mlrun 1.10.0rc3__py3-none-any.whl → 1.10.0rc5__py3-none-any.whl

mlrun/common/schemas/project.py

@@ -16,7 +16,6 @@ import datetime
 import typing
 
 import pydantic.v1
-from deprecated import deprecated
 
 import mlrun.common.types
 
@@ -24,22 +23,6 @@ from .common import ImageBuilder
 from .object import ObjectKind, ObjectStatus
 
 
-@deprecated(
-    version="1.7.0",
-    reason="mlrun.common.schemas.ProjectsFormat is deprecated and will be removed in 1.10.0. "
-    "Use mlrun.common.formatters.ProjectFormat instead.",
-    category=FutureWarning,
-)
-class ProjectsFormat(mlrun.common.types.StrEnum):
-    full = "full"
-    name_only = "name_only"
-    # minimal format removes large fields from the response (e.g. functions, workflows, artifacts)
-    # and is used for faster response times (in the UI)
-    minimal = "minimal"
-    # internal - allowed only in follower mode, only for the leader for upgrade purposes
-    leader = "leader"
-
-
 class ProjectMetadata(pydantic.v1.BaseModel):
     name: str
     created: typing.Optional[datetime.datetime] = None

mlrun/common/schemas/runs.py

@@ -15,26 +15,9 @@
 import typing
 
 import pydantic.v1
-from deprecated import deprecated
-
-import mlrun.common.types
 
 
 class RunIdentifier(pydantic.v1.BaseModel):
     kind: typing.Literal["run"] = "run"
     uid: typing.Optional[str]
     iter: typing.Optional[int]
-
-
-@deprecated(
-    version="1.7.0",
-    reason="mlrun.common.schemas.RunsFormat is deprecated and will be removed in 1.10.0. "
-    "Use mlrun.common.formatters.RunFormat instead.",
-    category=FutureWarning,
-)
-class RunsFormat(mlrun.common.types.StrEnum):
-    # No enrichment, data is pulled as-is from the database.
-    standard = "standard"
-
-    # Performs run enrichment, including the run's artifacts. Only available for the `get` run API.
-    full = "full"

mlrun/config.py

@@ -234,7 +234,10 @@ default_config = {
                 "model_endpoint_creation": "600",
                 "model_endpoint_tsdb_leftovers": "900",
             },
-            "runtimes": {"dask": "600"},
+            "runtimes": {
+                "dask": "600",
+                "dask_cluster_start": "300",
+            },
             "push_notifications": "60",
         },
     },
@@ -482,7 +485,7 @@ default_config = {
             "project_owners_cache_ttl": "30 seconds",
             # access key to be used when the leader is iguazio and polling is done from it
             "iguazio_access_key": "",
-            "iguazio_list_projects_default_page_size": 500,
+            "iguazio_list_projects_default_page_size": 200,
             "iguazio_client_job_cache_ttl": "20 minutes",
             "nuclio_project_deletion_verification_timeout": "300 seconds",
             "nuclio_project_deletion_verification_interval": "5 seconds",

mlrun/datastore/base.py

@@ -104,10 +104,10 @@ class DataStore:
         """Whether the data store supports isdir"""
         return True
 
-    def _get_secret_or_env(self, key, default=None, prefix=None):
+    def _get_secret_or_env(self, key, default=None):
         # Project-secrets are mounted as env variables whose name can be retrieved from SecretsStore
         return mlrun.get_secret_or_env(
-            key, secret_provider=self._get_secret, default=default, prefix=prefix
+            key, secret_provider=self._get_secret, default=default
         )
 
     def get_storage_options(self):

mlrun/datastore/datastore.py

@@ -54,7 +54,7 @@ def parse_url(url):
     return schema, endpoint, parsed_url
 
 
-def schema_to_store(schema):
+def schema_to_store(schema) -> DataStore.__subclasses__():
     # import store classes inside to enable making their dependencies optional (package extras)
 
     if not schema or schema in get_local_file_schema():

mlrun/datastore/datastore_profile.py

@@ -141,19 +141,11 @@ class ConfigProfile(DatastoreProfile):
 class DatastoreProfileKafkaTarget(DatastoreProfile):
     type: str = pydantic.v1.Field("kafka_target")
     _private_attributes = "kwargs_private"
-    brokers: typing.Optional[str] = None
+    brokers: str
     topic: str
     kwargs_public: typing.Optional[dict]
     kwargs_private: typing.Optional[dict]
 
