mlrun 1.10.0rc37__py3-none-any.whl → 1.10.0rc41__py3-none-any.whl

mlrun/artifacts/document.py

@@ -359,7 +359,12 @@ class DocumentArtifact(Artifact):
         self,
         splitter: Optional["TextSplitter"] = None,  # noqa: F821
     ) -> list["Document"]:  # noqa: F821
-        from langchain.schema import Document
+        # Try new langchain 1.0+ import path first
+        try:
+            from langchain_core.documents import Document
+        except ImportError:
+            # Fall back to old langchain <1.0 import path
+            from langchain.schema import Document
 
         """
         Create LC documents from the artifact

mlrun/common/constants.py

@@ -27,6 +27,12 @@ DASK_LABEL_PREFIX = "dask.org/"
 NUCLIO_LABEL_PREFIX = "nuclio.io/"
 RESERVED_TAG_NAME_LATEST = "latest"
 
+# Kubernetes DNS-1123 label name length limit
+K8S_DNS_1123_LABEL_MAX_LENGTH = 63
+
+
+RESERVED_BATCH_JOB_SUFFIX = "-batch"
+
 JOB_TYPE_WORKFLOW_RUNNER = "workflow-runner"
 JOB_TYPE_PROJECT_LOADER = "project-loader"
 JOB_TYPE_RERUN_WORKFLOW_RUNNER = "rerun-workflow-runner"

mlrun/common/model_monitoring/helpers.py

@@ -170,6 +170,6 @@ def log_background_task_state(
             f"Model endpoint creation task is still in progress with the current state: "
             f"{background_task_state}. Events will not be monitored for the next "
             f"{mlrun.mlconf.model_endpoint_monitoring.model_endpoint_creation_check_period} seconds",
-            function_name=server.function.name,
+            function_name=server.function_name,
             background_task_check_timestamp=background_task_check_timestamp.isoformat(),
         )

mlrun/common/schemas/model_monitoring/constants.py

@@ -486,8 +486,6 @@ class ModelMonitoringLabels:
 
 _RESERVED_FUNCTION_NAMES = MonitoringFunctionNames.list() + [SpecialApps.MLRUN_INFRA]
 
-_RESERVED_EVALUATE_FUNCTION_SUFFIX = "-batch"
-
 
 class ModelEndpointMonitoringMetricType(StrEnum):
     RESULT = "result"

mlrun/common/secrets.py

@@ -11,10 +11,31 @@
 # 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 re
 from abc import ABC, abstractmethod
 
 import mlrun.common.schemas
+from mlrun.config import config as mlconf
+
+_AUTH_SECRET_NAME_TEMPLATE = re.escape(
+    mlconf.secret_stores.kubernetes.auth_secret_name.format(
+        hashed_access_key="",
+    )
+)
+AUTH_SECRET_PATTERN = re.compile(f"^{_AUTH_SECRET_NAME_TEMPLATE}.*")
+
+
+def validate_not_forbidden_secret(secret_name: str) -> None:
+    """
+    Forbid client-supplied references to internal MLRun auth/project secrets.
+    No-op when running inside the API server (API enrichments are allowed).
+    """
+    if not secret_name or mlrun.config.is_running_as_api():
+        return
+    if AUTH_SECRET_PATTERN.match(secret_name):
+        raise mlrun.errors.MLRunInvalidArgumentError(
+            f"Forbidden secret '{secret_name}' matches MLRun auth-secret pattern."
+        )
 
 
 class SecretProviderInterface(ABC):

mlrun/launcher/local.py

@@ -243,6 +243,8 @@ class ClientLocalLauncher(launcher.ClientBaseLauncher):
 
         # if the handler has module prefix force "local" (vs "handler") runtime
         kind = "local" if isinstance(handler, str) and "." in handler else ""
+
+        # Create temporary local function for execution
         fn = mlrun.new_function(meta.name, command=command, args=args, kind=kind)
         fn.metadata = meta
         setattr(fn, "_is_run_local", True)

mlrun/model.py

@@ -29,6 +29,7 @@ import pydantic.v1.error_wrappers
 import mlrun
 import mlrun.common.constants as mlrun_constants
 import mlrun.common.schemas.notification
+import mlrun.common.secrets
 import mlrun.utils.regex
 
 from .utils import (
@@ -1616,7 +1617,12 @@ class RunTemplate(ModelObj):
 
