anyscale 0.26.70__py3-none-any.whl → 0.26.72__py3-none-any.whl

anyscale/controllers/job_controller.py

@@ -29,6 +29,7 @@ from anyscale.client.openapi_client.models.decorated_production_job import (
     DecoratedProductionJob,
 )
 from anyscale.client.openapi_client.models.ha_job_states import HaJobStates
+from anyscale.commands.util import flatten_tag_dict_to_api_list
 from anyscale.controllers.base_controller import BaseController
 from anyscale.models.job_model import JobConfig
 from anyscale.project_utils import infer_project_id
@@ -213,6 +214,7 @@ class JobController(BaseController):
                 workspace_id=job_config.workspace_id,
                 config=config_object,
                 job_queue_config=job_queue_config,
+                tags=getattr(job_config, "tags", None),
             )
         ).result
         self.log.info(
@@ -288,6 +290,13 @@ class JobController(BaseController):
         ).result
         return job.state.current_state
 
+    def resolve_job_id(self, job_id: Optional[str], job_name: Optional[str]) -> str:
+        """Resolve and return a job's ID from either an explicit ID or a name.
+
+        Raises click.ClickException if neither is provided or if the name is ambiguous.
+        """
+        return self._resolve_job_object(job_id, job_name).id
+
     def list(  # noqa: PLR0913
         self,
         include_all_users: bool,
@@ -297,6 +306,7 @@ class JobController(BaseController):
         include_archived: bool,
         max_items: int,
         states: List[HaJobStates],
+        tags: Optional[Dict[str, List[str]]] = None,
     ) -> None:
         """
         This function will list jobs.
@@ -349,6 +359,7 @@ class JobController(BaseController):
                 archive_status="ALL" if include_archived else "NOT_ARCHIVED",
                 count=DEFAULT_PAGE_LIMIT,
                 state_filter=states,
+                tag_filter=flatten_tag_dict_to_api_list(tags),
             )
             jobs_list.extend(resp.results)
             paging_token = resp.metadata.next_paging_token
@@ -363,6 +374,7 @@ class JobController(BaseController):
                     count=DEFAULT_PAGE_LIMIT,
                     paging_token=paging_token,
                     state_filter=states,
+                    tag_filter=flatten_tag_dict_to_api_list(tags),
                 )
                 jobs_list.extend(resp.results)
                 paging_token = resp.metadata.next_paging_token

anyscale/controllers/workspace_controller.py

@@ -1,7 +1,7 @@
 import os
 import shlex
 import subprocess
-from typing import Any, List, Optional
+from typing import Any, Dict, List, Optional
 
 import click
 import tabulate
@@ -10,9 +10,19 @@ from anyscale.cli_logger import BlockLogger
 from anyscale.client.openapi_client.models.create_experimental_workspace import (
     CreateExperimentalWorkspace,
 )
+from anyscale.client.openapi_client.models.delete_resource_tags_request import (
+    DeleteResourceTagsRequest,
+)
 from anyscale.client.openapi_client.models.experimental_workspace import (
     ExperimentalWorkspace,
 )
+from anyscale.client.openapi_client.models.resource_tag_resource_type import (
+    ResourceTagResourceType,
+)
+from anyscale.client.openapi_client.models.upsert_resource_tags_request import (
+    UpsertResourceTagsRequest,
+)
+from anyscale.commands.util import flatten_tag_dict_to_api_list
 from anyscale.controllers.base_controller import BaseController
 from anyscale.feature_flags import FLAG_DEFAULT_WORKING_DIR_FOR_PROJ
 from anyscale.project_utils import get_default_project
@@ -84,13 +94,21 @@ class WorkspaceController(BaseController):
             )
         )
 
-    def list(self) -> None:
+    def list(self, tags_filter: Optional[Dict[str, List[str]]] = None) -> None:
         """
         prints a non-exhaustive tabular list of information about non-deleted workspaces.
         """
-        workspaces_data: List[
-            ExperimentalWorkspace
-        ] = self.api_client.list_workspaces_api_v2_experimental_workspaces_get().results
+        params = {}
+        if tags_filter:
+            api_tag_filter = flatten_tag_dict_to_api_list(tags_filter)
+            if api_tag_filter:
+                params["tag_filter"] = api_tag_filter
+
+        workspaces_data: List[ExperimentalWorkspace] = (
+            self.api_client.list_workspaces_api_v2_experimental_workspaces_get(
+                **params
+            ).results
+        )
 
         workspaces_table: List[List[Any]] = [
             [
@@ -109,6 +127,50 @@ class WorkspaceController(BaseController):
 
         print(f"Workspaces:\n{table}")
 
