mlrun 1.7.0rc2__py3-none-any.whl → 1.7.0rc4__py3-none-any.whl

mlrun/db/httpdb.py

@@ -152,7 +152,7 @@ class HTTPRunDB(RunDBInterface):
     @staticmethod
     def get_api_path_prefix(version: str = None) -> str:
         """
-        :param version: API version to use, None (the default) will mean to use the default value from mlconf,
+        :param version: API version to use, None (the default) will mean to use the default value from mlrun.config,
          for un-versioned api set an empty string.
         """
         if version is not None:
@@ -250,7 +250,11 @@ class HTTPRunDB(RunDBInterface):
 
         try:
             response = self.session.request(
-                method, url, timeout=timeout, verify=False, **kw
+                method,
+                url,
+                timeout=timeout,
+                verify=config.httpdb.http.verify,
+                **kw,
             )
         except requests.RequestException as exc:
             error = f"{err_to_str(exc)}: {error}" if error else err_to_str(exc)
@@ -302,11 +306,11 @@ class HTTPRunDB(RunDBInterface):
 
     def connect(self, secrets=None):
         """Connect to the MLRun API server. Must be called prior to executing any other method.
-        The code utilizes the URL for the API server from the configuration - ``mlconf.dbpath``.
+        The code utilizes the URL for the API server from the configuration - ``config.dbpath``.
 
         For example::
 
-            mlconf.dbpath = mlconf.dbpath or 'http://mlrun-api:8080'
+            config.dbpath = config.dbpath or 'http://mlrun-api:8080'
             db = get_run_db().connect()
         """
         # hack to allow unit tests to instantiate HTTPRunDB without a real server behind
@@ -454,6 +458,10 @@ class HTTPRunDB(RunDBInterface):
                     setattr(
                         config.feature_store.data_prefixes, prefix, server_prefix_value
                     )
+            config.feature_store.default_targets = (
+                server_cfg.get("feature_store_default_targets")
+                or config.feature_store.default_targets
+            )
 
         except Exception as exc:
             logger.warning(
@@ -500,7 +508,7 @@ class HTTPRunDB(RunDBInterface):
         if offset < 0:
             raise MLRunInvalidArgumentError("Offset cannot be negative")
         if size is None:
-            size = int(mlrun.mlconf.httpdb.logs.pull_logs_default_size_limit)
+            size = int(config.httpdb.logs.pull_logs_default_size_limit)
         elif size == -1:
             logger.warning(
                 "Retrieving all logs. This may be inefficient and can result in a large log."
@@ -546,25 +554,23 @@ class HTTPRunDB(RunDBInterface):
 
         state, text = self.get_log(uid, project, offset=offset)
         if text:
-            print(text.decode(errors=mlrun.mlconf.httpdb.logs.decode.errors))
+            print(text.decode(errors=config.httpdb.logs.decode.errors))
         nil_resp = 0
         while True:
             offset += len(text)
             # if we get 3 nil responses in a row, increase the sleep time to 10 seconds
             # TODO: refactor this to use a conditional backoff mechanism
             if nil_resp < 3:
-                time.sleep(int(mlrun.mlconf.httpdb.logs.pull_logs_default_interval))
+                time.sleep(int(config.httpdb.logs.pull_logs_default_interval))
             else:
                 time.sleep(
-                    int(
-                        mlrun.mlconf.httpdb.logs.pull_logs_backoff_no_logs_default_interval
-                    )
+                    int(config.httpdb.logs.pull_logs_backoff_no_logs_default_interval)
                 )
             state, text = self.get_log(uid, project, offset=offset)
             if text:
                 nil_resp = 0
                 print(
-                    text.decode(errors=mlrun.mlconf.httpdb.logs.decode.errors),
+                    text.decode(errors=config.httpdb.logs.decode.errors),
                     end="",
                 )
             else:
@@ -1135,17 +1141,17 @@ class HTTPRunDB(RunDBInterface):
             structured_dict = {}
             for project, job_runtime_resources_map in response.json().items():
                 for job_id, runtime_resources in job_runtime_resources_map.items():
-                    structured_dict.setdefault(project, {})[
-                        job_id
-                    ] = mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                    structured_dict.setdefault(project, {})[job_id] = (
+                        mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                    )
             return structured_dict
         elif group_by == mlrun.common.schemas.ListRuntimeResourcesGroupByField.project:
             structured_dict = {}
             for project, kind_runtime_resources_map in response.json().items():
                 for kind, runtime_resources in kind_runtime_resources_map.items():
-                    structured_dict.setdefault(project, {})[
-                        kind
-                    ] = mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                    structured_dict.setdefault(project, {})[kind] = (
+                        mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                    )
             return structured_dict
         else:
             raise NotImplementedError(
@@ -1173,7 +1179,8 @@ class HTTPRunDB(RunDBInterface):
         :param force: Force deletion - delete the runtime resource even if it's not in terminal state or if the grace
             period didn't pass.
         :param grace_period: Grace period given to the runtime resource before they are actually removed, counted from
-            the moment they moved to terminal state (defaults to mlrun.mlconf.runtime_resources_deletion_grace_period).
+            the moment they moved to terminal state
+            (defaults to mlrun.config.config.runtime_resources_deletion_grace_period).
 