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-
-        if not self.brokers:
-            raise mlrun.errors.MLRunInvalidArgumentError(
-                "DatastoreProfileKafkaTarget requires the 'brokers' field to be set"
-            )
-
     def get_topic(self) -> typing.Optional[str]:
         return self.topic
 

mlrun/datastore/redis.py

@@ -48,9 +48,8 @@ class RedisStore(DataStore):
             raise mlrun.errors.MLRunInvalidArgumentError(
                 "Provide Redis username and password only via secrets"
             )
-        credentials_prefix = self._get_secret_or_env("CREDENTIALS_PREFIX")
-        user = self._get_secret_or_env("REDIS_USER", "", credentials_prefix)
-        password = self._get_secret_or_env("REDIS_PASSWORD", "", credentials_prefix)
+        user = self._get_secret_or_env("REDIS_USER", "")
+        password = self._get_secret_or_env("REDIS_PASSWORD", "")
         host = parsed_endpoint.hostname
         port = parsed_endpoint.port if parsed_endpoint.port else redis_default_port
         schema = parsed_endpoint.scheme

mlrun/datastore/sources.py

@@ -768,7 +768,6 @@ class SnowflakeSource(BaseSourceDriver):
     :parameter url: URL of the snowflake cluster
     :parameter user: snowflake user
     :parameter database: snowflake database
-    :parameter schema: snowflake schema - deprecated, use db_schema
     :parameter db_schema: snowflake schema
     :parameter warehouse: snowflake warehouse
     """
@@ -790,18 +789,10 @@ class SnowflakeSource(BaseSourceDriver):
         url: Optional[str] = None,
         user: Optional[str] = None,
         database: Optional[str] = None,
-        schema: Optional[str] = None,
         db_schema: Optional[str] = None,
         warehouse: Optional[str] = None,
         **kwargs,
     ):
-        # TODO: Remove in 1.10.0
-        if schema:
-            warnings.warn(
-                "schema is deprecated in 1.7.0, and will be removed in 1.10.0, please use db_schema"
-            )
-        db_schema = db_schema or schema  # TODO: Remove in 1.10.0
-
         attributes = attributes or {}
         if url:
             attributes["url"] = url

mlrun/datastore/storeytargets.py

@@ -152,9 +152,7 @@ class KafkaStoreyTarget(storey.KafkaTarget):
                 parsed.path.strip("/") if parsed.path else datastore_profile.get_topic()
             )
         else:
-            brokers = attributes.pop(
-                "brokers", attributes.pop("bootstrap_servers", None)
-            )
+            brokers = attributes.pop("brokers", None)
             topic, brokers = parse_kafka_url(path, brokers)
 
         if not topic:
@@ -175,8 +173,7 @@ class RedisNoSqlStoreyTarget(storey.NoSqlTarget):
     def __init__(self, *args, **kwargs):
         path = kwargs.pop("path")
         endpoint, uri = mlrun.datastore.targets.RedisNoSqlTarget.get_server_endpoint(
-            path,
-            kwargs.pop("credentials_prefix", None),
+            path
         )
         kwargs["path"] = endpoint + "/" + uri
         super().__init__(*args, **kwargs)

