrun/mlrun
 Author: Yaron Haviv
@@ -105,7 +105,7 @@ Requires-Dist: snowflake-connector-python~=3.7; extra == "snowflake"
 Provides-Extra: dev-postgres
 Requires-Dist: pytest-mock-resources[postgres]~=2.12; extra == "dev-postgres"
 Provides-Extra: kfp18
-Requires-Dist: mlrun_pipelines_kfp_v1_8[kfp]>=0.5.7; extra == "kfp18"
+Requires-Dist: mlrun_pipelines_kfp_v1_8[kfp]~=0.5.8; extra == "kfp18"
 Provides-Extra: api
 Requires-Dist: uvicorn~=0.32.1; extra == "api"
 Requires-Dist: dask-kubernetes~=0.11.0; extra == "api"

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/dependencies.py

@@ -87,7 +87,7 @@ def extra_requirements() -> dict[str, list[str]]:
     extras_require.update(
         {
             "dev-postgres": ["pytest-mock-resources[postgres]~=2.12"],
-            "kfp18": ["mlrun_pipelines_kfp_v1_8[kfp]>=0.5.7"],
+            "kfp18": ["mlrun_pipelines_kfp_v1_8[kfp]~=0.5.8"],
             # TODO uncomment when KFP 1.8 support is removed
             # "kfp2": ["mlrun_pipelines_kfp_v2[kfp]>=0.5.0 ; python_version >= '3.11'"],
             "api": api_deps,

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/common/schemas/__init__.py

@@ -154,6 +154,7 @@ from .model_monitoring import (
     ModelEndpointSchema,
     ModelEndpointSpec,
     ModelEndpointStatus,
+    ModelMonitoringInfraLabel,
     ModelMonitoringMode,
     MonitoringFunctionNames,
     TSDBTarget,

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/common/schemas/model_monitoring/__init__.py

@@ -30,6 +30,7 @@ from .constants import (
     ModelEndpointMonitoringMetricType,
     ModelEndpointSchema,
     ModelMonitoringAppLabel,
+    ModelMonitoringInfraLabel,
     ModelMonitoringMode,
     MonitoringFunctionNames,
     PredictionsQueryConstants,

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/model_monitoring/applications/base.py

@@ -332,36 +332,11 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
             project=project,
         ) as (endpoints_output, application_schedules):
 
-            def call_do_tracking(event: Optional[dict] = None):
+            def call_do_tracking(
+                monitoring_context: mm_context.MonitoringApplicationContext,
+            ):
                 nonlocal endpoints_output
 
-                if event is None:
-                    event = {}
-                monitoring_context = (
-                    mm_context.MonitoringApplicationContext._from_ml_ctx(
-                        event=event,
-                        application_name=application_name,
-                        context=context,
-                        project=project,
-                        sample_df=sample_data,
-                        feature_stats=feature_stats,
-                    )
-                )
-
-                if (
-                    monitoring_context.endpoint_id
-                    and monitoring_context.sample_df.empty
-                ):
-                    # The current sample is empty
-                    context.logger.warning(
-                        "No sample data available for tracking",
-                        application_name=application_name,
-                        endpoint_id=monitoring_context.endpoint_id,
-                        start_time=monitoring_context.start_infer_time,
-                        end_time=monitoring_context.end_infer_time,
-                    )
-                    return
-
                 result = self.do_tracking(monitoring_context)
                 endpoints_output[monitoring_context.endpoint_id].append(
                     (monitoring_context, result)
@@ -369,7 +344,7 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
                 return result
 
             if endpoints is not None:
-                resolved_endpoints = self._handle_endpoints_type_evaluate(
+                resolved_endpoints = self._validate_endpoints(
                     project=project, endpoints=endpoints
                 )
                 if (
@@ -391,26 +366,24 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
                         application_schedules=application_schedules,
                     )
                 for endpoint_name, endpoint_id in resolved_endpoints:
-                    for window_start, window_end in self._window_generator(
+                    for monitoring_ctx in self._window_generator(
                         start=start,
                         end=end,
                         base_period=base_period,
                         application_schedules=application_schedules,
                         endpoint_id=endpoint_id,
+                        endpoint_name=endpoint_name,
                         application_name=application_name,
                         existing_data_handling=existing_data_handling,
+                        sample_data=sample_data,
+                        context=context,
+                        project=project,
                     ):
-                        result = call_do_tracking(
-                            event={
-                                mm_constants.ApplicationEvent.ENDPOINT_NAME: endpoint_name,
-                                mm_constants.ApplicationEvent.ENDPOINT_ID: endpoint_id,
-                                mm_constants.ApplicationEvent.START_INFER_TIME: window_start,
-                                mm_constants.ApplicationEvent.END_INFER_TIME: window_end,
-                            }
-                        )
+                        result = call_do_tracking(monitoring_ctx)
                         result_key = (
-                            f"{endpoint_name}-{endpoint_id}_{window_start.isoformat()}_{window_end.isoformat()}"
-                            if window_start and window_end
+                            f"{endpoint_name}-{endpoint_id}_{monitoring_ctx.start_infer_time.isoformat()}_{monitoring_ctx.end_infer_time.isoformat()}"
+                            if monitoring_ctx.start_infer_time
+                            and monitoring_ctx.end_infer_time
                             else f"{endpoint_name}-{endpoint_id}"
                         )
 