         :returns: The RunTemplate object
         """
-
+        if kind == "azure_vault" and isinstance(source, dict):
+            candidate_secret_name = (source.get("k8s_secret") or "").strip()
+            if candidate_secret_name:
+                mlrun.common.secrets.validate_not_forbidden_secret(
+                    candidate_secret_name
+                )
         if kind == "vault" and isinstance(source, list):
             source = {"project": self.metadata.project, "secrets": source}
 

mlrun/model_monitoring/api.py

@@ -563,9 +563,10 @@ def _create_model_monitoring_function_base(
             "An application cannot have the following names: "
             f"{mm_constants._RESERVED_FUNCTION_NAMES}"
         )
-    if name and name.endswith(mm_constants._RESERVED_EVALUATE_FUNCTION_SUFFIX):
+    _, has_valid_suffix, suffix = mlrun.utils.helpers.ensure_batch_job_suffix(name)
+    if name and not has_valid_suffix:
         raise mlrun.errors.MLRunValueError(
-            "Model monitoring application names cannot end with `-batch`"
+            f"Model monitoring application names cannot end with `{suffix}`"
         )
     if func is None:
         func = ""

mlrun/model_monitoring/applications/base.py

@@ -799,10 +799,13 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
                 f"`{mm_constants.APP_NAME_REGEX.pattern}`. "
                 "Please choose another `func_name`."
             )
-        if not job_name.endswith(mm_constants._RESERVED_EVALUATE_FUNCTION_SUFFIX):
-            job_name += mm_constants._RESERVED_EVALUATE_FUNCTION_SUFFIX
+        job_name, was_renamed, suffix = mlrun.utils.helpers.ensure_batch_job_suffix(
+            job_name
+        )
+        if was_renamed:
             mlrun.utils.logger.info(
-                'Changing function name - adding `"-batch"` suffix', func_name=job_name
+                f'Changing function name - adding `"{suffix}"` suffix',
+                func_name=job_name,
             )
 
         return job_name

mlrun/model_monitoring/applications/context.py

@@ -249,6 +249,7 @@ class MonitoringApplicationContext:
                 project=self.project_name,
                 endpoint_id=self.endpoint_id,
                 feature_analysis=True,
+                tsdb_metrics=False,
             )
         return self._model_endpoint
 

mlrun/model_monitoring/db/tsdb/base.py

@@ -14,7 +14,7 @@
 
 from abc import ABC, abstractmethod
 from datetime import datetime, timedelta
-from typing import Callable, ClassVar, Literal, Optional, Union
+from typing import ClassVar, Literal, Optional, Union
 
 import pandas as pd
 import pydantic.v1
@@ -444,11 +444,9 @@ class TSDBConnector(ABC):
                                    ]
         """
 
-    async def add_basic_metrics(
+    def add_basic_metrics(
         self,
         model_endpoint_objects: list[mlrun.common.schemas.ModelEndpoint],
-        project: str,
-        run_in_threadpool: Callable,
         metric_list: Optional[list[str]] = None,
     ) -> list[mlrun.common.schemas.ModelEndpoint]:
         raise NotImplementedError()

mlrun/model_monitoring/db/tsdb/tdengine/tdengine_connector.py

@@ -14,7 +14,7 @@
 
 import threading
 from datetime import datetime, timedelta
-from typing import Callable, Final, Literal, Optional, Union
+from typing import Final, Literal, Optional, Union
 
 import pandas as pd
 import taosws
@@ -55,14 +55,12 @@ class TDEngineConnector(TSDBConnector):
     """
 
     type: str = mm_schemas.TSDBTarget.TDEngine
-    database = f"{tdengine_schemas._MODEL_MONITORING_DATABASE}_{mlrun.mlconf.system_id}"
 
     def __init__(
         self,
         project: str,
         profile: DatastoreProfile,
         timestamp_precision: TDEngineTimestampPrecision = TDEngineTimestampPrecision.MICROSECOND,
-        **kwargs,
     ):
         super().__init__(project=project)
 
@@ -72,6 +70,15 @@ class TDEngineConnector(TSDBConnector):
             timestamp_precision
         )
 
+        if not mlrun.mlconf.system_id:
+            raise mlrun.errors.MLRunInvalidArgumentError(
+                "system_id is not set in mlrun.mlconf. "
+                "TDEngineConnector requires system_id to be configured for database name construction. "
+                "Please ensure MLRun configuration is properly loaded before creating TDEngineConnector."
+            )
+        self.database = (
+            f"{tdengine_schemas._MODEL_MONITORING_DATABASE}_{mlrun.mlconf.system_id}"
+        )
         self._init_super_tables()
 
