mlrun 1.7.0rc3__py3-none-any.whl → 1.7.0rc5__py3-none-any.whl

mlrun/artifacts/manager.py

@@ -17,7 +17,11 @@ from os.path import exists, isdir
 from urllib.parse import urlparse
 
 import mlrun.config
-from mlrun.utils.helpers import get_local_file_schema, template_artifact_path
+from mlrun.utils.helpers import (
+    get_local_file_schema,
+    template_artifact_path,
+    validate_inline_artifact_body_size,
+)
 
 from ..utils import (
     is_legacy_artifact,
@@ -212,6 +216,7 @@ class ArtifactManager:
             target_path = target_path or item.target_path
 
         validate_artifact_key_name(key, "artifact.key")
+        validate_inline_artifact_body_size(item.spec.inline)
         src_path = local_path or item.src_path  # TODO: remove src_path
         self.ensure_artifact_source_file_exists(item=item, path=src_path, body=body)
         if format == "html" or (src_path and pathlib.Path(src_path).suffix == "html"):

mlrun/common/constants.py

@@ -13,3 +13,5 @@
 # limitations under the License.
 #
 IMAGE_NAME_ENRICH_REGISTRY_PREFIX = "."  # prefix for image name to enrich with registry
+MLRUN_CREATED_LABEL = "mlrun-created"
+MYSQL_MEDIUMBLOB_SIZE_BYTES = 16 * 1024 * 1024

mlrun/common/model_monitoring/helpers.py

@@ -16,6 +16,7 @@ import sys
 import typing
 
 import mlrun.common
+import mlrun.common.schemas.model_monitoring.constants as mm_constants
 from mlrun.common.schemas.model_monitoring import (
     EndpointUID,
     FunctionURI,
@@ -64,7 +65,7 @@ def parse_model_endpoint_store_prefix(store_prefix: str):
 
 
 def parse_monitoring_stream_path(
-    stream_uri: str, project: str, application_name: str = None
+    stream_uri: str, project: str, function_name: str = None
 ):
     if stream_uri.startswith("kafka://"):
         if "?topic" in stream_uri:
@@ -72,23 +73,28 @@ def parse_monitoring_stream_path(
                 "Custom kafka topic is not allowed"
             )
         # Add topic to stream kafka uri
-        if application_name is None:
+        if (
+            function_name is None
+            or function_name == mm_constants.MonitoringFunctionNames.STREAM
+        ):
             stream_uri += f"?topic=monitoring_stream_{project}"
         else:
-            stream_uri += f"?topic=monitoring_stream_{project}_{application_name}"
+            stream_uri += f"?topic=monitoring_stream_{project}_{function_name}"
 
     elif stream_uri.startswith("v3io://") and mlrun.mlconf.is_ce_mode():
         # V3IO is not supported in CE mode, generating a default http stream path
-        if application_name is None:
+        if function_name is None:
             stream_uri = (
                 mlrun.mlconf.model_endpoint_monitoring.default_http_sink.format(
-                    project=project
+                    project=project, namespace=mlrun.mlconf.namespace
                 )
             )
         else:
             stream_uri = (
                 mlrun.mlconf.model_endpoint_monitoring.default_http_sink_app.format(
-                    project=project, application_name=application_name
+                    project=project,
+                    application_name=function_name,
+                    namespace=mlrun.mlconf.namespace,
                 )
             )
     return stream_uri

mlrun/common/schemas/__init__.py

@@ -14,6 +14,16 @@
 #
 # flake8: noqa  - this is until we take care of the F401 violations with respect to __all__ & sphinx
 
+from .api_gateway import (
+    APIGateway,
+    APIGatewayAuthenticationMode,
+    APIGatewayBasicAuth,
+    APIGatewayMetadata,
+    APIGatewaysOutput,
+    APIGatewaySpec,
+    APIGatewayStatus,
+    APIGatewayUpstream,
+)
 from .artifact import (
     Artifact,
     ArtifactCategories,
@@ -43,6 +53,7 @@ from .clusterization_spec import (
     ClusterizationSpec,
     WaitForChiefToReachOnlineStateFeatureFlag,
 )
+from .common import ImageBuilder
 from .constants import (
     APIStates,
     ClusterizationRole,