mlrun/datastore/targets.py

@@ -17,7 +17,6 @@ import os
 import random
 import sys
 import time
-import warnings
 from collections import Counter
 from copy import copy
 from typing import Any, Optional, Union
@@ -409,7 +408,6 @@ class BaseStoreTarget(DataTargetBase):
         flush_after_seconds: Optional[int] = None,
         storage_options: Optional[dict[str, str]] = None,
         schema: Optional[dict[str, Any]] = None,
-        credentials_prefix=None,
     ):
         super().__init__(
             self.kind,
@@ -424,7 +422,6 @@ class BaseStoreTarget(DataTargetBase):
             max_events,
             flush_after_seconds,
             schema=schema,
-            credentials_prefix=credentials_prefix,
         )
 
         self.name = name or self.kind
@@ -440,13 +437,6 @@ class BaseStoreTarget(DataTargetBase):
         self.flush_after_seconds = flush_after_seconds
         self.storage_options = storage_options
         self.schema = schema or {}
-        self.credentials_prefix = credentials_prefix
-        if credentials_prefix:
-            warnings.warn(
-                "The 'credentials_prefix' parameter is deprecated in 1.7.0 and will be removed in "
-                "1.10.0. Please use datastore profiles instead.",
-                FutureWarning,
-            )
 
         self._target = None
         self._resource = None
@@ -457,18 +447,11 @@ class BaseStoreTarget(DataTargetBase):
             key,
             secret_provider=self._secrets,
             default=default_value,
-            prefix=self.credentials_prefix,
         )
 
     def _get_store_and_path(self):
-        credentials_prefix_secrets = (
-            {"CREDENTIALS_PREFIX": self.credentials_prefix}
-            if self.credentials_prefix
-            else None
-        )
         store, resolved_store_path, url = mlrun.store_manager.get_or_create_store(
-            self.get_target_path(),
-            credentials_prefix_secrets,
+            self.get_target_path()
         )
         return store, resolved_store_path, url
 
@@ -621,7 +604,6 @@ class BaseStoreTarget(DataTargetBase):
         driver.path = spec.path
         driver.attributes = spec.attributes
         driver.schema = spec.schema
-        driver.credentials_prefix = spec.credentials_prefix
 
         if hasattr(spec, "columns"):
             driver.columns = spec.columns
@@ -638,7 +620,6 @@ class BaseStoreTarget(DataTargetBase):
         driver.max_events = spec.max_events
         driver.flush_after_seconds = spec.flush_after_seconds
         driver.storage_options = spec.storage_options
-        driver.credentials_prefix = spec.credentials_prefix
 
         driver._resource = resource
         driver.run_id = spec.run_id
@@ -720,7 +701,6 @@ class BaseStoreTarget(DataTargetBase):
         target.key_bucketing_number = self.key_bucketing_number
         target.partition_cols = self.partition_cols
         target.time_partitioning_granularity = self.time_partitioning_granularity
-        target.credentials_prefix = self.credentials_prefix
 
         self._resource.status.update_target(target)
         return target
@@ -1213,7 +1193,6 @@ class SnowflakeTarget(BaseStoreTarget):
         flush_after_seconds: Optional[int] = None,
         storage_options: Optional[dict[str, str]] = None,
         schema: Optional[dict[str, Any]] = None,
-        credentials_prefix=None,
         url: Optional[str] = None,
         user: Optional[str] = None,
         db_schema: Optional[str] = None,
@@ -1249,7 +1228,6 @@ class SnowflakeTarget(BaseStoreTarget):
             flush_after_seconds=flush_after_seconds,
             storage_options=storage_options,
             schema=schema,
-            credentials_prefix=credentials_prefix,
         )
 
     def get_spark_options(self, key_column=None, timestamp_key=None, overwrite=True):
@@ -1488,7 +1466,7 @@ class RedisNoSqlTarget(NoSqlBaseTarget):
     writer_step_name = "RedisNoSqlTarget"
 
