mlrun 1.7.2rc4__py3-none-any.whl → 1.8.0__py3-none-any.whl

mlrun/render.py

@@ -361,9 +361,6 @@ def get_tblframe(df, display, classes=None):
     return ipython_display(html, display)
 
 
-uid_template = '<div title="{}"><a href="{}/{}/{}/jobs/monitor/{}/overview" target="_blank" >...{}</a></div>'
-
-
 def runs_to_html(
     df: pd.DataFrame,
     display: bool = True,
@@ -379,15 +376,14 @@ def runs_to_html(
     df["results"] = df["results"].apply(dict_html)
     df["start"] = df["start"].apply(time_str)
     df["parameters"] = df["parameters"].apply(dict_html)
+    uid_template = '<div title="{}"><a href="{}" target="_blank" >...{}</a></div>'
+
     if config.resolve_ui_url():
         df["uid"] = df.apply(
             lambda x: uid_template.format(
-                x.uid,
-                config.resolve_ui_url(),
-                config.ui.projects_prefix,
-                x.project,
-                x.uid,
-                x.uid[-8:],
+                x["uid"],
+                mlrun.utils.get_run_url(x["project"], x["uid"], x["name"]),
+                x["uid"][-8:],
             ),
             axis=1,
         )

mlrun/run.py

@@ -30,15 +30,15 @@ from typing import Optional, Union
 
 import nuclio
 import yaml
-from mlrun_pipelines.common.models import RunStatuses
-from mlrun_pipelines.common.ops import format_summary_from_kfp_run, show_kfp_run
-from mlrun_pipelines.utils import get_client
 
 import mlrun.common.constants as mlrun_constants
 import mlrun.common.formatters
 import mlrun.common.schemas
 import mlrun.errors
 import mlrun.utils.helpers
+import mlrun_pipelines.utils
+from mlrun_pipelines.common.models import RunStatuses
+from mlrun_pipelines.common.ops import format_summary_from_kfp_run, show_kfp_run
 
 from .common.helpers import parse_versioned_object_uri
 from .config import config as mlconf
@@ -73,6 +73,9 @@ from .utils import (
     update_in,
 )
 
+if typing.TYPE_CHECKING:
+    from mlrun.datastore import DataItem
+
 
 def function_to_module(code="", workdir=None, secrets=None, silent=False):
     """Load code, notebook or mlrun function as .py module
@@ -200,7 +203,7 @@ def get_or_create_ctx(
     event=None,
     spec: Optional[dict] = None,
     with_env: bool = True,
-    rundb: str = "",
+    rundb: Union[str, "mlrun.db.RunDBInterface"] = "",
     project: str = "",
     upload_artifacts: bool = False,
     labels: Optional[dict] = None,
@@ -258,7 +261,7 @@ def get_or_create_ctx(
     """
     if labels:
         warnings.warn(
-            "The `labels` argument is deprecated and will be removed in 1.9.0. "
+            "The `labels` argument is deprecated in 1.7.0 and will be removed in 1.10.0. "
             "Please use `spec` instead, e.g.:\n"
             "spec={'metadata': {'labels': {'key': 'value'}}}",
             FutureWarning,
@@ -306,7 +309,7 @@ def get_or_create_ctx(
     out = rundb or mlconf.dbpath or environ.get("MLRUN_DBPATH")
     if out:
         autocommit = True
-        logger.info(f"logging run results to: {out}")
+        logger.info(f"Logging run results to: {out}")
 
     newspec["metadata"]["project"] = (
         newspec["metadata"].get("project") or project or mlconf.default_project
@@ -434,7 +437,7 @@ def new_function(
     mode: Optional[str] = None,
     handler: Optional[str] = None,
     source: Optional[str] = None,
-    requirements: Union[str, list[str]] = None,
+    requirements: Optional[list[str]] = None,
     kfp: Optional[bool] = None,
     requirements_file: str = "",
 ):
@@ -634,7 +637,7 @@ def code_to_function(
     - databricks: run code on Databricks cluster (python scripts, Spark etc.)
     - application: run a long living application (e.g. a web server, UI, etc.)
 
