mlrun 1.7.0rc13__py3-none-any.whl → 1.7.0rc15__py3-none-any.whl

mlrun/datastore/targets.py

@@ -656,6 +656,29 @@ class BaseStoreTarget(DataTargetBase):
     def _target_path_object(self):
         """return the actual/computed target path"""
         is_single_file = hasattr(self, "is_single_file") and self.is_single_file()
+
+        if self._resource and self.path:
+            parsed_url = urlparse(self.path)
+            # When the URL consists only from scheme and endpoint and no path,
+            # make a default path for DS and redis targets.
+            # Also ignore KafkaTarget when it uses the ds scheme (no default path for KafkaTarget)
+            if (
+                not isinstance(self, KafkaTarget)
+                and parsed_url.scheme in ["ds", "redis", "rediss"]
+                and (not parsed_url.path or parsed_url.path == "/")
+            ):
+                return TargetPathObject(
+                    _get_target_path(
+                        self,
+                        self._resource,
+                        self.run_id is not None,
+                        netloc=parsed_url.netloc,
+                        scheme=parsed_url.scheme,
+                    ),
+                    self.run_id,
+                    is_single_file,
+                )
+
         return self.get_path() or (
             TargetPathObject(
                 _get_target_path(self, self._resource, self.run_id is not None),
@@ -714,9 +737,13 @@ class BaseStoreTarget(DataTargetBase):
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
         """return the target data as dataframe"""
+        mlrun.utils.helpers.additional_filters_warning(
+            additional_filters, self.__class__
+        )
         return mlrun.get_dataitem(self.get_target_path()).as_df(
             columns=columns,
             df_module=df_module,
@@ -961,6 +988,7 @@ class ParquetTarget(BaseStoreTarget):
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
         """return the target data as dataframe"""
@@ -971,6 +999,7 @@ class ParquetTarget(BaseStoreTarget):
             start_time=start_time,
             end_time=end_time,
             time_column=time_column,
+            additional_filters=additional_filters,
             **kwargs,
         )
         if not columns:
@@ -1101,8 +1130,12 @@ class CSVTarget(BaseStoreTarget):
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
+        mlrun.utils.helpers.additional_filters_warning(
+            additional_filters, self.__class__
+        )
         df = super().as_df(
             columns=columns,
             df_module=df_module,
@@ -1209,6 +1242,7 @@ class SnowflakeTarget(BaseStoreTarget):
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
         raise NotImplementedError()
@@ -1275,7 +1309,17 @@ class NoSqlBaseTarget(BaseStoreTarget):
     def get_dask_options(self):
         return {"format": "csv"}
 
-    def as_df(self, columns=None, df_module=None, **kwargs):
+    def as_df(
+        self,
+        columns=None,
+        df_module=None,
+        entities=None,
+        start_time=None,
+        end_time=None,
+        time_column=None,
+        additional_filters=None,
+        **kwargs,
+    ):
         raise NotImplementedError()
 
     def write_dataframe(
@@ -1511,11 +1555,40 @@ class StreamTarget(BaseStoreTarget):
             **self.attributes,
         )
 
-    def as_df(self, columns=None, df_module=None, **kwargs):
+    def as_df(
+        self,
+        columns=None,
+        df_module=None,
+        entities=None,
+        start_time=None,
+        end_time=None,
+        time_column=None,
+        additional_filters=None,
+        **kwargs,
+    ):
         raise NotImplementedError()
 
 
 class KafkaTarget(BaseStoreTarget):
+    """
+    Kafka target storage driver, used to write data into kafka topics.
+    example::
+        # define target
+        kafka_target = KafkaTarget(
+            name="kafka", path="my_topic", brokers="localhost:9092"
+        )
+        # ingest
+        stocks_set.ingest(stocks, [kafka_target])
+    :param name:                target name
+    :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"]
+    """
+
     kind = TargetTypes.kafka
     is_table = False
     is_online = False
@@ -1597,7 +1670,17 @@ class KafkaTarget(BaseStoreTarget):
             **attributes,
         )
 