         :returns: :py:class:`~mlrun.common.schemas.GroupedByProjectRuntimeResourcesOutput` listing the runtime resources
             that were removed.
@@ -1203,9 +1210,9 @@ class HTTPRunDB(RunDBInterface):
         structured_dict = {}
         for project, kind_runtime_resources_map in response.json().items():
             for kind, runtime_resources in kind_runtime_resources_map.items():
-                structured_dict.setdefault(project, {})[
-                    kind
-                ] = mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                structured_dict.setdefault(project, {})[kind] = (
+                    mlrun.common.schemas.RuntimeResources(**runtime_resources)
+                )
         return structured_dict
 
     def create_schedule(
@@ -1340,7 +1347,7 @@ class HTTPRunDB(RunDBInterface):
             logger.warning(
                 "Building a function image to ECR and loading an S3 source to the image may require conflicting access "
                 "keys. Only the permissions granted to the platform's configured secret will take affect "
-                "(see mlrun.mlconf.httpdb.builder.docker_registry_secret). "
+                "(see mlrun.config.config.httpdb.builder.docker_registry_secret). "
                 "In case the permissions are limited to ECR scope, you may use pull_at_runtime=True instead",
                 source=func.spec.build.source,
                 load_source_on_run=func.spec.build.load_source_on_run,
@@ -1495,7 +1502,7 @@ class HTTPRunDB(RunDBInterface):
         Retrieve updated information on project background tasks being executed.
         If no filter is provided, will return background tasks from the last week.
 
-        :param project: Project name (defaults to mlrun.mlconf.default_project).
+        :param project: Project name (defaults to mlrun.config.config.default_project).
         :param state:   List only background tasks whose state is specified.
         :param created_from: Filter by background task created time in ``[created_from, created_to]``.
         :param created_to:  Filter by background task created time in ``[created_from, created_to]``.
@@ -3071,36 +3078,58 @@ class HTTPRunDB(RunDBInterface):
         resp = self.api_call(method="POST", path=path, params=params)
         return resp.json()["func"]
 
-    def create_model_monitoring_controller(
+    def update_model_monitoring_controller(
         self,
-        project: str = "",
-        default_controller_image: str = "mlrun/mlrun",
+        project: str,
         base_period: int = 10,
+        image: str = "mlrun/mlrun",
     ):
         """
-        Submit model monitoring application controller job along with deploying the model monitoring writer function.
-        While the main goal of the controller job is to handle the monitoring processing and triggering applications,
-        the goal of the model monitoring writer function is to write all the monitoring application results to the
-        databases. Note that the default scheduling policy of the controller job is to run every 10 min.
+        Redeploy model monitoring application controller function.
 