-    Learn more about [Kinds of function (runtimes)](../concepts/functions-overview.html).
+    Learn more about :doc:`../../concepts/functions-overview`
 
     :param name:         function name, typically best to use hyphen-case
     :param project:      project used to namespace the function, defaults to 'default'
@@ -906,13 +909,53 @@ def _run_pipeline(
     return pipeline_run_id
 
 
+def retry_pipeline(
+    run_id: str,
+    project: str,
+    namespace: Optional[str] = None,
+) -> str:
+    """Retry a pipeline run.
+
+    This function retries a previously executed pipeline run using the specified run ID. If the run is not in a
+    retryable state, a new run is created as a clone of the original run.
+
+    :param run_id: ID of the pipeline run to retry.
+    :param project: name of the project associated with the pipeline run.
+    :param namespace: Optional; Kubernetes namespace to use if not the default.
+
+    :returns: ID of the retried pipeline run or the ID of a cloned run if the original run is not retryable.
+    :raises ValueError: If access to the remote API service is not available.
+    """
+    mldb = mlrun.db.get_run_db()
+    if mldb.kind != "http":
+        raise ValueError(
+            "Retrying a pipeline requires access to remote API service. "
+            "Please set the dbpath URL."
+        )
+
+    pipeline_run_id = mldb.retry_pipeline(
+        run_id=run_id,
+        project=project,
+        namespace=namespace,
+    )
+    if pipeline_run_id == run_id:
+        logger.info(
+            f"Retried pipeline run ID={pipeline_run_id}, check UI for progress."
+        )
+    else:
+        logger.info(
+            f"Copy of pipeline {run_id} was retried as run ID={pipeline_run_id}, check UI for progress."
+        )
+    return pipeline_run_id
+
+
 def wait_for_pipeline_completion(
     run_id,
     timeout=60 * 60,
-    expected_statuses: list[str] = None,
+    expected_statuses: Optional[list[str]] = None,
     namespace=None,
     remote=True,
-    project: str = None,
+    project: Optional[str] = None,
 ):
     """Wait for Pipeline status, timeout in sec
 
@@ -972,7 +1015,7 @@ def wait_for_pipeline_completion(
             _wait_for_pipeline_completion,
         )
     else:
-        client = get_client(namespace=namespace)
+        client = mlrun_pipelines.utils.get_client(namespace=namespace)
         resp = client.wait_for_run_completion(run_id, timeout)
         if resp:
             resp = resp.to_dict()
@@ -1004,7 +1047,7 @@ def get_pipeline(
     format_: Union[
         str, mlrun.common.formatters.PipelineFormat
     ] = mlrun.common.formatters.PipelineFormat.summary,
-    project: str = None,
+    project: Optional[str] = None,
     remote: bool = True,
 ):
     """Get Pipeline status
@@ -1033,7 +1076,7 @@ def get_pipeline(
         )
 
     else:
-        client = get_client(namespace=namespace)
+        client = mlrun_pipelines.utils.get_client(namespace=namespace)
         resp = client.get_run(run_id)
         if resp:
             resp = resp.to_dict()
@@ -1089,7 +1132,7 @@ def get_object(url, secrets=None, size=None, offset=0, db=None):
     return stores.object(url=url).get(size, offset)
 
 
-def get_dataitem(url, secrets=None, db=None) -> mlrun.datastore.DataItem:
+def get_dataitem(url, secrets=None, db=None) -> "DataItem":
     """get mlrun dataitem object (from path/url)"""
     stores = store_manager.set(secrets, db=db)
     return stores.object(url=url)

mlrun/runtimes/__init__.py

@@ -12,8 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# flake8: noqa  - this is until we take care of the F401 violations with respect to __all__ & sphinx
-
 __all__ = [
     "BaseRuntime",
     "KubejobRuntime",
@@ -60,7 +58,7 @@ from ..serving import MLModelServer, new_v1_model_server  # noqa isort: skip
 def new_model_server(
     name,
     model_class: str,
-    models: dict = None,
+    models: typing.Optional[dict] = None,
     filename="",
     protocol="",
     image="",

mlrun/runtimes/base.py

@@ -16,31 +16,38 @@ import http
 import re
 import typing
 import warnings
-from base64 import b64encode
 from os import environ
 from typing import Callable, Optional, Union
 
 import requests.exceptions
-from mlrun_pipelines.common.ops import mlrun_op
 from nuclio.build import mlrun_footer
 