     @staticmethod
-    def get_server_endpoint(path, credentials_prefix=None):
+    def get_server_endpoint(path):
         endpoint, uri = parse_path(path)
         endpoint = endpoint or mlrun.mlconf.redis.url
         if endpoint.startswith("ds://"):
@@ -1506,15 +1484,8 @@ class RedisNoSqlTarget(NoSqlBaseTarget):
                 raise mlrun.errors.MLRunInvalidArgumentError(
                     "Provide Redis username and password only via secrets"
                 )
-            credentials_prefix = credentials_prefix or mlrun.get_secret_or_env(
-                key="CREDENTIALS_PREFIX"
-            )
-            user = mlrun.get_secret_or_env(
-                "REDIS_USER", default="", prefix=credentials_prefix
-            )
-            password = mlrun.get_secret_or_env(
-                "REDIS_PASSWORD", default="", prefix=credentials_prefix
-            )
+            user = mlrun.get_secret_or_env("REDIS_USER", default="")
+            password = mlrun.get_secret_or_env("REDIS_PASSWORD", default="")
             host = parsed_endpoint.hostname
             port = parsed_endpoint.port if parsed_endpoint.port else "6379"
             scheme = parsed_endpoint.scheme
@@ -1528,9 +1499,7 @@ class RedisNoSqlTarget(NoSqlBaseTarget):
         from storey import Table
         from storey.redis_driver import RedisDriver
 
-        endpoint, uri = self.get_server_endpoint(
-            self.get_target_path(), self.credentials_prefix
-        )
+        endpoint, uri = self.get_server_endpoint(self.get_target_path())
 
         return Table(
             uri,
@@ -1539,9 +1508,7 @@ class RedisNoSqlTarget(NoSqlBaseTarget):
         )
 
     def get_spark_options(self, key_column=None, timestamp_key=None, overwrite=True):
-        endpoint, uri = self.get_server_endpoint(
-            self.get_target_path(), self.credentials_prefix
-        )
+        endpoint, uri = self.get_server_endpoint(self.get_target_path())
         parsed_endpoint = urlparse(endpoint)
         store, path_in_store, path = self._get_store_and_path()
         return {
@@ -1592,7 +1559,6 @@ class RedisNoSqlTarget(NoSqlBaseTarget):
             class_name="mlrun.datastore.storeytargets.RedisNoSqlStoreyTarget",
             columns=column_list,
             table=table,
-            credentials_prefix=self.credentials_prefix,
             **self.attributes,
         )
 
@@ -1648,7 +1614,6 @@ class KafkaTarget(BaseStoreTarget):
     :param path:                topic name e.g. "my_topic"
     :param after_step:          optional, after what step in the graph to add the target
     :param columns:             optional, which columns from data to write
-    :param bootstrap_servers:   Deprecated. Use the brokers parameter instead
     :param producer_options:    additional configurations for kafka producer
     :param brokers:             kafka broker as represented by a host:port pair, or a list of kafka brokers, e.g.
         "localhost:9092", or ["kafka-broker-1:9092", "kafka-broker-2:9092"]
@@ -1664,27 +1629,12 @@ class KafkaTarget(BaseStoreTarget):
     def __init__(
         self,
         *args,
-        bootstrap_servers=None,
         producer_options=None,
         brokers=None,
         **kwargs,
     ):
         attrs = {}
 
-        # TODO: Remove this in 1.10.0
-        if bootstrap_servers:
-            if brokers:
-                raise mlrun.errors.MLRunInvalidArgumentError(
-                    "KafkaTarget cannot be created with both the 'brokers' parameter and the deprecated "
-                    "'bootstrap_servers' parameter. Please use 'brokers' only."
-                )
-            warnings.warn(
-                "'bootstrap_servers' parameter is deprecated in 1.7.0 and will be removed in 1.10.0, "
-                "use 'brokers' instead.",
-                FutureWarning,
-            )
-            brokers = bootstrap_servers
-
         if brokers:
             attrs["brokers"] = brokers
         if producer_options is not None:

mlrun/datastore/utils.py

@@ -16,7 +16,6 @@ import math
 import tarfile
 import tempfile
 import typing
-import warnings
 from urllib.parse import parse_qs, urlparse
 