+    def _resolve_workspace_id(
+        self, *, workspace_id: Optional[str], name: Optional[str]
+    ) -> str:
+        if workspace_id:
+            return workspace_id
+        if not name:
+            raise click.ClickException("Provide either --id or --name for workspace.")
+        results = self.api_client.list_workspaces_api_v2_experimental_workspaces_get(
+            name=name
+        ).results
+        if len(results) == 0:
+            raise click.ClickException(f"No workspace with name '{name}' found.")
+        if len(results) > 1:
+            raise click.ClickException(
+                f"Multiple workspaces with name '{name}' found. Please use --id."
+            )
+        return results[0].id
+
+    def add_tags(
+        self,
+        *,
+        workspace_id: Optional[str] = None,
+        name: Optional[str] = None,
+        tags: Dict[str, str],
+    ) -> None:
+        wid = self._resolve_workspace_id(workspace_id=workspace_id, name=name)
+        req = UpsertResourceTagsRequest(
+            resource_type=ResourceTagResourceType.WORKSPACE, resource_id=wid, tags=tags,
+        )
+        self.api_client.upsert_resource_tags_api_v2_tags_resource_put(req)
+
+    def remove_tags(
+        self,
+        *,
+        workspace_id: Optional[str] = None,
+        name: Optional[str] = None,
+        keys: List[str],
+    ) -> None:
+        wid = self._resolve_workspace_id(workspace_id=workspace_id, name=name)
+        req = DeleteResourceTagsRequest(
+            resource_type=ResourceTagResourceType.WORKSPACE, resource_id=wid, keys=keys,
+        )
+        self.api_client.delete_resource_tags_api_v2_tags_resource_delete(req)
+
     def clone(self, workspace: ExperimentalWorkspace) -> None:
         dir_name = workspace.name
         os.makedirs(dir_name)

anyscale/job/_private/job_sdk.py

@@ -46,7 +46,8 @@ HA_JOB_STATE_TO_JOB_STATE = {
     HaJobStates.PENDING: JobState.STARTING,
     HaJobStates.AWAITING_CLUSTER_START: JobState.STARTING,
     HaJobStates.SUCCESS: JobState.SUCCEEDED,
-    HaJobStates.ERRORED: JobState.FAILED,
+    # ERRORED is a transient state that can transition to RESTARTING when retries remain.
+    HaJobStates.ERRORED: JobState.RUNNING,
     HaJobStates.TERMINATED: JobState.FAILED,
     HaJobStates.BROKEN: JobState.FAILED,
     HaJobStates.OUT_OF_RETRIES: JobState.FAILED,
@@ -280,6 +281,7 @@ class PrivateJobSDK(WorkloadSDK):
                 workspace_id=self.client.get_current_workspace_id(),
                 config=prod_job_config,
                 job_queue_config=job_queue_config,
+                tags=getattr(config, "tags", None),
             )
         )
 

anyscale/job/models.py

@@ -300,6 +300,9 @@ py_modules: # (Optional) A list of local directories or remote URIs that will be
 cloud: anyscale-prod # (Optional) The name of the Anyscale Cloud.
 project: my-project # (Optional) The name of the Anyscale Project.
 max_retries: 3 # (Optional) Maximum number of times the job will be retried before being marked failed. Defaults to `1`.
+tags:
+    team: mlops
+    purpose: training
 
 """
 
@@ -377,6 +380,19 @@ max_retries: 3 # (Optional) Maximum number of times the job will be retried befo
             if timeout_s < 0:
                 raise ValueError("'timeout_s' must be >= 0.")
 