-    def as_df(self, columns=None, df_module=None, **kwargs):
+    def as_df(
+        self,
+        columns=None,
+        df_module=None,
+        entities=None,
+        start_time=None,
+        end_time=None,
+        time_column=None,
+        additional_filters=None,
+        **kwargs,
+    ):
         raise NotImplementedError()
 
     def purge(self):
@@ -1644,7 +1727,17 @@ class TSDBTarget(BaseStoreTarget):
             **self.attributes,
         )
 
-    def as_df(self, columns=None, df_module=None, **kwargs):
+    def as_df(
+        self,
+        columns=None,
+        df_module=None,
+        entities=None,
+        start_time=None,
+        end_time=None,
+        time_column=None,
+        additional_filters=None,
+        **kwargs,
+    ):
         raise NotImplementedError()
 
     def write_dataframe(
@@ -1755,11 +1848,16 @@ class DFTarget(BaseStoreTarget):
         self,
         columns=None,
         df_module=None,
+        entities=None,
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
+        mlrun.utils.helpers.additional_filters_warning(
+            additional_filters, self.__class__
+        )
         return select_columns_from_df(
             filter_df_start_end_time(
                 self._df,
@@ -1934,6 +2032,7 @@ class SQLTarget(BaseStoreTarget):
         start_time=None,
         end_time=None,
         time_column=None,
+        additional_filters=None,
         **kwargs,
     ):
         try:
@@ -1942,6 +2041,10 @@ class SQLTarget(BaseStoreTarget):
         except (ModuleNotFoundError, ImportError) as exc:
             self._raise_sqlalchemy_import_error(exc)
 
+        mlrun.utils.helpers.additional_filters_warning(
+            additional_filters, self.__class__
+        )
+
         db_path, table_name, _, _, _, _ = self._parse_url()
         engine = sqlalchemy.create_engine(db_path)
         parse_dates: Optional[list[str]] = self.attributes.get("parse_dates")
@@ -2088,7 +2191,7 @@ kind_to_driver = {
 }
 
 
-def _get_target_path(driver, resource, run_id_mode=False):
+def _get_target_path(driver, resource, run_id_mode=False, netloc=None, scheme=""):
     """return the default target path given the resource and target kind"""
     kind = driver.kind
     suffix = driver.suffix
@@ -2105,11 +2208,27 @@ def _get_target_path(driver, resource, run_id_mode=False):
     )
     name = resource.metadata.name
     project = resource.metadata.project or mlrun.mlconf.default_project
-    data_prefix = get_default_prefix_for_target(kind).format(
+
+    default_kind_name = kind
+    if scheme == "ds":
+        # "dsnosql" is not an actual target like Parquet or Redis; rather, it serves
+        # as a placeholder that can be used in any specified target
+        default_kind_name = "dsnosql"
+    if scheme == "redis" or scheme == "rediss":
+        default_kind_name = TargetTypes.redisnosql
+
+    netloc = netloc or ""
+    data_prefix = get_default_prefix_for_target(default_kind_name).format(
+        ds_profile_name=netloc,  # In case of ds profile, set its the name
+        authority=netloc,  # In case of redis, replace {authority} with netloc
         project=project,
         kind=kind,
         name=name,
     )
+
+    if scheme == "rediss":
+        data_prefix = data_prefix.replace("redis://", "rediss://", 1)
+
     # todo: handle ver tag changes, may need to copy files?
     if not run_id_mode:
         version = resource.metadata.tag

mlrun/datastore/v3io.py

@@ -29,7 +29,7 @@ from .base import (
 )
 