 import pandas as pd
@@ -171,16 +170,7 @@ def _generate_sql_query_with_time_filter(
 def get_kafka_brokers_from_dict(options: dict, pop=False) -> typing.Optional[str]:
     get_or_pop = options.pop if pop else options.get
     kafka_brokers = get_or_pop("kafka_brokers", None)
-    if kafka_brokers:
-        return kafka_brokers
-    kafka_bootstrap_servers = get_or_pop("kafka_bootstrap_servers", None)
-    if kafka_bootstrap_servers:
-        warnings.warn(
-            "The 'kafka_bootstrap_servers' parameter is deprecated in 1.7.0 and will be removed in "
-            "1.10.0. Please pass the 'kafka_brokers' parameter instead.",
-            FutureWarning,
-        )
-    return kafka_bootstrap_servers
+    return kafka_brokers
 
 
 def transform_list_filters_to_tuple(additional_filters):

mlrun/db/base.py

@@ -185,6 +185,7 @@ class RunDBInterface(ABC):
         kind: Optional[str] = None,
         category: Union[str, mlrun.common.schemas.ArtifactCategories] = None,
         tree: Optional[str] = None,
+        parent: Optional[str] = None,
         format_: mlrun.common.formatters.ArtifactFormat = mlrun.common.formatters.ArtifactFormat.full,
         limit: Optional[int] = None,
         partition_by: Optional[

mlrun/db/httpdb.py

@@ -1225,6 +1225,7 @@ class HTTPRunDB(RunDBInterface):
         category: Union[str, mlrun.common.schemas.ArtifactCategories] = None,
         tree: Optional[str] = None,
         producer_uri: Optional[str] = None,
+        parent: Optional[str] = None,
         format_: Optional[
             mlrun.common.formatters.ArtifactFormat
         ] = mlrun.common.formatters.ArtifactFormat.full,
@@ -1253,6 +1254,8 @@ class HTTPRunDB(RunDBInterface):
                 "results", tag="*", project="iris", labels=["uploaded", "type=binary"]
             )
 
+        :param parent: The URI or <parent_name>:<parent_tag> string of the parent artifact.
+                Used to filter and return only artifacts that are direct children of the specified parent.
         :param name: Name of artifacts to retrieve. Name with '~' prefix is used as a like query, and is not
             case-sensitive. This means that querying for ``~name`` may return artifacts named
             ``my_Name_1`` or ``surname``.
@@ -1309,6 +1312,7 @@ class HTTPRunDB(RunDBInterface):
             partition_sort_by=partition_sort_by,
             partition_order=partition_order,
             return_all=not limit,
+            parent=parent,
         )
         return artifacts
 
@@ -5080,6 +5084,7 @@ class HTTPRunDB(RunDBInterface):
         category: Union[str, mlrun.common.schemas.ArtifactCategories] = None,
         tree: Optional[str] = None,
         producer_uri: Optional[str] = None,
+        parent: Optional[str] = None,
         format_: Optional[
             mlrun.common.formatters.ArtifactFormat
         ] = mlrun.common.formatters.ArtifactFormat.full,
@@ -5128,6 +5133,7 @@ class HTTPRunDB(RunDBInterface):
             "page": page,
             "page-size": page_size,
             "page-token": page_token,
+            "parent": parent,
         }
 
         if partition_by:

mlrun/db/nopdb.py

@@ -208,6 +208,7 @@ class NopDB(RunDBInterface):
         kind: Optional[str] = None,
         category: Union[str, mlrun.common.schemas.ArtifactCategories] = None,
         tree: Optional[str] = None,
+        parent: Optional[str] = None,
         format_: mlrun.common.formatters.ArtifactFormat = mlrun.common.formatters.ArtifactFormat.full,
         limit: Optional[int] = None,
         partition_by: Optional[