         :param project:                  Project name.
-        :param default_controller_image: The default image of the model monitoring controller job. Note that the writer
-                                         function, which is a real time nuclio functino, will be deployed with the same
-                                         image. By default, the image is mlrun/mlrun.
-        :param base_period:              Minutes to determine the frequency in which the model monitoring controller job
-                                         is running. By default, the base period is 5 minutes.
-        :returns: model monitoring controller job as a dictionary. You can easily convert the returned function into a
-                  runtime object by calling ~mlrun.new_function.
+        :param base_period:              The time period in minutes in which the model monitoring controller function
+                                         triggers. By default, the base period is 10 minutes.
+        :param image: The image of the model monitoring controller function.
+                                         By default, the image is mlrun/mlrun.
         """
 
         params = {
-            "default_controller_image": default_controller_image,
+            "image": image,
             "base_period": base_period,
         }
-        path = f"projects/{project}/jobs/model-monitoring-controller"
+        path = f"projects/{project}/model-monitoring/model-monitoring-controller"
+        self.api_call(method="POST", path=path, params=params)
 
-        resp = self.api_call(method="POST", path=path, params=params)
-        return resp.json()["func"]
+    def enable_model_monitoring(
+        self,
+        project: str,
+        base_period: int = 10,
+        image: str = "mlrun/mlrun",
+    ):
+        """
+        Deploy model monitoring application controller, writer and stream functions.
+        While the main goal of the controller function is to handle the monitoring processing and triggering
+        applications, the goal of the model monitoring writer function is to write all the monitoring
+        application results to the databases.
+        The stream function goal is to monitor the log of the data stream. It is triggered when a new log entry
+        is detected. It processes the new events into statistics that are then written to statistics databases.
+
+
+        :param project:                  Project name.
+        :param base_period:              The time period in minutes in which the model monitoring controller function
+                                         triggers. By default, the base period is 10 minutes.
+        :param image:                    The image of the model monitoring controller, writer & monitoring
+                                         stream functions, which are real time nuclio functions.
+                                         By default, the image is mlrun/mlrun.
+        """
+
+        params = {
+            "base_period": base_period,
+            "image": image,
+        }
+        path = f"projects/{project}/model-monitoring/enable-model-monitoring"
+        self.api_call(method="POST", path=path, params=params)
 
     def create_hub_source(
         self, source: Union[dict, mlrun.common.schemas.IndexedHubSource]
@@ -3451,7 +3480,7 @@ class HTTPRunDB(RunDBInterface):
         run_name: Optional[str] = None,
         namespace: Optional[str] = None,
         notifications: list[mlrun.model.Notification] = None,
-    ):
+    ) -> mlrun.common.schemas.WorkflowResponse:
         """
         Submitting workflow for a remote execution.
 

mlrun/execution.py

@@ -559,9 +559,9 @@ class MLClientCtx:
             for k, v in get_in(task, ["status", "results"], {}).items():
                 self._results[k] = v
             for artifact in get_in(task, ["status", run_keys.artifacts], []):
-                self._artifacts_manager.artifacts[
-                    artifact["metadata"]["key"]
-                ] = artifact
+                self._artifacts_manager.artifacts[artifact["metadata"]["key"]] = (
+                    artifact
+                )
                 self._artifacts_manager.link_artifact(
                     self.project,
                     self.name,

mlrun/feature_store/feature_vector.py

@@ -490,10 +490,10 @@ class FeatureVector(ModelObj):
             vector = fstore.FeatureVector("my-vec", features)
 
             # get the vector as a dataframe
-            df = fstore.get_offline_features(vector).to_dataframe()
+            df = vector.get_offline_features().to_dataframe()
 
             # return an online/real-time feature service
-            svc = fstore.get_online_feature_service(vector, impute_policy={"*": "$mean"})
+            svc = vector.get_online_feature_service(impute_policy={"*": "$mean"})
             resp = svc.get([{"stock": "GOOG"}])
 