     @property
@@ -1050,7 +1057,6 @@ class TDEngineConnector(TSDBConnector):
             ]
         ):
             metric_objects = []
-
             if not df_results.empty:
                 df_results.rename(
                     columns={
@@ -1062,7 +1068,9 @@ class TDEngineConnector(TSDBConnector):
                     metric_objects.append(
                         mm_schemas.ApplicationResultRecord(
                             time=datetime.fromisoformat(
-                                row[mm_schemas.WriterEvent.END_INFER_TIME]
+                                row[mm_schemas.WriterEvent.END_INFER_TIME].replace(
+                                    " +", "+"
+                                )
                             ),
                             result_name=row[mm_schemas.ResultData.RESULT_NAME],
                             kind=row[mm_schemas.ResultData.RESULT_KIND],
@@ -1082,7 +1090,9 @@ class TDEngineConnector(TSDBConnector):
                     metric_objects.append(
                         mm_schemas.ApplicationMetricRecord(
                             time=datetime.fromisoformat(
-                                row[mm_schemas.WriterEvent.END_INFER_TIME]
+                                row[mm_schemas.WriterEvent.END_INFER_TIME].replace(
+                                    " +", "+"
+                                )
                             ),
                             metric_name=row[mm_schemas.MetricData.METRIC_NAME],
                             value=row[mm_schemas.MetricData.METRIC_VALUE],
@@ -1241,11 +1251,9 @@ class TDEngineConnector(TSDBConnector):
             df.dropna(inplace=True)
         return df
 
-    async def add_basic_metrics(
+    def add_basic_metrics(
         self,
         model_endpoint_objects: list[mlrun.common.schemas.ModelEndpoint],
-        project: str,
-        run_in_threadpool: Callable,
         metric_list: Optional[list[str]] = None,
     ) -> list[mlrun.common.schemas.ModelEndpoint]:
         """
@@ -1253,8 +1261,6 @@ class TDEngineConnector(TSDBConnector):
 
         :param model_endpoint_objects: A list of `ModelEndpoint` objects that will
                                         be filled with the relevant basic metrics.
-        :param project:                The name of the project.
-        :param run_in_threadpool:      A function that runs another function in a thread pool.
         :param metric_list:            List of metrics to include from the time series DB. Defaults to all metrics.
 
         :return: A list of `ModelEndpointMonitoringMetric` objects.

mlrun/model_monitoring/db/tsdb/v3io/v3io_connector.py

@@ -12,9 +12,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import math
-from datetime import datetime, timedelta
+from datetime import datetime, timedelta, timezone
 from io import StringIO
-from typing import Callable, Literal, Optional, Union
+from typing import Literal, Optional, Union
 
 import pandas as pd
 import v3io_frames
@@ -1230,11 +1230,9 @@ class V3IOTSDBConnector(TSDBConnector):
             )
         return df.reset_index(drop=True)
 
-    async def add_basic_metrics(
+    def add_basic_metrics(
         self,
         model_endpoint_objects: list[mlrun.common.schemas.ModelEndpoint],
-        project: str,
-        run_in_threadpool: Callable,
         metric_list: Optional[list[str]] = None,
     ) -> list[mlrun.common.schemas.ModelEndpoint]:
         """
@@ -1242,8 +1240,6 @@ class V3IOTSDBConnector(TSDBConnector):
 
         :param model_endpoint_objects: A list of `ModelEndpoint` objects that will
                                        be filled with the relevant basic metrics.
-        :param project:                The name of the project.
-        :param run_in_threadpool:      A function that runs another function in a thread pool.
         :param metric_list:            List of metrics to include from the time series DB. Defaults to all metrics.
 
         :return: A list of `ModelEndpointMonitoringMetric` objects.
@@ -1272,8 +1268,7 @@ class V3IOTSDBConnector(TSDBConnector):
             function,
             _,
         ) in metric_name_to_function_and_column_name.items():
-            metric_name_to_result[metric_name] = await run_in_threadpool(
-                function,
+            metric_name_to_result[metric_name] = function(
                 endpoint_ids=uids,
                 get_raw=True,
             )
@@ -1344,7 +1339,7 @@ class V3IOTSDBConnector(TSDBConnector):
             else:
                 filter_query = app_filter_query
 