mlrun/execution.py

@@ -31,6 +31,7 @@ from mlrun.artifacts import (
     DatasetArtifact,
     DocumentArtifact,
     DocumentLoaderSpec,
+    LLMPromptArtifact,
     ModelArtifact,
 )
 from mlrun.datastore.store_resources import get_store_resource
@@ -808,6 +809,8 @@ class MLClientCtx:
         label_column: Optional[Union[str, list]] = None,
         extra_data=None,
         db_key=None,
+        model_url: Optional[str] = None,
+        default_config=None,
         **kwargs,
     ) -> ModelArtifact:
         """Log a model artifact and optionally upload it to datastore
@@ -850,6 +853,9 @@ class MLClientCtx:
                                 value can be absolute path | relative path (to model dir) | bytes | artifact object
         :param db_key:          The key to use in the artifact DB table, by default its run name + '_' + key
                                 db_key=False will not register it in the artifacts table
+        :param model_url:       Remote model url.
+        :param default_config:  Default configuration for client building
+                                Saved as a sub-dictionary under the parameter.
 
         :returns: Model artifact object
         """
@@ -858,7 +864,6 @@ class MLClientCtx:
             raise MLRunInvalidArgumentError(
                 "Cannot specify inputs and training set together"
             )
-
         model = ModelArtifact(
             key,
             body,
@@ -873,6 +878,8 @@ class MLClientCtx:
             feature_vector=feature_vector,
             feature_weights=feature_weights,
             extra_data=extra_data,
+            model_url=model_url,
+            default_config=default_config,
             **kwargs,
         )
         if training_set is not None:
@@ -893,6 +900,85 @@ class MLClientCtx:
         self._update_run()
         return item
 
+    def log_llm_prompt(
+        self,
+        key,
+        prompt_string: Optional[str] = None,
+        prompt_path: Optional[str] = None,
+        prompt_legend: Optional[dict] = None,
+        model_artifact: Union[ModelArtifact, str] = None,
+        model_configuration: Optional[dict] = None,
+        description: Optional[str] = None,
+        target_path: Optional[str] = None,
+        artifact_path: Optional[str] = None,
+        tag: Optional[str] = None,
+        labels: Optional[Union[list[str], str]] = None,
+        upload: Optional[bool] = None,
+        **kwargs,
+    ) -> LLMPromptArtifact:
+        """Log an LLM prompt artifact and optionally upload it to the artifact store.
+
+        This function allows you to log a prompt artifact for large language model (LLM) usage. Prompts can be defined
+        as a string or by referencing a file path. Optionally, you can link the prompt to a parent model artifact and
+        provide metadata like a prompt legend (e.g., input variable mapping) and generation configuration.
+
+        Examples::
+
+            # Log an inline prompt
+            context.log_llm_prompt(
+                key="qa-prompt",
+                prompt_string="Q: {question}",
+                model_artifact=model,
+                prompt_legend={"question": "user_input"},
+                model_configuration={"temperature": 0.7, "max_tokens": 128},
+                tag="latest",
+            )
+
+        :param key: Unique name of the artifact.
+        :param prompt_string: Raw prompt text as a string. Cannot be used with `prompt_path`.
+        :param prompt_path: Path to a file containing the prompt content. Cannot be used with `prompt_string`.
+        :param prompt_legend: A dictionary where each key is a placeholder in the prompt (e.g., ``{user_name}``)
+               and the value is a description or explanation of what that placeholder represents.
+               Useful for documenting and clarifying dynamic parts of the prompt.
+        :param model_artifact: Reference to the parent model (either `ModelArtifact` or model URI string).
+        :param model_configuration: Dictionary of generation parameters (e.g., temperature, max_tokens).
+        :param description: Optional description of the prompt.
+        :param target_path: Path to write the artifact locally.
+        :param artifact_path: Path in the artifact store (defaults to project artifact path).
+        :param tag: Tag/version to assign to the prompt artifact.
+        :param labels: Labels to tag the artifact (e.g., list or dict of key-value pairs).
+        :param upload: Whether to upload the artifact to the store (defaults to True).
+        :param kwargs: Additional fields to pass to the `LLMPromptArtifact` constructor.
+
+        :returns: The logged `LLMPromptArtifact` object.
+        """
+
+        llm_prompt = LLMPromptArtifact(
+            key=key,
+            project=self.project or "",
+            prompt_string=prompt_string,
+            prompt_path=prompt_path,
+            prompt_legend=prompt_legend,
+            model_artifact=model_artifact,
+            model_configuration=model_configuration,
+            target_path=target_path,
+            description=description,
+            **kwargs,
+        )
+
+        item = cast(
+            LLMPromptArtifact,
+            self.log_artifact(
+                llm_prompt,
+                artifact_path=artifact_path,
+                tag=tag,
+                upload=upload,
+                labels=labels,
+            ),
+        )
+        self._update_run()
+        return item
+
     def log_document(
         self,
         key: str = "",