 import mlrun.common.constants
 import mlrun.common.constants as mlrun_constants
+import mlrun.common.formatters
+import mlrun.common.runtimes
 import mlrun.common.schemas
 import mlrun.common.schemas.model_monitoring.constants as mm_constants
-import mlrun.db
 import mlrun.errors
 import mlrun.launcher.factory
 import mlrun.utils.helpers
 import mlrun.utils.notifications
 import mlrun.utils.regex
+from mlrun.model import (
+    BaseMetadata,
+    HyperParamOptions,
+    ImageBuilder,
+    ModelObj,
+    RunObject,
+    RunTemplate,
+)
 from mlrun.utils.helpers import generate_object_uri, verify_field_regex
+from mlrun_pipelines.common.ops import mlrun_op
 
 from ..config import config
 from ..datastore import store_manager
 from ..errors import err_to_str
 from ..lists import RunList
-from ..model import BaseMetadata, HyperParamOptions, ImageBuilder, ModelObj, RunObject
 from ..utils import (
     dict_to_json,
     dict_to_yaml,
@@ -141,10 +148,10 @@ class FunctionSpec(ModelObj):
 
     @property
     def clone_target_dir(self):
-        # TODO: remove this property in 1.9.0
+        # TODO: remove this property in 1.10.0
         if self.build.source_code_target_dir:
             warnings.warn(
-                "The clone_target_dir attribute is deprecated in 1.6.2 and will be removed in 1.9.0. "
+                "The clone_target_dir attribute is deprecated in 1.6.2 and will be removed in 1.10.0. "
                 "Use spec.build.source_code_target_dir instead.",
                 FutureWarning,
             )
@@ -152,10 +159,10 @@ class FunctionSpec(ModelObj):
 
     @clone_target_dir.setter
     def clone_target_dir(self, clone_target_dir):
-        # TODO: remove this property in 1.9.0
+        # TODO: remove this property in 1.10.0
         if clone_target_dir:
             warnings.warn(
-                "The clone_target_dir attribute is deprecated in 1.6.2 and will be removed in 1.9.0. "
+                "The clone_target_dir attribute is deprecated in 1.6.2 and will be removed in 1.10.0. "
                 "Use spec.build.source_code_target_dir instead.",
                 FutureWarning,
             )
@@ -517,7 +524,7 @@ class BaseRuntime(ModelObj):
                 run_results = resp["status"].get("results", {})
                 if generator.eval_stop_condition(run_results):
                     logger.info(
-                        f"reached early stop condition ({generator.options.stop_condition}), stopping iterations!"
+                        f"Reached early stop condition ({generator.options.stop_condition}), stopping iterations!"
                     )
                     results.append(resp)
                     break
@@ -529,7 +536,7 @@ class BaseRuntime(ModelObj):
                 resp = self._update_run_state(task=task, err=error_string)
                 num_errors += 1
                 if num_errors > generator.max_errors:
-                    logger.error("too many errors, stopping iterations!")
+                    logger.error("Too many errors, stopping iterations!")
                     results.append(resp)
                     break
 
@@ -553,9 +560,9 @@ class BaseRuntime(ModelObj):
 
     def _update_run_state(
         self,
-        resp: dict = None,
+        resp: Optional[dict] = None,
         task: RunObject = None,
-        err: Union[Exception, str] = None,
+        err: Optional[Union[Exception, str]] = None,
         run_format: mlrun.common.formatters.RunFormat = mlrun.common.formatters.RunFormat.full,
     ) -> typing.Optional[dict]:
         """update the task state in the DB"""
@@ -644,7 +651,10 @@ class BaseRuntime(ModelObj):
         return updates
 
     def full_image_path(
-        self, image=None, client_version: str = None, client_python_version: str = None
+        self,
+        image=None,
+        client_version: Optional[str] = None,
+        client_python_version: Optional[str] = None,
     ):
         image = image or self.spec.image or ""
 
@@ -665,20 +675,20 @@ class BaseRuntime(ModelObj):
 
     def as_step(
         self,
-        runspec: RunObject = None,
+        runspec: Union[RunObject, RunTemplate] = None,
         handler=None,
         name: str = "",
         project: str = "",
-        params: dict = None,
+        params: Optional[dict] = None,
         hyperparams=None,
         selector="",
         hyper_param_options: HyperParamOptions = None,
-        inputs: dict = None,
-        outputs: list = None,
+        inputs: Optional[dict] = None,
+        outputs: Optional[list] = None,
         workdir: str = "",
         artifact_path: str = "",
         image: str = "",
-        labels: dict = None,
+        labels: Optional[dict] = None,
         use_db=True,
         verbose=None,
         scrape_metrics=False,
@@ -791,9 +801,7 @@ class BaseRuntime(ModelObj):
                     mlrun.runtimes.nuclio.serving.serving_subkind
                 )
 