 V3IO_LOCAL_ROOT = "v3io"
-V3IO_DEFAULT_UPLOAD_CHUNK_SIZE = 1024 * 1024 * 100
+V3IO_DEFAULT_UPLOAD_CHUNK_SIZE = 1024 * 1024 * 10
 
 
 class V3ioStore(DataStore):

mlrun/db/auth_utils.py

@@ -0,0 +1,152 @@
+# 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.
+
+from abc import ABC, abstractmethod
+from datetime import datetime, timedelta
+
+import requests
+
+import mlrun.errors
+from mlrun.utils import logger
+
+
+class TokenProvider(ABC):
+    @abstractmethod
+    def get_token(self):
+        pass
+
+    @abstractmethod
+    def is_iguazio_session(self):
+        pass
+
+
+class StaticTokenProvider(TokenProvider):
+    def __init__(self, token: str):
+        self.token = token
+
+    def get_token(self):
+        return self.token
+
+    def is_iguazio_session(self):
+        return mlrun.platforms.iguazio.is_iguazio_session(self.token)
+
+
+class OAuthClientIDTokenProvider(TokenProvider):
+    def __init__(
+        self, token_endpoint: str, client_id: str, client_secret: str, timeout=5
+    ):
+        if not token_endpoint or not client_id or not client_secret:
+            raise mlrun.errors.MLRunValueError(
+                "Invalid client_id configuration for authentication. Must provide token endpoint, client-id and secret"
+            )
+        self.token_endpoint = token_endpoint
+        self.client_id = client_id
+        self.client_secret = client_secret
+        self.timeout = timeout
+
+        # Since we're only issuing POST requests, which are actually a disguised GET, then it's ok to allow retries
+        # on them.
+        self._session = mlrun.utils.HTTPSessionWithRetry(
+            retry_on_post=True,
+            verbose=True,
+        )
+
+        self._cleanup()
+        self._refresh_token_if_needed()
+
+    def get_token(self):
+        self._refresh_token_if_needed()
+        return self.token
+
+    def is_iguazio_session(self):
+        return False
+
+    def _cleanup(self):
+        self.token = self.token_expiry_time = self.token_refresh_time = None
+
+    def _refresh_token_if_needed(self):
+        now = datetime.now()
+        if self.token:
+            if self.token_refresh_time and now <= self.token_refresh_time:
+                return self.token
+
+            # We only cleanup if token was really expired - even if we fail in refreshing the token, we can still
+            # use the existing one given that it's not expired.
+            if now >= self.token_expiry_time:
+                self._cleanup()
+
+        self._issue_token_request()
+        return self.token
+
+    def _issue_token_request(self, raise_on_error=False):
+        try:
+            headers = {"Content-Type": "application/x-www-form-urlencoded"}
+            request_body = {
+                "grant_type": "client_credentials",
+                "client_id": self.client_id,
+                "client_secret": self.client_secret,
+            }
+            response = self._session.request(
+                "POST",
+                self.token_endpoint,
+                timeout=self.timeout,
+                headers=headers,
+                data=request_body,
+            )
+        except requests.RequestException as exc:
+            error = f"Retrieving token failed: {mlrun.errors.err_to_str(exc)}"
+            if raise_on_error:
+                raise mlrun.errors.MLRunRuntimeError(error) from exc
+            else:
+                logger.warning(error)
+                return
+
+        if not response.ok:
+            error = "No error available"
+            if response.content:
+                try:
+                    data = response.json()
+                    error = data.get("error")
+                except Exception:
+                    pass
+            logger.warning(
+                "Retrieving token failed", status=response.status_code, error=error
+            )
+            if raise_on_error:
+                mlrun.errors.raise_for_status(response)
+            return
+
+        self._parse_response(response.json())
+
+    def _parse_response(self, data: dict):
+        # Response is described in https://datatracker.ietf.org/doc/html/rfc6749#section-4.4.3
+        # According to spec, there isn't a refresh token - just the access token and its expiry time (in seconds).
+        self.token = data.get("access_token")
+        expires_in = data.get("expires_in")
+        if not self.token or not expires_in:
+            token_str = "****" if self.token else "missing"
+            logger.warning(
+                "Failed to parse token response", token=token_str, expires_in=expires_in
+            )
+            return
+
+        now = datetime.now()
+        self.token_expiry_time = now + timedelta(seconds=expires_in)
+        self.token_refresh_time = now + timedelta(seconds=expires_in / 2)
+        logger.info(
+            "Successfully retrieved client-id token",
+            expires_in=expires_in,
+            expiry=str(self.token_expiry_time),
+            refresh=str(self.token_refresh_time),
+        )

mlrun/db/base.py

@@ -802,7 +802,7 @@ class RunDBInterface(ABC):
         project: str,
         base_period: int = 10,
         image: str = "mlrun/mlrun",
-    ):
+    ) -> None:
         pass
 