-        df = self._get_records(
+        raw_frames: list[v3io_frames.client.RawFrame] = self._get_records(
             table=mm_schemas.V3IOTSDBTables.APP_RESULTS,
             start=start,
             end=end,
@@ -1353,39 +1348,33 @@ class V3IOTSDBConnector(TSDBConnector):
                 mm_schemas.ResultData.RESULT_STATUS,
             ],
             filter_query=filter_query,
+            get_raw=True,
         )
 
-        # filter result status
-        if result_status_list and not df.empty:
-            df = df[df[mm_schemas.ResultData.RESULT_STATUS].isin(result_status_list)]
-
-        if df.empty:
+        if not raw_frames:
             return {}
-        else:
-            # convert application name to lower case
-            df[mm_schemas.ApplicationEvent.APPLICATION_NAME] = df[
-                mm_schemas.ApplicationEvent.APPLICATION_NAME
-            ].str.lower()
-
-            df = (
-                df[
-                    [
-                        mm_schemas.ApplicationEvent.APPLICATION_NAME,
-                        mm_schemas.ResultData.RESULT_STATUS,
-                        mm_schemas.ResultData.RESULT_VALUE,
-                    ]
-                ]
-                .groupby(
-                    [
-                        mm_schemas.ApplicationEvent.APPLICATION_NAME,
-                        mm_schemas.ResultData.RESULT_STATUS,
-                    ],
-                    observed=True,
-                )
-                .count()
-            )
 
-            return df[mm_schemas.ResultData.RESULT_VALUE].to_dict()
+        # Count occurrences by (application_name, result_status) from RawFrame objects
+        count_dict = {}
+
+        for frame in raw_frames:
+            # Extract column data from each RawFrame
+            app_name = frame.column_data(mm_schemas.ApplicationEvent.APPLICATION_NAME)[
+                0
+            ]
+            statuses = frame.column_data(mm_schemas.ResultData.RESULT_STATUS)
+
+            for status in statuses:
+                # Filter by result status if specified
+                if result_status_list and status not in result_status_list:
+                    continue
+
+                # Convert application name to lower case
+                key = (app_name.lower(), status)
+
+                # Update the count in the dictionary
+                count_dict[key] = count_dict.get(key, 0) + 1
+        return count_dict
 
     def count_processed_model_endpoints(
         self,
@@ -1543,51 +1532,140 @@ class V3IOTSDBConnector(TSDBConnector):
     ) -> mm_schemas.ModelEndpointDriftValues:
         table = mm_schemas.V3IOTSDBTables.APP_RESULTS
         start, end, interval = self._prepare_aligned_start_end(start, end)
-        df = self._get_records(
+        raw_frames: list[v3io_frames.client.RawFrame] = self._get_records(
             table=table,
             start=start,
             end=end,
             columns=[mm_schemas.ResultData.RESULT_STATUS],
+            get_raw=True,
         )
-        df = self._aggregate_raw_drift_data(df, start, end, interval)
-        if df.empty:
+
+        if not raw_frames:
+            return mm_schemas.ModelEndpointDriftValues(values=[])
+
+        aggregated_data = self._aggregate_raw_drift_data(
+            raw_frames=raw_frames, start=start, end=end, interval=interval
+        )
+        if not aggregated_data:
+            return mm_schemas.ModelEndpointDriftValues(values=[])
+
+        # Filter to only include entries with max result_status >= 1
+        filtered_data = [
+            (endpoint_id, timestamp, max_status)
+            for endpoint_id, timestamp, max_status in aggregated_data
+            if max_status >= 1
+        ]
+
+        if not filtered_data:
             return mm_schemas.ModelEndpointDriftValues(values=[])
-        df = df[df[f"max({mm_schemas.ResultData.RESULT_STATUS})"] >= 1]
-        return self._df_to_drift_data(df)
+
+        return self._convert_drift_data_to_values(aggregated_data=filtered_data)
 