-        self.spec.build.functionSourceCode = b64encode(body.encode("utf-8")).decode(
-            "utf-8"
-        )
+        self.spec.build.functionSourceCode = mlrun.utils.helpers.encode_user_code(body)
         if with_doc:
             update_function_entry_points(self, body)
         return self
@@ -851,7 +859,7 @@ class BaseRuntime(ModelObj):
 
     def requires_build(self) -> bool:
         build = self.spec.build
-        return (
+        return bool(
             build.commands
             or build.requirements
             or (build.source and not build.load_source_on_run)

mlrun/runtimes/daskjob.py

@@ -363,7 +363,7 @@ class DaskCluster(KubejobRuntime):
         skip_deployed=False,
         is_kfp=False,
         mlrun_version_specifier=None,
-        builder_env: dict = None,
+        builder_env: Optional[dict] = None,
         show_on_failure: bool = False,
         force_build: bool = False,
     ):
@@ -406,9 +406,9 @@ class DaskCluster(KubejobRuntime):
 
     def with_scheduler_limits(
         self,
-        mem: str = None,
-        cpu: str = None,
-        gpus: int = None,
+        mem: Optional[str] = None,
+        cpu: Optional[str] = None,
+        gpus: Optional[int] = None,
         gpu_type: str = "nvidia.com/gpu",
         patch: bool = False,
     ):
@@ -422,9 +422,9 @@ class DaskCluster(KubejobRuntime):
 
     def with_worker_limits(
         self,
-        mem: str = None,
-        cpu: str = None,
-        gpus: int = None,
+        mem: Optional[str] = None,
+        cpu: Optional[str] = None,
+        gpus: Optional[int] = None,
         gpu_type: str = "nvidia.com/gpu",
         patch: bool = False,
     ):
@@ -442,7 +442,7 @@ class DaskCluster(KubejobRuntime):
         )
 
     def with_scheduler_requests(
-        self, mem: str = None, cpu: str = None, patch: bool = False
+        self, mem: Optional[str] = None, cpu: Optional[str] = None, patch: bool = False
     ):
         """
         set scheduler pod resources requests
@@ -451,7 +451,7 @@ class DaskCluster(KubejobRuntime):
         self.spec._verify_and_set_requests("scheduler_resources", mem, cpu, patch=patch)
 
     def with_worker_requests(
-        self, mem: str = None, cpu: str = None, patch: bool = False
+        self, mem: Optional[str] = None, cpu: Optional[str] = None, patch: bool = False
     ):
         """
         set worker pod resources requests

mlrun/runtimes/databricks_job/databricks_runtime.py

@@ -11,14 +11,15 @@
 # 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 typing
 from ast import FunctionDef, parse, unparse
-from base64 import b64decode, b64encode
+from base64 import b64decode
 from typing import Callable, Optional, Union
 
 import mlrun
 import mlrun.runtimes.kubejob as kubejob
 import mlrun.runtimes.pod as pod
+import mlrun.utils.helpers
 from mlrun.errors import MLRunInvalidArgumentError
 from mlrun.model import HyperParamOptions, RunObject
 
@@ -138,7 +139,7 @@ class DatabricksRuntime(kubejob.KubejobRuntime):
             )
 
     def _get_modified_user_code(self, original_handler: str, log_artifacts_code: str):
-        encoded_code = (
+        encoded_code: typing.Optional[str] = (
             self.spec.build.functionSourceCode if hasattr(self.spec, "build") else None
         )
         if not encoded_code:
@@ -162,7 +163,7 @@ class DatabricksRuntime(kubejob.KubejobRuntime):
         if original_handler:
             decoded_code += f"\nresult = {original_handler}(**handler_arguments)\n"
             decoded_code += _return_artifacts_code
-        return b64encode(decoded_code.encode("utf-8")).decode("utf-8")
+        return mlrun.utils.helpers.encode_user_code(decoded_code)
 
     def get_internal_parameters(self, runobj: RunObject):
         """
@@ -202,7 +203,7 @@ from mlrun.runtimes.databricks_job import databricks_wrapper
 def run_mlrun_databricks_job(context,task_parameters: dict, **kwargs):
         databricks_wrapper.run_mlrun_databricks_job(context, task_parameters, **kwargs)
 """
-        wrap_code = b64encode(wrap_code).decode("utf-8")
+        wrap_code = mlrun.utils.helpers.encode_user_code(wrap_code)
         self.spec.build.functionSourceCode = wrap_code
         runspec.spec.handler = "run_mlrun_databricks_job"
 

mlrun/runtimes/function_reference.py

@@ -13,11 +13,11 @@
 # limitations under the License.
 