+    tags: Optional[Dict[str, str]] = field(
+        default=None, metadata={"docstring": "Tags to associate with the job."},
+    )
+
+    def _validate_tags(self, tags: Optional[Dict[str, str]]):
+        if tags is None:
+            return
+        if not isinstance(tags, dict):
+            raise TypeError("'tags' must be a Dict[str, str].")
+        for k, v in tags.items():
+            if not isinstance(k, str) or not isinstance(v, str):
+                raise TypeError("'tags' must be a Dict[str, str].")
+
 
 class JobRunState(ModelEnum):
     """Current state of an individual job run."""

anyscale/job_queue/__init__.py

@@ -1,4 +1,4 @@
-from typing import List, Optional
+from typing import Dict, List, Optional
 
 from anyscale._private.anyscale_client import AnyscaleClient
 from anyscale._private.models.model_base import ResultIterator
@@ -16,9 +16,15 @@ from anyscale.job_queue.commands import (
     _LIST_EXAMPLE,
     _STATUS_ARG_DOCSTRINGS,
     _STATUS_EXAMPLE,
+    _TAGS_ADD_ARG_DOCSTRINGS,
+    _TAGS_ADD_EXAMPLE,
+    _TAGS_REMOVE_ARG_DOCSTRINGS,
+    _TAGS_REMOVE_EXAMPLE,
     _UPDATE_ARG_DOCSTRINGS,
     _UPDATE_EXAMPLE,
+    add_tags,
     list,
+    remove_tags,
     status,
     update,
 )
@@ -49,6 +55,7 @@ class JobQueueSDK:
         cloud: Optional[str] = None,
         project: Optional[str] = None,
         cluster_status: Optional[SessionState] = None,
+        tags_filter: Optional[Dict[str, List[str]]] = None,
         page_size: Optional[int] = None,
         max_items: Optional[int] = None,
         sorting_directives: Optional[List[JobQueueSortDirective]] = None,
@@ -61,6 +68,7 @@ class JobQueueSDK:
             cloud=cloud,
             project=project,
             cluster_status=cluster_status,
+            tags_filter=tags_filter,
             page_size=page_size,
             max_items=max_items,
             sorting_directives=sorting_directives,
@@ -87,3 +95,31 @@ class JobQueueSDK:
             max_concurrency=max_concurrency,
             idle_timeout_s=idle_timeout_s,
         )
+
+    @sdk_docs(doc_py_example=_TAGS_ADD_EXAMPLE, arg_docstrings=_TAGS_ADD_ARG_DOCSTRINGS)
+    def add_tags(  # noqa: F811
+        self,
+        *,
+        job_queue_id: Optional[str] = None,
+        name: Optional[str] = None,
+        tags: Dict[str, str],
+    ) -> None:
+        """Upsert (add/update) tag key/value pairs for a job queue."""
+        return self._private_sdk.add_tags(
+            job_queue_id=job_queue_id, name=name, tags=tags
+        )
+
+    @sdk_docs(
+        doc_py_example=_TAGS_REMOVE_EXAMPLE, arg_docstrings=_TAGS_REMOVE_ARG_DOCSTRINGS
+    )
+    def remove_tags(  # noqa: F811
+        self,
+        *,
+        job_queue_id: Optional[str] = None,
+        name: Optional[str] = None,
+        keys: List[str],
+    ) -> None:
+        """Remove tags by key from a job queue."""
+        return self._private_sdk.remove_tags(
+            job_queue_id=job_queue_id, name=name, keys=keys
+        )

anyscale/job_queue/_private/job_queue_sdk.py

@@ -1,4 +1,4 @@
-from typing import List, Optional
+from typing import Dict, List, Optional
 
 from anyscale._private.models.model_base import ResultIterator
 from anyscale._private.workload import WorkloadSDK
@@ -12,6 +12,9 @@ from anyscale.client.openapi_client.models.job_queue_sort_directive import (
 from anyscale.client.openapi_client.models.list_response_metadata import (
     ListResponseMetadata,
 )
+from anyscale.client.openapi_client.models.resource_tag_resource_type import (
+    ResourceTagResourceType,
+)
 from anyscale.client.openapi_client.models.session_state import SessionState
 from anyscale.job_queue.models import JobQueueStatus
 