mlrun/common/schemas/api_gateway.py

@@ -0,0 +1,85 @@
+# Copyright 2023 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.
+#
+import typing
+from typing import Optional
+
+import pydantic
+
+import mlrun.common.types
+
+
+class APIGatewayAuthenticationMode(mlrun.common.types.StrEnum):
+    basic = "basicAuth"
+    none = "none"
+
+    @classmethod
+    def from_str(cls, authentication_mode: str):
+        if authentication_mode == "none":
+            return cls.none
+        elif authentication_mode == "basicAuth":
+            return cls.basic
+        else:
+            raise mlrun.errors.MLRunInvalidArgumentError(
+                f"Authentication mode `{authentication_mode}` is not supported",
+            )
+
+
+class _APIGatewayBaseModel(pydantic.BaseModel):
+    class Config:
+        extra = pydantic.Extra.allow
+
+
+class APIGatewayMetadata(_APIGatewayBaseModel):
+    name: str
+    namespace: Optional[str]
+    labels: Optional[dict] = {}
+
+
+class APIGatewayBasicAuth(_APIGatewayBaseModel):
+    username: str
+    password: str
+
+
+class APIGatewayUpstream(_APIGatewayBaseModel):
+    kind: Optional[str] = "nucliofunction"
+    nucliofunction: dict[str, str]
+    percentage: Optional[int] = 0
+
+
+class APIGatewaySpec(_APIGatewayBaseModel):
+    name: str
+    description: Optional[str]
+    path: Optional[str] = "/"
+    authenticationMode: Optional[APIGatewayAuthenticationMode] = (
+        APIGatewayAuthenticationMode.none
+    )
+    upstreams: list[APIGatewayUpstream]
+    authentication: Optional[dict[str, Optional[APIGatewayBasicAuth]]]
+    host: Optional[str]
+
+
+class APIGatewayStatus(_APIGatewayBaseModel):
+    name: Optional[str]
+    state: Optional[str]
+
+
+class APIGateway(_APIGatewayBaseModel):
+    metadata: APIGatewayMetadata
+    spec: APIGatewaySpec
+    status: Optional[APIGatewayStatus]
+
+
+class APIGatewaysOutput(_APIGatewayBaseModel):
+    api_gateways: typing.Optional[dict[str, APIGateway]] = {}

mlrun/common/schemas/auth.py

@@ -59,7 +59,7 @@ class AuthorizationResourceTypes(mlrun.common.types.StrEnum):
     hub_source = "hub-source"
     workflow = "workflow"
     datastore_profile = "datastore-profile"
-    api_gateways = "api-gateways"
+    api_gateway = "api-gateway"
 
     def to_resource_string(
         self,
@@ -94,7 +94,7 @@ class AuthorizationResourceTypes(mlrun.common.types.StrEnum):
             AuthorizationResourceTypes.hub_source: "/marketplace/sources",
             # workflow define how to run a pipeline and can be considered as the specification of a pipeline.
             AuthorizationResourceTypes.workflow: "/projects/{project_name}/workflows/{resource_name}",
-            AuthorizationResourceTypes.api_gateways: "/projects/{project_name}/api-gateways",
+            AuthorizationResourceTypes.api_gateway: "/projects/{project_name}/api-gateways/{resource_name}",
         }[self].format(project_name=project_name, resource_name=resource_name)
 
 

mlrun/common/schemas/client_spec.py

@@ -29,6 +29,7 @@ class ClientSpec(pydantic.BaseModel):
     ui_url: typing.Optional[str]
     artifact_path: typing.Optional[str]
     feature_store_data_prefixes: typing.Optional[dict[str, str]]
+    feature_store_default_targets: typing.Optional[str]
     spark_app_image: typing.Optional[str]
     spark_app_image_tag: typing.Optional[str]
     spark_history_server_path: typing.Optional[str]

mlrun/common/schemas/common.py

@@ -0,0 +1,40 @@
+# Copyright 2023 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.
+#
+import typing
+
+import pydantic
+
+
+class ImageBuilder(pydantic.BaseModel):
+    functionSourceCode: typing.Optional[str] = None
+    codeEntryType: typing.Optional[str] = None
+    codeEntryAttributes: typing.Optional[str] = None
+    source: typing.Optional[str] = None
+    code_origin: typing.Optional[str] = None
+    origin_filename: typing.Optional[str] = None
+    image: typing.Optional[str] = None
+    base_image: typing.Optional[str] = None
+    commands: typing.Optional[list] = None
+    extra: typing.Optional[str] = None
+    extra_args: typing.Optional[dict] = None
+    builder_env: typing.Optional[dict] = None
+    secret: typing.Optional[str] = None
+    registry: typing.Optional[str] = None
+    load_source_on_run: typing.Optional[bool] = None
+    with_mlrun: typing.Optional[bool] = None
+    auto_build: typing.Optional[bool] = None
+    build_pod: typing.Optional[str] = None
+    requirements: typing.Optional[list] = None
+    source_code_target_dir: typing.Optional[str] = None

mlrun/common/schemas/model_monitoring/constants.py

@@ -181,7 +181,7 @@ class MonitoringFunctionNames:
     WRITER = "model-monitoring-writer"
     BATCH = "model-monitoring-batch"
     APPLICATION_CONTROLLER = "model-monitoring-controller"
-    STREAM = None
+    STREAM = "model-monitoring-stream"
 