mlrun/model.py

@@ -2155,7 +2155,6 @@ class DataSource(ModelObj):
         "max_age",
         "start_time",
         "end_time",
-        "credentials_prefix",
     ]
     kind = None
 
@@ -2218,7 +2217,6 @@ class DataTargetBase(ModelObj):
         "storage_options",
         "run_id",
         "schema",
-        "credentials_prefix",
     ]
 
     @classmethod
@@ -2253,7 +2251,6 @@ class DataTargetBase(ModelObj):
         flush_after_seconds: Optional[int] = None,
         storage_options: Optional[dict[str, str]] = None,
         schema: Optional[dict[str, Any]] = None,
-        credentials_prefix=None,
     ):
         self.name = name
         self.kind: str = kind
@@ -2270,7 +2267,6 @@ class DataTargetBase(ModelObj):
         self.storage_options = storage_options
         self.run_id = None
         self.schema = schema
-        self.credentials_prefix = credentials_prefix
 
 
 class FeatureSetProducer(ModelObj):
@@ -2303,7 +2299,6 @@ class DataTarget(DataTargetBase):
         "key_bucketing_number",
         "partition_cols",
         "time_partitioning_granularity",
-        "credentials_prefix",
     ]
 
     def __init__(

mlrun/model_monitoring/applications/base.py

@@ -409,8 +409,8 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
         tag: Optional[str] = None,
         run_local: bool = True,
         auto_build: bool = True,
-        sample_data: Optional[pd.DataFrame] = None,
-        reference_data: Optional[pd.DataFrame] = None,
+        sample_data: Optional[Union[pd.DataFrame, str]] = None,
+        reference_data: Optional[Union[pd.DataFrame, str]] = None,
         image: Optional[str] = None,
         with_repo: Optional[bool] = False,
         class_handler: Optional[str] = None,
@@ -434,9 +434,11 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
         :param tag:               Tag for the function.
         :param run_local:         Whether to run the function locally or remotely.
         :param auto_build:        Whether to auto build the function.
-        :param sample_data:       Pandas data-frame as the current dataset.
+        :param sample_data:       Pandas data-frame or :py:class:`~mlrun.artifacts.dataset.DatasetArtifact` URI as
+                                  the current dataset.
                                   When set, it replaces the data read from the model endpoint's offline source.
-        :param reference_data:    Pandas data-frame of the reference dataset.
+        :param reference_data:    Pandas data-frame or :py:class:`~mlrun.artifacts.dataset.DatasetArtifact` URI as
+                                  the reference dataset.
                                   When set, its statistics override the model endpoint's feature statistics.
         :param image:             Docker image to run the job on (when running remotely).
         :param with_repo:         Whether to clone the current repo to the build source.
@@ -515,7 +517,9 @@ class ModelMonitoringApplicationBase(MonitoringApplicationToDict, ABC):
             (sample_data, "sample_data"),
             (reference_data, "reference_data"),
         ]:
-            if data is not None:
+            if isinstance(data, str):
+                inputs[identifier] = data
+            elif data is not None:
                 key = f"{job.metadata.name}_{identifier}"
                 inputs[identifier] = project.log_dataset(
                     key,