     @abstractmethod

mlrun/db/httpdb.py

@@ -38,6 +38,7 @@ import mlrun.platforms
 import mlrun.projects
 import mlrun.runtimes.nuclio.api_gateway
 import mlrun.utils
+from mlrun.db.auth_utils import OAuthClientIDTokenProvider, StaticTokenProvider
 from mlrun.errors import MLRunInvalidArgumentError, err_to_str
 
 from ..artifacts import Artifact
@@ -138,17 +139,28 @@ class HTTPRunDB(RunDBInterface):
             endpoint += f":{parsed_url.port}"
         base_url = f"{parsed_url.scheme}://{endpoint}{parsed_url.path}"
 
+        self.base_url = base_url
         username = parsed_url.username or config.httpdb.user
         password = parsed_url.password or config.httpdb.password
+        self.token_provider = None
 
-        username, password, token = mlrun.platforms.add_or_refresh_credentials(
-            parsed_url.hostname, username, password, config.httpdb.token
-        )
+        if config.auth_with_client_id.enabled:
+            self.token_provider = OAuthClientIDTokenProvider(
+                token_endpoint=mlrun.get_secret_or_env("MLRUN_AUTH_TOKEN_ENDPOINT"),
+                client_id=mlrun.get_secret_or_env("MLRUN_AUTH_CLIENT_ID"),
+                client_secret=mlrun.get_secret_or_env("MLRUN_AUTH_CLIENT_SECRET"),
+                timeout=config.auth_with_client_id.request_timeout,
+            )
+        else:
+            username, password, token = mlrun.platforms.add_or_refresh_credentials(
+                parsed_url.hostname, username, password, config.httpdb.token
+            )
+
+            if token:
+                self.token_provider = StaticTokenProvider(token)
 
-        self.base_url = base_url
         self.user = username
         self.password = password
-        self.token = token
 
     def __repr__(self):
         cls = self.__class__.__name__
@@ -218,17 +230,19 @@ class HTTPRunDB(RunDBInterface):
 
         if self.user:
             kw["auth"] = (self.user, self.password)
-        elif self.token:
-            # Iguazio auth doesn't support passing token through bearer, so use cookie instead
-            if mlrun.platforms.iguazio.is_iguazio_session(self.token):
-                session_cookie = f'j:{{"sid": "{self.token}"}}'
-                cookies = {
-                    "session": session_cookie,
-                }
-                kw["cookies"] = cookies
-            else:
-                if "Authorization" not in kw.setdefault("headers", {}):
-                    kw["headers"].update({"Authorization": "Bearer " + self.token})
+        elif self.token_provider:
+            token = self.token_provider.get_token()
+            if token:
+                # Iguazio auth doesn't support passing token through bearer, so use cookie instead
+                if self.token_provider.is_iguazio_session():
+                    session_cookie = f'j:{{"sid": "{token}"}}'
+                    cookies = {
+                        "session": session_cookie,
+                    }
+                    kw["cookies"] = cookies
+                else:
+                    if "Authorization" not in kw.setdefault("headers", {}):
+                        kw["headers"].update({"Authorization": "Bearer " + token})
 
         if mlrun.common.schemas.HeaderNames.client_version not in kw.setdefault(
             "headers", {}
@@ -645,10 +659,10 @@ class HTTPRunDB(RunDBInterface):
                 nil_resp += 1
 