     @staticmethod
     def all():
@@ -289,3 +289,6 @@ class ModelMonitoringAppLabel:
 
 class ControllerPolicy:
     BASE_PERIOD = "base_period"
+
+
+MLRUN_HISTOGRAM_DATA_DRIFT_APP_NAME = "histogram-data-drift"

mlrun/common/schemas/project.py

@@ -19,6 +19,7 @@ import pydantic
 
 import mlrun.common.types
 
+from .common import ImageBuilder
 from .object import ObjectKind, ObjectStatus
 
 
@@ -85,6 +86,7 @@ class ProjectSpec(pydantic.BaseModel):
     desired_state: typing.Optional[ProjectDesiredState] = ProjectDesiredState.online
     custom_packagers: typing.Optional[list[tuple[str, bool]]] = None
     default_image: typing.Optional[str] = None
+    build: typing.Optional[ImageBuilder] = None
 
     class Config:
         extra = pydantic.Extra.allow

mlrun/config.py

@@ -287,6 +287,12 @@ default_config = {
         "state": "online",
         "retry_api_call_on_exception": "enabled",
         "http_connection_timeout_keep_alive": 11,
+        # http client used by httpdb
+        "http": {
+            # when True, the client will verify the server's TLS
+            # set to False for backwards compatibility.
+            "verify": False,
+        },
         "db": {
             "commit_retry_timeout": 30,
             "commit_retry_interval": 3,
@@ -484,8 +490,8 @@ default_config = {
         "offline_storage_path": "model-endpoints/{kind}",
         # Default http path that points to the monitoring stream nuclio function. Will be used as a stream path
         # when the user is working in CE environment and has not provided any stream path.
-        "default_http_sink": "http://nuclio-{project}-model-monitoring-stream.mlrun.svc.cluster.local:8080",
-        "default_http_sink_app": "http://nuclio-{project}-{application_name}.mlrun.svc.cluster.local:8080",
+        "default_http_sink": "http://nuclio-{project}-model-monitoring-stream.{namespace}.svc.cluster.local:8080",
+        "default_http_sink_app": "http://nuclio-{project}-{application_name}.{namespace}.svc.cluster.local:8080",
         "batch_processing_function_branch": "master",
         "parquet_batching_max_events": 10_000,
         "parquet_batching_timeout_secs": timedelta(minutes=1).total_seconds(),
@@ -605,7 +611,7 @@ default_config = {
     "workflows": {
         "default_workflow_runner_name": "workflow-runner-{}",
         # Default timeout seconds for retrieving workflow id after execution:
-        "timeouts": {"local": 120, "kfp": 30, "remote": 30},
+        "timeouts": {"local": 120, "kfp": 30, "remote": 90},
     },
     "log_collector": {
         "address": "localhost:8282",
@@ -957,10 +963,10 @@ class Config:
             with_gpu = (
                 with_gpu_requests if requirement == "requests" else with_gpu_limits
             )
-            resources[
-                requirement
-            ] = self.get_default_function_pod_requirement_resources(
-                requirement, with_gpu
+            resources[requirement] = (
+                self.get_default_function_pod_requirement_resources(
+                    requirement, with_gpu
+                )
             )
         return resources
 
@@ -1053,7 +1059,7 @@ class Config:
         kind: str = "",
         target: str = "online",
         artifact_path: str = None,
-        application_name: str = None,
+        function_name: str = None,
     ) -> str:
         """Get the full path from the configuration based on the provided project and kind.
 
@@ -1068,7 +1074,7 @@ class Config:
                                 artifact path instead.
         :param artifact_path:   Optional artifact path that will be used as a relative path. If not provided, the
                                 relative artifact path will be taken from the global MLRun artifact path.
-        :param application_name:    Application name, None for model_monitoring_stream.
+        :param function_name:    Application name, None for model_monitoring_stream.
 