         :param name:           List of names of targets to delete (default: delete all ingested targets)

mlrun/frameworks/tf_keras/callbacks/logging_callback.py

@@ -389,9 +389,9 @@ class LoggingCallback(Callback):
         ):
             try:
                 self._get_hyperparameter(key_chain=learning_rate_key_chain)
-                self._dynamic_hyperparameters_keys[
-                    learning_rate_key
-                ] = learning_rate_key_chain
+                self._dynamic_hyperparameters_keys[learning_rate_key] = (
+                    learning_rate_key_chain
+                )
             except (KeyError, IndexError, ValueError):
                 pass
 

mlrun/frameworks/tf_keras/model_handler.py

@@ -263,13 +263,13 @@ class TFKerasModelHandler(DLModelHandler):
         # Update the paths and log artifacts if context is available:
         if self._weights_file is not None:
             if self._context is not None:
-                artifacts[
-                    self._get_weights_file_artifact_name()
-                ] = self._context.log_artifact(
-                    self._weights_file,
-                    local_path=self._weights_file,
-                    artifact_path=output_path,
-                    db_key=False,
+                artifacts[self._get_weights_file_artifact_name()] = (
+                    self._context.log_artifact(
+                        self._weights_file,
+                        local_path=self._weights_file,
+                        artifact_path=output_path,
+                        db_key=False,
+                    )
                 )
 
         return artifacts if self._context is not None else None

mlrun/k8s_utils.py

@@ -133,13 +133,13 @@ def sanitize_label_value(value: str) -> str:
     return re.sub(r"([^a-zA-Z0-9_.-]|^[^a-zA-Z0-9]|[^a-zA-Z0-9]$)", "-", value[:63])
 
 
-def verify_label_key(key):
+def verify_label_key(key: str):
+    """
+    Verify that the label key is valid for Kubernetes.
+    Refer to https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/#syntax-and-character-set
+    """
     if not key:
         raise mlrun.errors.MLRunInvalidArgumentError("label key cannot be empty")
-    if key.startswith("k8s.io") or key.startswith("kubernetes.io"):
-        raise mlrun.errors.MLRunInvalidArgumentError(
-            "Labels cannot start with 'k8s.io' or 'kubernetes.io'"
-        )
 
     mlrun.utils.helpers.verify_field_regex(
         f"project.metadata.labels.'{key}'",
@@ -147,6 +147,11 @@ def verify_label_key(key):
         mlrun.utils.regex.k8s_character_limit,
     )
 
+    if key.startswith("k8s.io/") or key.startswith("kubernetes.io/"):
+        raise mlrun.errors.MLRunInvalidArgumentError(
+            "Labels cannot start with 'k8s.io/' or 'kubernetes.io/'"
+        )
+
     parts = key.split("/")
     if len(parts) == 1:
         name = parts[0]

mlrun/kfpops.py

@@ -41,8 +41,8 @@ from .utils import (
 
 # default KFP artifacts and output (ui metadata, metrics etc.)
 # directories to /tmp to allow running with security context
-KFPMETA_DIR = os.environ.get("KFPMETA_OUT_DIR", "/tmp")
-KFP_ARTIFACTS_DIR = os.environ.get("KFP_ARTIFACTS_DIR", "/tmp")
+KFPMETA_DIR = "/tmp"
+KFP_ARTIFACTS_DIR = "/tmp"
 
 project_annotation = "mlrun/project"
 run_annotation = "mlrun/pipeline-step-type"
@@ -71,7 +71,7 @@ def write_kfpmeta(struct):
             {"name": k, "numberValue": v} for k, v in results.items() if is_num(v)
         ],
     }
-    with open(KFPMETA_DIR + "/mlpipeline-metrics.json", "w") as f:
+    with open(os.path.join(KFPMETA_DIR, "mlpipeline-metrics.json"), "w") as f:
         json.dump(metrics, f)
 
     struct = deepcopy(struct)
@@ -91,7 +91,14 @@ def write_kfpmeta(struct):
         elif key in results:
             val = results[key]
         try:
-            path = "/".join([KFP_ARTIFACTS_DIR, key])
+            # NOTE: if key has "../x", it would fail on path traversal
+            path = os.path.join(KFP_ARTIFACTS_DIR, key)
+            if not mlrun.utils.helpers.is_safe_path(KFP_ARTIFACTS_DIR, path):
+                logger.warning(
+                    "Path traversal is not allowed ignoring", path=path, key=key
+                )
+                continue
+            path = os.path.abspath(path)
             logger.info("Writing artifact output", path=path, val=val)
             with open(path, "w") as fp:
                 fp.write(str(val))
@@ -109,7 +116,7 @@ def write_kfpmeta(struct):
         "outputs": output_artifacts
         + [{"type": "markdown", "storage": "inline", "source": text}]
     }
-    with open(KFPMETA_DIR + "/mlpipeline-ui-metadata.json", "w") as f:
+    with open(os.path.join(KFPMETA_DIR, "mlpipeline-ui-metadata.json"), "w") as f:
         json.dump(metadata, f)
 