@@ -418,10 +391,38 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
                             result_key, self._flatten_data_result(result)
                         )
             else:
-                return self._flatten_data_result(call_do_tracking())
+                result = call_do_tracking(
+                    mm_context.MonitoringApplicationContext._from_ml_ctx(
+                        context=context,
+                        project=project,
+                        application_name=application_name,
+                        event={},
+                        sample_df=sample_data,
+                        feature_stats=feature_stats,
+                    )
+                )
+                return self._flatten_data_result(result)
 
     @staticmethod
-    def _handle_endpoints_type_evaluate(
+    def _check_endpoints_first_request(
+        endpoints: list[mlrun.common.schemas.ModelEndpoint],
+    ) -> None:
+        """Make sure that all the endpoints have had at least one request"""
+        endpoints_no_requests = [
+            (endpoint.metadata.name, endpoint.metadata.uid)
+            for endpoint in endpoints
+            if not endpoint.status.first_request
+        ]
+        if endpoints_no_requests:
+            raise mlrun.errors.MLRunValueError(
+                "The following model endpoints have not had any requests yet and "
+                "have no data, cannot run the model monitoring application on them: "
+                f"{endpoints_no_requests}"
+            )
+
+    @classmethod
+    def _validate_endpoints(
+        cls,
         project: "mlrun.MlrunProject",
         endpoints: Union[
             list[tuple[str, str]], list[list[str]], list[str], Literal["all"]
@@ -457,6 +458,9 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
         endpoints_list = project.list_model_endpoints(
             names=endpoint_names, latest_only=True
         ).endpoints
+
+        cls._check_endpoints_first_request(endpoints_list)
+
         if endpoints_list:
             list_endpoints_result = [
                 (endpoint.metadata.name, endpoint.metadata.uid)
@@ -596,13 +600,51 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
         application_schedules: Optional[
             mm_schedules.ModelMonitoringSchedulesFileApplication
         ],
+        endpoint_name: str,
         endpoint_id: str,
         application_name: str,
         existing_data_handling: ExistingDataHandling,
-    ) -> Iterator[tuple[Optional[datetime], Optional[datetime]]]:
+        context: "mlrun.MLClientCtx",
+        project: "mlrun.MlrunProject",
+        sample_data: Optional[pd.DataFrame],
+    ) -> Iterator[mm_context.MonitoringApplicationContext]:
+        def yield_monitoring_ctx(
+            window_start: Optional[datetime], window_end: Optional[datetime]
+        ) -> Iterator[mm_context.MonitoringApplicationContext]:
+            ctx = mm_context.MonitoringApplicationContext._from_ml_ctx(
+                event={
+                    mm_constants.ApplicationEvent.ENDPOINT_NAME: endpoint_name,
+                    mm_constants.ApplicationEvent.ENDPOINT_ID: endpoint_id,
+                    mm_constants.ApplicationEvent.START_INFER_TIME: window_start,
+                    mm_constants.ApplicationEvent.END_INFER_TIME: window_end,
+                },
+                application_name=application_name,
+                context=context,
+                project=project,
+                sample_df=sample_data,
+            )
+
+            if ctx.sample_df.empty:
+                # The current sample is empty
+                context.logger.debug(
+                    "No sample data available for tracking",
+                    application_name=application_name,
+                    endpoint_id=ctx.endpoint_id,
+                    start_time=ctx.start_infer_time,
+                    end_time=ctx.end_infer_time,
+                )
+                return
+
+            yield ctx
+
+            if application_schedules and window_end:
+                application_schedules.update_endpoint_last_analyzed(
+                    endpoint_uid=endpoint_id, last_analyzed=window_end
+                )
+
         if start is None or end is None:
             # A single window based on the `sample_data` input - see `_handler`.
-            yield None, None
+            yield from yield_monitoring_ctx(None, None)
             return
 
         start_dt = datetime.fromisoformat(start)
@@ -630,11 +672,7 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
             )
 
         if base_period is None:
-            yield start_dt, end_dt
-            if application_schedules:
-                application_schedules.update_endpoint_last_analyzed(
-                    endpoint_uid=endpoint_id, last_analyzed=end_dt
-                )
+            yield from yield_monitoring_ctx(start_dt, end_dt)
             return
 
         window_length = cls._validate_and_get_window_length(
@@ -644,11 +682,7 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
         current_start_time = start_dt
         while current_start_time < end_dt:
             current_end_time = min(current_start_time + window_length, end_dt)
-            yield current_start_time, current_end_time
-            if application_schedules:
-                application_schedules.update_endpoint_last_analyzed(
-                    endpoint_uid=endpoint_id, last_analyzed=current_end_time
-                )
+            yield from yield_monitoring_ctx(current_start_time, current_end_time)
             current_start_time = current_end_time
 