         :return:                Full configured path for the provided kind.
         """
@@ -1082,20 +1088,19 @@ class Config:
                 return store_prefix_dict[kind].format(project=project)
 
             if (
-                application_name
+                function_name
+                and function_name
                 != mlrun.common.schemas.model_monitoring.constants.MonitoringFunctionNames.STREAM
             ):
                 return mlrun.mlconf.model_endpoint_monitoring.store_prefixes.user_space.format(
                     project=project,
                     kind=kind
-                    if application_name is None
-                    else f"{kind}-{application_name.lower()}",
+                    if function_name is None
+                    else f"{kind}-{function_name.lower()}",
                 )
             return mlrun.mlconf.model_endpoint_monitoring.store_prefixes.default.format(
                 project=project,
-                kind=kind
-                if application_name is None
-                else f"{kind}-{application_name.lower()}",
+                kind=kind,
             )
 
         # Get the current offline path from the configuration
@@ -1343,12 +1348,21 @@ def read_env(env=None, prefix=env_prefix):
         if igz_domain:
             config["ui_url"] = f"https://mlrun-ui.{igz_domain}"
 
-    if config.get("log_level"):
+    if log_level := config.get("log_level"):
         import mlrun.utils.logger
 
         # logger created (because of imports mess) before the config is loaded (in tests), therefore we're changing its
         # level manually
-        mlrun.utils.logger.set_logger_level(config["log_level"])
+        mlrun.utils.logger.set_logger_level(log_level)
+
+    if log_formatter_name := config.get("log_formatter"):
+        import mlrun.utils.logger
+
+        log_formatter = mlrun.utils.create_formatter_instance(
+            mlrun.utils.FormatterKinds(log_formatter_name)
+        )
+        mlrun.utils.logger.get_handler("default").setFormatter(log_formatter)
+
     # The default function pod resource values are of type str; however, when reading from environment variable numbers,
     # it converts them to type int if contains only number, so we want to convert them to str.
     _convert_resources_to_str(config)

mlrun/datastore/azure_blob.py

@@ -175,9 +175,9 @@ class AzureBlobStore(DataStore):
 
         if "client_secret" in st or "client_id" in st or "tenant_id" in st:
             res[f"spark.hadoop.fs.azure.account.auth.type.{host}"] = "OAuth"
-            res[
-                f"spark.hadoop.fs.azure.account.oauth.provider.type.{host}"
-            ] = "org.apache.hadoop.fs.azurebfs.oauth2.ClientCredsTokenProvider"
+            res[f"spark.hadoop.fs.azure.account.oauth.provider.type.{host}"] = (
+                "org.apache.hadoop.fs.azurebfs.oauth2.ClientCredsTokenProvider"
+            )
             if "client_id" in st:
                 res[f"spark.hadoop.fs.azure.account.oauth2.client.id.{host}"] = st[
                     "client_id"
@@ -188,14 +188,27 @@ class AzureBlobStore(DataStore):
                 ]
             if "tenant_id" in st:
                 tenant_id = st["tenant_id"]
-                res[
-                    f"spark.hadoop.fs.azure.account.oauth2.client.endpoint.{host}"
-                ] = f"https://login.microsoftonline.com/{tenant_id}/oauth2/token"
+                res[f"spark.hadoop.fs.azure.account.oauth2.client.endpoint.{host}"] = (
+                    f"https://login.microsoftonline.com/{tenant_id}/oauth2/token"
+                )
 
         if "sas_token" in st:
             res[f"spark.hadoop.fs.azure.account.auth.type.{host}"] = "SAS"
-            res[
-                f"spark.hadoop.fs.azure.sas.token.provider.type.{host}"
-            ] = "org.apache.hadoop.fs.azurebfs.sas.FixedSASTokenProvider"
+            res[f"spark.hadoop.fs.azure.sas.token.provider.type.{host}"] = (
+                "org.apache.hadoop.fs.azurebfs.sas.FixedSASTokenProvider"
+            )
             res[f"spark.hadoop.fs.azure.sas.fixed.token.{host}"] = st["sas_token"]
         return res
+
+    @property
+    def spark_url(self):
+        spark_options = self.get_spark_options()
+        url = f"wasbs://{self.endpoint}"
+        prefix = "spark.hadoop.fs.azure.account.key."
+        if spark_options:
+            for key in spark_options:
+                if key.startswith(prefix):
+                    account_key = key[len(prefix) :]
+                    url += f"@{account_key}"
+                    break
+        return url

mlrun/datastore/base.py

@@ -147,6 +147,10 @@ class DataStore:
     def url(self):
         return f"{self.kind}://{self.endpoint}"
 