     @staticmethod
     def _aggregate_raw_drift_data(
-        df: pd.DataFrame, start: datetime, end: datetime, interval: str
-    ) -> pd.DataFrame:
-        if df.empty:
-            return df
-        if not isinstance(df.index, pd.DatetimeIndex):
-            raise TypeError("Expected a DatetimeIndex on the DataFrame (time index).")
-        df[EventFieldType.ENDPOINT_ID] = (
-            df[EventFieldType.ENDPOINT_ID].astype("string").str.strip()
-        )  # remove extra data carried by the category dtype
-        window = df.loc[
-            (df.index >= start) & (df.index < end),
-            [mm_schemas.ResultData.RESULT_STATUS, EventFieldType.ENDPOINT_ID],
+        raw_frames: list[v3io_frames.client.RawFrame],
+        start: datetime,
+        end: datetime,
+        interval: str,
+    ) -> list[tuple[str, datetime, float]]:
+        """
+        Aggregate raw drift data from RawFrame objects.
+
+        :param raw_frames: List of RawFrame objects containing drift data.
+        :param start:      Start datetime for filtering data.
+        :param end:        End datetime for filtering data.
+        :param interval:   Time interval string (e.g., '5min') for aggregation
+
+        :returns: list of tuples: (endpoint_id, timestamp, max_result_status)
+        """
+        if not raw_frames:
+            return []
+
+        # Parse interval to get timedelta
+        interval_td = pd.Timedelta(interval)
+
+        # Collect all data points from RawFrame objects
+        data_points = []
+        for frame in raw_frames:
+            endpoint_id = frame.column_data(EventFieldType.ENDPOINT_ID)[0]
+            result_statuses = frame.column_data(mm_schemas.ResultData.RESULT_STATUS)
+            timestamps = frame.indices()[0].times
+
+            # Combine data from this frame
+            for i, (status, timestamp) in enumerate(zip(result_statuses, timestamps)):
+                # V3IO TSDB returns timestamps in nanoseconds
+                timestamp_dt = pd.Timestamp(
+                    timestamp, unit="ns", tzinfo=timezone.utc
+                ).to_pydatetime()
+
+                # Filter by time window
+                if start <= timestamp_dt < end:
+                    data_points.append((endpoint_id, timestamp_dt, status))
+
+        if not data_points:
+            return []
+
+        # Group by endpoint_id and time intervals, then find max status
+        # Create time buckets aligned to start
+        grouped_data = {}
+        for endpoint_id, timestamp, status in data_points:
+            # Calculate which interval bucket this timestamp falls into
+            time_diff = timestamp - start
+            bucket_index = int(time_diff / interval_td)
+            bucket_start = start + (bucket_index * interval_td)
+
+            key = (endpoint_id, bucket_start)
+            if key not in grouped_data:
+                grouped_data[key] = status
+            else:
+                # Keep the maximum status value
+                grouped_data[key] = max(grouped_data[key], status)
+
+        # Convert to list of tuples
+        result = [
+            (endpoint_id, timestamp, max_status)
+            for (endpoint_id, timestamp), max_status in grouped_data.items()
         ]
-        out = (
-            window.groupby(
-                [
-                    EventFieldType.ENDPOINT_ID,
-                    pd.Grouper(
-                        freq=interval, origin=start, label="left", closed="left"
-                    ),
-                ]
-                # align to start, [start, end) intervals
-            )[mm_schemas.ResultData.RESULT_STATUS]
-            .max()
-            .reset_index()
-            .rename(
-                columns={
-                    mm_schemas.ResultData.RESULT_STATUS: f"max({mm_schemas.ResultData.RESULT_STATUS})"
+
+        return result
+
+    @staticmethod
+    def _convert_drift_data_to_values(
+        aggregated_data: list[tuple[str, datetime, float]],
+    ) -> mm_schemas.ModelEndpointDriftValues:
+        """
+        Convert aggregated drift data to ModelEndpointDriftValues format.
+
+        :param aggregated_data: List of tuples (endpoint_id, timestamp, max_result_status)
+        :return: ModelEndpointDriftValues with counts of suspected and detected per timestamp
+        """
+        suspected_val = mm_schemas.constants.ResultStatusApp.potential_detection.value
+        detected_val = mm_schemas.constants.ResultStatusApp.detected.value
+
+        # Group by timestamp and result status, then count occurrences
+        timestamp_status_counts = {}
+        for _, timestamp, max_status in aggregated_data:
+            key = (timestamp, max_status)
+            timestamp_status_counts[key] = timestamp_status_counts.get(key, 0) + 1
+
+        # Organize by timestamp with counts for suspected and detected
+        timestamp_counts = {}
+        for (timestamp, status), count in timestamp_status_counts.items():
+            if timestamp not in timestamp_counts:
+                timestamp_counts[timestamp] = {
+                    "count_suspected": 0,
+                    "count_detected": 0,
                 }
-            )
-        )
-        return out.rename(
-            columns={"time": "_wstart"}
-        )  # rename datetime column to _wstart to align with the tdengine result
+
+            if status == suspected_val:
+                timestamp_counts[timestamp]["count_suspected"] = count
+            elif status == detected_val:
+                timestamp_counts[timestamp]["count_detected"] = count
+
+        # Convert to the expected format: list of (timestamp, count_suspected, count_detected)
+        values = [
+            (timestamp, counts["count_suspected"], counts["count_detected"])
+            for timestamp, counts in sorted(timestamp_counts.items())
+        ]
+
+        return mm_schemas.ModelEndpointDriftValues(values=values)