             if watch and state in [
-                mlrun.runtimes.constants.RunStates.pending,
-                mlrun.runtimes.constants.RunStates.running,
-                mlrun.runtimes.constants.RunStates.created,
-                mlrun.runtimes.constants.RunStates.aborting,
+                mlrun.common.runtimes.constants.RunStates.pending,
+                mlrun.common.runtimes.constants.RunStates.running,
+                mlrun.common.runtimes.constants.RunStates.created,
+                mlrun.common.runtimes.constants.RunStates.aborting,
             ]:
                 continue
             else:
@@ -1142,7 +1156,29 @@ class HTTPRunDB(RunDBInterface):
         project = project or config.default_project
         path = f"projects/{project}/functions/{name}"
         error_message = f"Failed deleting function {project}/{name}"
-        self.api_call("DELETE", path, error_message)
+        response = self.api_call("DELETE", path, error_message, version="v2")
+        if response.status_code == http.HTTPStatus.ACCEPTED:
+            logger.info(
+                "Function is being deleted", project_name=project, function_name=name
+            )
+            background_task = mlrun.common.schemas.BackgroundTask(**response.json())
+            background_task = self._wait_for_background_task_to_reach_terminal_state(
+                background_task.metadata.name, project=project
+            )
+            if (
+                background_task.status.state
+                == mlrun.common.schemas.BackgroundTaskState.succeeded
+            ):
+                logger.info(
+                    "Function deleted", project_name=project, function_name=name
+                )
+            elif (
+                background_task.status.state
+                == mlrun.common.schemas.BackgroundTaskState.failed
+            ):
+                logger.info(
+                    "Function deletion failed", project_name=project, function_name=name
+                )
 
     def list_functions(self, name=None, project=None, tag=None, labels=None):
         """Retrieve a list of functions, filtered by specific criteria.
@@ -1488,16 +1524,15 @@ class HTTPRunDB(RunDBInterface):
         """
 
         try:
+            normalized_name = normalize_name(func.metadata.name)
             params = {
-                "name": normalize_name(func.metadata.name),
+                "name": normalized_name,
                 "project": func.metadata.project,
                 "tag": func.metadata.tag,
                 "last_log_timestamp": str(last_log_timestamp),
                 "verbose": bool2str(verbose),
             }
-            _path = (
-                f"projects/{func.metadata.project}/nuclio/{func.metadata.name}/deploy"
-            )
+            _path = f"projects/{func.metadata.project}/nuclio/{normalized_name}/deploy"
             resp = self.api_call("GET", _path, params=params)
         except OSError as err:
             logger.error(f"error getting deploy status: {err_to_str(err)}")
@@ -3214,7 +3249,7 @@ class HTTPRunDB(RunDBInterface):
         project: str,
         base_period: int = 10,
         image: str = "mlrun/mlrun",
-    ):
+    ) -> None:
         """
         Redeploy model monitoring application controller function.
 
@@ -3224,13 +3259,14 @@ class HTTPRunDB(RunDBInterface):
         :param image: The image of the model monitoring controller function.
                                          By default, the image is mlrun/mlrun.
         """
-
-        params = {
-            "image": image,
-            "base_period": base_period,
-        }
-        path = f"projects/{project}/model-monitoring/model-monitoring-controller"
-        self.api_call(method="POST", path=path, params=params)
+        self.api_call(
+            method=mlrun.common.types.HTTPMethod.POST,
+            path=f"projects/{project}/model-monitoring/model-monitoring-controller",
+            params={
+                "base_period": base_period,
+                "image": image,
+            },
+        )
 
     def enable_model_monitoring(
         self,

mlrun/feature_store/__init__.py

@@ -19,7 +19,6 @@ __all__ = [
     "get_online_feature_service",
     "ingest",
     "preview",
-    "deploy_ingestion_service",
     "deploy_ingestion_service_v2",
     "delete_feature_set",
     "delete_feature_vector",
@@ -41,7 +40,6 @@ from ..features import Entity, Feature
 from .api import (
     delete_feature_set,
     delete_feature_vector,
-    deploy_ingestion_service,
     deploy_ingestion_service_v2,
     get_feature_set,
     get_feature_vector,