 
@@ -401,9 +408,9 @@ def mlrun_op(
         cmd += ["--label", f"{label}={val}"]
     for output in outputs:
         cmd += ["-o", str(output)]
-        file_outputs[
-            output.replace(".", "_")
-        ] = f"/tmp/{output}"  # not using path.join to avoid windows "\"
+        file_outputs[output.replace(".", "_")] = (
+            f"/tmp/{output}"  # not using path.join to avoid windows "\"
+        )
     if project:
         cmd += ["--project", project]
     if handler:
@@ -450,8 +457,10 @@ def mlrun_op(
         command=cmd + [command],
         file_outputs=file_outputs,
         output_artifact_paths={
-            "mlpipeline-ui-metadata": KFPMETA_DIR + "/mlpipeline-ui-metadata.json",
-            "mlpipeline-metrics": KFPMETA_DIR + "/mlpipeline-metrics.json",
+            "mlpipeline-ui-metadata": os.path.join(
+                KFPMETA_DIR, "mlpipeline-ui-metadata.json"
+            ),
+            "mlpipeline-metrics": os.path.join(KFPMETA_DIR, "mlpipeline-metrics.json"),
         },
     )
     cop = add_default_function_resources(cop)

mlrun/model.py

@@ -501,6 +501,7 @@ class ImageBuilder(ModelObj):
         requirements: list = None,
         extra_args=None,
         builder_env=None,
+        source_code_target_dir=None,
     ):
         self.functionSourceCode = functionSourceCode  #: functionSourceCode
         self.codeEntryType = ""  #: codeEntryType
@@ -521,6 +522,7 @@ class ImageBuilder(ModelObj):
         self.auto_build = auto_build  #: auto_build
         self.build_pod = None
         self.requirements = requirements or []  #: pip requirements
+        self.source_code_target_dir = source_code_target_dir or None
 
     @property
     def source(self):
@@ -557,6 +559,7 @@ class ImageBuilder(ModelObj):
         overwrite=False,
         builder_env=None,
         extra_args=None,
+        source_code_target_dir=None,
     ):
         if image:
             self.image = image
@@ -582,6 +585,8 @@ class ImageBuilder(ModelObj):
             self.builder_env = builder_env
         if extra_args:
             self.extra_args = extra_args
+        if source_code_target_dir:
+            self.source_code_target_dir = source_code_target_dir
 
     def with_commands(
         self,

mlrun/model_monitoring/api.py

@@ -436,9 +436,9 @@ def _generate_model_endpoint(
         ] = possible_drift_threshold
 
     model_endpoint.spec.monitoring_mode = monitoring_mode
-    model_endpoint.status.first_request = (
-        model_endpoint.status.last_request
-    ) = datetime_now().isoformat()
+    model_endpoint.status.first_request = model_endpoint.status.last_request = (
+        datetime_now().isoformat()
+    )
     if sample_set_statistics:
         model_endpoint.status.feature_stats = sample_set_statistics
 

mlrun/model_monitoring/application.py

@@ -267,7 +267,7 @@ class PushToMonitoringWriter(StepToDict):
         self.project = project
         self.application_name_to_push = writer_application_name
         self.stream_uri = stream_uri or get_stream_path(
-            project=self.project, application_name=self.application_name_to_push
+            project=self.project, function_name=self.application_name_to_push
         )
         self.output_stream = None
         self.name = name or "PushToMonitoringWriter"

mlrun/model_monitoring/applications/__init__.py

@@ -0,0 +1,13 @@
+# Copyright 2024 Iguazio
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.