mlrun/projects/project.py

@@ -45,6 +45,7 @@ import mlrun.common.runtimes.constants
 import mlrun.common.schemas.alert
 import mlrun.common.schemas.artifact
 import mlrun.common.schemas.model_monitoring.constants as mm_constants
+import mlrun.common.secrets
 import mlrun.datastore.datastore_profile
 import mlrun.db
 import mlrun.errors
@@ -3418,7 +3419,12 @@ class MlrunProject(ModelObj):
         self._initialized = True
         return self.spec._function_objects
 
-    def with_secrets(self, kind, source, prefix=""):
+    def with_secrets(
+        self,
+        kind,
+        source,
+        prefix="",
+    ):
         """register a secrets source (file, env or dict)
 
         read secrets from a source provider to be used in workflows, example::
@@ -3440,12 +3446,19 @@ class MlrunProject(ModelObj):
 
         This will enable access to all secrets in vault registered to the current project.
 
-        :param kind:   secret type (file, inline, env, vault)
+        :param kind:   secret type (file, inline, env, vault, azure_vault)
         :param source: secret data or link (see example)
         :param prefix: add a prefix to the keys in this source
 
         :returns: project object
         """
+        # Block using mlrun-auth-secrets.* via azure_vault's k8s_secret param (client-side only)
+        if kind == "azure_vault" and isinstance(source, dict):
+            candidate_secret_name = (source.get("k8s_secret") or "").strip()
+            if candidate_secret_name:
+                mlrun.common.secrets.validate_not_forbidden_secret(
+                    candidate_secret_name
+                )
 
         if kind == "vault" and isinstance(source, list):
             source = {"project": self.metadata.name, "secrets": source}

mlrun/run.py

@@ -118,7 +118,25 @@ 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)
-    mod_name = mlrun.utils.helpers.get_module_name_from_path(command)
+
+    source_file_path_object, working_dir_path_object = (
+        mlrun.utils.helpers.get_source_and_working_dir_paths(command)
+    )
+    if source_file_path_object.is_relative_to(working_dir_path_object):
+        mod_name = mlrun.utils.helpers.get_relative_module_name_from_path(
+            source_file_path_object, working_dir_path_object
+        )
+    elif source_file_path_object.is_relative_to(
+        pathlib.Path(tempfile.gettempdir()).resolve()
+    ):
+        mod_name = Path(command).stem
+    else:
+        raise mlrun.errors.MLRunRuntimeError(
+            f"Cannot run source file '{command}': it must be located either under the current working "
+            f"directory ('{working_dir_path_object}') or the system temporary directory ('{tempfile.gettempdir()}'). "
+            f"This is required when running with local=True."
+        )
+
     spec = imputil.spec_from_file_location(mod_name, command)
     if spec is None:
         raise OSError(f"cannot import from {command!r}")
@@ -537,6 +555,7 @@ def new_function(
 
     # make sure function name is valid
     name = mlrun.utils.helpers.normalize_name(name)
+    mlrun.utils.helpers.validate_function_name(name)
 
     runner.metadata.name = name
     runner.metadata.project = (
@@ -576,6 +595,7 @@ def new_function(
         )
 
     runner.prepare_image_for_deploy()
+
     return runner
 
 
@@ -780,6 +800,9 @@ def code_to_function(
         kind=sub_kind,
         ignored_tags=ignored_tags,
     )
+
+    mlrun.utils.helpers.validate_function_name(name)
+
     spec["spec"]["env"].append(
         {
             "name": "MLRUN_HTTPDB__NUCLIO__EXPLICIT_ACK",
@@ -832,6 +855,7 @@ def code_to_function(
         runtime.spec.build.code_origin = code_origin
         runtime.spec.build.origin_filename = filename or (name + ".ipynb")
         update_common(runtime, spec)
+
         return runtime
 
     if kind is None or kind in ["", "Function"]:
@@ -845,6 +869,7 @@ def code_to_function(
 
     if not name:
         raise ValueError("name must be specified")
+
     h = get_in(spec, "spec.handler", "").split(":")
     runtime.handler = h[0] if len(h) <= 1 else h[1]
     runtime.metadata = get_in(spec, "spec.metadata")