+    @property
+    def spark_url(self):
+        return self.url
+
     def get(self, key, size=None, offset=0):
         pass
 
@@ -320,31 +324,17 @@ class DataStore:
             raise Exception(f"File type unhandled {url}")
 
         if file_system:
-            if (
-                self.supports_isdir()
-                and file_system.isdir(file_url)
-                or self._is_dd(df_module)
-            ):
-                storage_options = self.get_storage_options()
-                if url.startswith("ds://"):
-                    parsed_url = urllib.parse.urlparse(url)
-                    url = parsed_url.path
-                    if self.using_bucket:
-                        url = url[1:]
-                    # Pass the underlying file system
-                    kwargs["filesystem"] = file_system
-                elif storage_options:
-                    kwargs["storage_options"] = storage_options
-                df = reader(url, **kwargs)
-            else:
-                file = url
-                # Workaround for ARROW-12472 affecting pyarrow 3.x and 4.x.
-                if file_system.protocol != "file":
-                    # If not dir, use file_system.open() to avoid regression when pandas < 1.2 and does not
-                    # support the storage_options parameter.
-                    file = file_system.open(url)
-
-                df = reader(file, **kwargs)
+            storage_options = self.get_storage_options()
+            if url.startswith("ds://"):
+                parsed_url = urllib.parse.urlparse(url)
+                url = parsed_url.path
+                if self.using_bucket:
+                    url = url[1:]
+                # Pass the underlying file system
+                kwargs["filesystem"] = file_system
+            elif storage_options:
+                kwargs["storage_options"] = storage_options
+            df = reader(url, **kwargs)
         else:
             temp_file = tempfile.NamedTemporaryFile(delete=False)
             self.download(self._join(subpath), temp_file.name)

mlrun/datastore/datastore.py

@@ -94,6 +94,10 @@ def schema_to_store(schema):
         from .dbfs_store import DBFSStore
 
         return DBFSStore
+    elif schema == "hdfs":
+        from .hdfs import HdfsStore
+
+        return HdfsStore
     else:
         raise ValueError(f"unsupported store scheme ({schema})")
 
@@ -170,7 +174,7 @@ class StoreManager:
             raise mlrun.errors.MLRunInvalidArgumentError(
                 f"resource {url} does not have a valid/persistent offline target"
             )
-        return resource, target
+        return resource, target or ""
 
     def object(
         self, url, key="", project="", allow_empty_resources=None, secrets: dict = None
@@ -182,14 +186,21 @@ class StoreManager:
                 url, project, allow_empty_resources, secrets
             )
 
-        store, subpath = self.get_or_create_store(
+        store, subpath, url = self.get_or_create_store(
             url, secrets=secrets, project_name=project
         )
-        return DataItem(key, store, subpath, url, meta=meta, artifact_url=artifact_url)
+        return DataItem(
+            key,
+            store,
+            subpath,
+            url,
+            meta=meta,
+            artifact_url=artifact_url,
+        )
 
     def get_or_create_store(
         self, url, secrets: dict = None, project_name=""
-    ) -> (DataStore, str):
+    ) -> (DataStore, str, str):
         schema, endpoint, parsed_url = parse_url(url)
         subpath = parsed_url.path
         store_key = f"{schema}://{endpoint}"
@@ -206,17 +217,17 @@ class StoreManager:
 
         if schema == "memory":
             subpath = url[len("memory://") :]
-            return in_memory_store, subpath
+            return in_memory_store, subpath, url
 
         if not schema and endpoint:
             if endpoint in self._stores.keys():
-                return self._stores[endpoint], subpath
+                return self._stores[endpoint], subpath, url
             else:
                 raise ValueError(f"no such store ({endpoint})")
 
         if not secrets and not mlrun.config.is_running_as_api():
             if store_key in self._stores.keys():
-                return self._stores[store_key], subpath
+                return self._stores[store_key], subpath, url
 
         # support u/p embedding in url (as done in redis) by setting netloc as the "endpoint" parameter
         # when running on server we don't cache the datastore, because there are multiple users and we don't want to
@@ -227,7 +238,7 @@ class StoreManager:
         if not secrets and not mlrun.config.is_running_as_api():
             self._stores[store_key] = store
         # in file stores in windows path like c:\a\b the drive letter is dropped from the path, so we return the url
-        return store, url if store.kind == "file" else subpath
+        return store, url if store.kind == "file" else subpath, url
 
     def reset_secrets(self):
         self._secrets = {}