     @classmethod

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/run.py

@@ -17,6 +17,7 @@ import json
 import os
 import pathlib
 import socket
+import sys
 import tempfile
 import time
 import typing
@@ -117,14 +118,13 @@ def function_to_module(code="", workdir=None, secrets=None, silent=False):
         raise ValueError("nothing to run, specify command or function")
 
     command = os.path.join(workdir or "", command)
-    path = Path(command)
-    mod_name = path.name
-    if path.suffix:
-        mod_name = mod_name[: -len(path.suffix)]
+    mod_name = mlrun.utils.helpers.get_module_name_from_path(command)
     spec = imputil.spec_from_file_location(mod_name, command)
     if spec is None:
         raise OSError(f"cannot import from {command!r}")
     mod = imputil.module_from_spec(spec)
+    # add to system modules, which can be necessary when running in a MockServer (ML-10937)
+    sys.modules[mod_name] = mod
     spec.loader.exec_module(mod)
 
     return mod
@@ -413,21 +413,29 @@ def import_function_to_dict(
             with open(code_file, "wb") as fp:
                 fp.write(code)
         elif cmd:
-            if not path.isfile(code_file):
-                slash_index = url.rfind("/")
-                if slash_index < 0:
-                    raise ValueError(f"no file in exec path (spec.command={code_file})")
-                base_dir = os.path.normpath(url[: slash_index + 1])
-                candidate_path = _ensure_path_confined_to_base_dir(
-                    base_directory=base_dir,
-                    relative_path=code_file,
-                    error_message_on_escape=f"exec file spec.command={code_file} is outside of allowed directory",
-                )
-                if path.isfile(candidate_path):
-                    raise ValueError(
-                        f"exec file spec.command={code_file} is relative, change working dir"
-                    )
+            slash_index = url.rfind("/")
+            if slash_index < 0:
+                raise ValueError(f"no file in exec path (spec.command={code_file})")
+            base_dir = os.path.normpath(url[: slash_index + 1])
+
+            # Validate and resolve the candidate path before checking existence
+            candidate_path = _ensure_path_confined_to_base_dir(
+                base_directory=base_dir,
+                relative_path=code_file,
+                error_message_on_escape=(
+                    f"exec file spec.command={code_file} is outside of allowed directory"
+                ),
+            )
+
+            # Only now it's safe to check file existence
+            if not path.isfile(candidate_path):
                 raise ValueError(f"no file in exec path (spec.command={code_file})")
+
+            # Check that the path is absolute
+            if not os.path.isabs(code_file):
+                raise ValueError(
+                    f"exec file spec.command={code_file} is relative, it must be absolute. Change working dir"
+                )
         else:
             raise ValueError("command or code not specified in function spec")
 

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/runtimes/nuclio/application/application.py

@@ -568,6 +568,12 @@ class ApplicationRuntime(RemoteRuntime):
                 "Authentication credentials not provided"
             )
 
+        if direct_port_access and port:
+            logger.warning(
+                "Ignoring 'port' because 'direct_port_access' is enabled. "
+                "The 'port' setting is only applicable when 'direct_port_access' is disabled."
+            )
+
         ports = (
             port or self.spec.internal_application_port if direct_port_access else []
         )

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/serving/server.py

@@ -20,7 +20,6 @@ import copy
 import importlib
 import json
 import os
-import pathlib
 import socket
 import traceback
 import uuid
@@ -51,7 +50,7 @@ from ..datastore.store_resources import ResourceCache
 from ..errors import MLRunInvalidArgumentError
 from ..execution import MLClientCtx
 from ..model import ModelObj
-from ..utils import get_caller_globals
+from ..utils import get_caller_globals, get_module_name_from_path
 from .states import (
     FlowStep,
     MonitoredStep,
@@ -598,17 +597,7 @@ async def async_execute_graph(
         #  gets set in local flow and not just in the remote pod
         source_file_path = spec.get("filename", None)
         if source_file_path:
-            source_file_path_object = pathlib.Path(source_file_path).resolve()
-            current_dir_path_object = pathlib.Path(".").resolve()
-            if not source_file_path_object.is_relative_to(current_dir_path_object):
-                raise mlrun.errors.MLRunRuntimeError(
-                    f"Source file path '{source_file_path}' is not under the current working directory "
-                    f"(which is required when running with local=True)"
-                )
-            relative_path_to_source_file = source_file_path_object.relative_to(
-                current_dir_path_object
-            )
-            modname = ".".join(relative_path_to_source_file.with_suffix("").parts)
+            modname = get_module_name_from_path(source_file_path)
 
     namespace = {}
     if modname:

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun/utils/helpers.py

@@ -21,6 +21,7 @@ import inspect
 import itertools
 import json
 import os
+import pathlib
 import re
 import string
 import sys
@@ -923,11 +924,23 @@ def enrich_image_url(
     )
     mlrun_version = config.images_tag or client_version or server_version
     tag = mlrun_version or ""
-    tag += resolve_image_tag_suffix(
-        mlrun_version=mlrun_version,
-        python_version=client_python_version,
+
+    # starting mlrun 1.10.0-rc0 we want to enrich the kfp image with the python version
+    # e.g for 1.9 we have a single mlrun-kfp image that supports only python 3.9
+    enrich_kfp_python_version = (
+        "mlrun-kfp" in image_url
+        and mlrun_version
+        and semver.VersionInfo.is_valid(mlrun_version)
+        and semver.VersionInfo.parse(mlrun_version)
+        >= semver.VersionInfo.parse("1.10.0-rc0")
     )
 
+    if "mlrun-kfp" not in image_url or enrich_kfp_python_version:
+        tag += resolve_image_tag_suffix(
+            mlrun_version=mlrun_version,
+            python_version=client_python_version,
+        )
+
     # it's an mlrun image if the repository is mlrun
     is_mlrun_image = image_url.startswith("mlrun/") or "/mlrun/" in image_url
     if ":" in image_url:
@@ -2408,3 +2421,17 @@ def set_data_by_path(
         raise mlrun.errors.MLRunInvalidArgumentError(
             "Expected path to be of type str or list of str"
         )
+
+
+def get_module_name_from_path(source_file_path: str) -> str:
+    source_file_path_object = pathlib.Path(source_file_path).resolve()
+    current_dir_path_object = pathlib.Path(".").resolve()
+    if not source_file_path_object.is_relative_to(current_dir_path_object):
+        raise mlrun.errors.MLRunRuntimeError(
+            f"Source file path '{source_file_path}' is not under the current working directory "
+            f"(which is required when running with local=True)"
+        )
+    relative_path_to_source_file = source_file_path_object.relative_to(
+        current_dir_path_object
+    )
+    return ".".join(relative_path_to_source_file.with_suffix("").parts)

mlrun-1.10.0rc35/mlrun/utils/version/version.json

@@ -0,0 +1,4 @@
+{
+  "git_commit": "6ab7c319ab2182ce9f0594c1a6c0be3c98175ebe",
+  "version": "1.10.0-rc35"
+}

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35/mlrun.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlrun
-Version: 1.10.0rc34
+Version: 1.10.0rc35
 Summary: Tracking and config of machine learning runs
 Home-page: https://github.com/mlrun/mlrun
 Author: Yaron Haviv
@@ -105,7 +105,7 @@ Requires-Dist: snowflake-connector-python~=3.7; extra == "snowflake"
 Provides-Extra: dev-postgres
 Requires-Dist: pytest-mock-resources[postgres]~=2.12; extra == "dev-postgres"
 Provides-Extra: kfp18
-Requires-Dist: mlrun_pipelines_kfp_v1_8[kfp]>=0.5.7; extra == "kfp18"
+Requires-Dist: mlrun_pipelines_kfp_v1_8[kfp]~=0.5.8; extra == "kfp18"
 Provides-Extra: api
 Requires-Dist: uvicorn~=0.32.1; extra == "api"
 Requires-Dist: dask-kubernetes~=0.11.0; extra == "api"

{mlrun-1.10.0rc34 → mlrun-1.10.0rc35}/mlrun.egg-info/requires.txt

@@ -231,7 +231,7 @@ kafka-python~=2.1.0
 avro~=1.11
 
 [kfp18]
-mlrun_pipelines_kfp_v1_8[kfp]>=0.5.7
+mlrun_pipelines_kfp_v1_8[kfp]~=0.5.8
 
 [mlflow]
 mlflow~=2.22

mlrun-1.10.0rc34/mlrun/utils/version/version.json

@@ -1,4 +0,0 @@
-{
-  "git_commit": "0d61caa563d5a76229ea18f1116fef794cd1058a",
-  "version": "1.10.0-rc34"
-}