 import os
-from base64 import b64encode
 
 from nuclio.build import mlrun_footer
 
 import mlrun
+import mlrun.utils.helpers
 
 from ..model import ModelObj
 from ..utils import generate_object_uri
@@ -36,6 +36,7 @@ class FunctionReference(ModelObj):
         spec=None,
         kind=None,
         name=None,
+        track_models=None,
     ):
         self.url = url
         self.kind = kind
@@ -46,6 +47,7 @@ class FunctionReference(ModelObj):
             spec = spec.to_dict()
         self.spec = spec
         self.code = code
+        self.track_models = track_models
 
         self._function = None
         self._address = None
@@ -116,7 +118,7 @@ class FunctionReference(ModelObj):
             func = mlrun.new_function(
                 self.name, kind=kind, image=self.image or default_image
             )
-            data = b64encode(code.encode("utf-8")).decode("utf-8")
+            data = mlrun.utils.helpers.encode_user_code(code)
             func.spec.build.functionSourceCode = data
             if kind not in mlrun.runtimes.RuntimeKinds.nuclio_runtimes():
                 func.spec.default_handler = "handler"
@@ -130,6 +132,7 @@ class FunctionReference(ModelObj):
         if self.requirements:
             func.with_requirements(self.requirements)
         self._function = func
+        func.spec.track_models = self.track_models
         return func
 
     @property

mlrun/runtimes/generators.py

@@ -15,6 +15,7 @@ import json
 import random
 import sys
 from copy import deepcopy
+from typing import Optional
 
 import pandas as pd
 
@@ -26,7 +27,7 @@ default_max_iterations = 10
 default_max_errors = 3
 
 
-def get_generator(spec: RunSpec, execution, param_file_secrets: dict = None):
+def get_generator(spec: RunSpec, execution, param_file_secrets: Optional[dict] = None):
     options = spec.hyper_param_options
     strategy = spec.strategy or options.strategy
     if not spec.is_hyper_job() or strategy == "custom":
@@ -181,7 +182,7 @@ class ListGenerator(TaskGenerator):
             yield newrun
 
 
-def get_run_copy(run):
+def get_run_copy(run: RunObject):
     newrun = deepcopy(run)
     newrun.spec.hyperparams = None
     newrun.spec.param_file = None

mlrun/runtimes/kubejob.py

@@ -14,11 +14,10 @@
 import typing
 import warnings
 
-from mlrun_pipelines.common.ops import build_op
-
 import mlrun.common.schemas
 import mlrun.db
 import mlrun.errors
+from mlrun_pipelines.common.ops import build_op
 
 from ..model import RunObject
 from .pod import KubeResource
@@ -76,7 +75,7 @@ class KubejobRuntime(KubeResource):
         self,
         image="",
         base_image=None,
-        commands: list = None,
+        commands: typing.Optional[list] = None,
         secret=None,
         source=None,
         extra=None,
@@ -115,9 +114,9 @@ class KubejobRuntime(KubeResource):
             e.g. builder_env={"GIT_TOKEN": token}
         """
         if not overwrite:
-            # TODO: change overwrite default to True in 1.8.0
+            # TODO: change overwrite default to True in 1.10.0
             warnings.warn(
-                "The `overwrite` parameter default will change from 'False' to 'True' in 1.8.0.",
+                "The `overwrite` parameter default will change from 'False' to 'True' in 1.10.0.",
                 mlrun.utils.OverwriteBuildParamsWarning,
             )
         image = mlrun.utils.helpers.remove_image_protocol_prefix(image)
@@ -148,7 +147,7 @@ class KubejobRuntime(KubeResource):
         skip_deployed: bool = False,
         is_kfp: bool = False,
         mlrun_version_specifier: typing.Optional[bool] = None,
-        builder_env: dict = None,
+        builder_env: typing.Optional[dict] = None,
         show_on_failure: bool = False,
         force_build: bool = False,
     ) -> bool:
@@ -190,7 +189,7 @@ class KubejobRuntime(KubeResource):
         self,
         image=None,
         base_image=None,
-        commands: list = None,
+        commands: typing.Optional[list] = None,
         secret_name="",
         with_mlrun=True,
         skip_deployed=False,