@@ -28,6 +31,7 @@ class PrivateJobQueueSDK(WorkloadSDK):
         cloud: Optional[str] = None,
         project: Optional[str] = None,
         cluster_status: Optional[SessionState] = None,
+        tags_filter: Optional[Dict[str, List[str]]] = None,
         page_size: Optional[int] = None,
         max_items: Optional[int] = None,
         sorting_directives: Optional[List[JobQueueSortDirective]] = None,
@@ -69,6 +73,7 @@ class PrivateJobQueueSDK(WorkloadSDK):
                 cloud=cloud,
                 project=project,
                 cluster_status=cluster_status,
+                tags_filter=tags_filter,
                 count=page_size,
                 paging_token=token,
                 sorting_directives=sorting_directives,
@@ -112,6 +117,28 @@ class PrivateJobQueueSDK(WorkloadSDK):
 
         return _parse_decorated_jq_to_status(updated_jq)
 
+    def add_tags(
+        self,
+        *,
+        job_queue_id: Optional[str] = None,
+        name: Optional[str] = None,
+        tags: Dict[str, str],
+    ) -> None:
+        jq = self._resolve_to_job_queue_model(job_queue_id=job_queue_id, name=name)
+        assert jq.id is not None
+        self.client.upsert_resource_tags(ResourceTagResourceType.JOB_QUEUE, jq.id, tags)
+
+    def remove_tags(
+        self,
+        *,
+        job_queue_id: Optional[str] = None,
+        name: Optional[str] = None,
+        keys: List[str],
+    ) -> None:
+        jq = self._resolve_to_job_queue_model(job_queue_id=job_queue_id, name=name)
+        assert jq.id is not None
+        self.client.delete_resource_tags(ResourceTagResourceType.JOB_QUEUE, jq.id, keys)
+
     def _resolve_to_job_queue_model(
         self, *, job_queue_id: Optional[str] = None, name: Optional[str] = None,
     ) -> DecoratedJobQueue:

anyscale/job_queue/commands.py

@@ -1,4 +1,4 @@
-from typing import List, Optional
+from typing import Dict, List, Optional
 
 from anyscale._private.models.model_base import ResultIterator
 from anyscale._private.sdk import sdk_command
@@ -58,6 +58,30 @@ _UPDATE_ARG_DOCSTRINGS = {
     "idle_timeout_s": "New idle timeout in seconds.",
 }
 
+_TAGS_ADD_EXAMPLE = """
+import anyscale
+
+anyscale.job_queue.add_tags(job_queue_id="jobq_abc123", tags={"team": "mlops"})
+"""
+
+_TAGS_ADD_ARG_DOCSTRINGS = {
+    "job_queue_id": "ID of the job queue to tag (alternative to name).",
+    "name": "Name of the job queue to tag (alternative to ID).",
+    "tags": "Key/value tags to upsert as a map {key: value}.",
+}
+
+_TAGS_REMOVE_EXAMPLE = """
+import anyscale
+
+anyscale.job_queue.remove_tags(job_queue_id="jobq_abc123", keys=["team"])
+"""
+
+_TAGS_REMOVE_ARG_DOCSTRINGS = {
+    "job_queue_id": "ID of the job queue to modify (alternative to name).",
+    "name": "Name of the job queue to modify (alternative to ID).",
+    "keys": "List of tag keys to remove.",
+}
+
 
 @sdk_command(
     _JOB_QUEUE_SDK_SINGLETON_KEY,
@@ -73,6 +97,7 @@ def list(  # noqa: A001
     cloud: Optional[str] = None,
     project: Optional[str] = None,
     cluster_status: Optional[SessionState] = None,
+    tags_filter: Optional[Dict[str, List[str]]] = None,
     page_size: Optional[int] = None,
     max_items: Optional[int] = None,
     sorting_directives: Optional[List[JobQueueSortDirective]] = None,
@@ -86,6 +111,7 @@ def list(  # noqa: A001
         cloud=cloud,
         project=project,
         cluster_status=cluster_status,
+        tags_filter=tags_filter,
         page_size=page_size,
         max_items=max_items,
         sorting_directives=sorting_directives,
@@ -128,3 +154,37 @@ def update(
         max_concurrency=max_concurrency,
         idle_timeout_s=idle_timeout_s,
     )
+
+
+@sdk_command(
+    _JOB_QUEUE_SDK_SINGLETON_KEY,
+    PrivateJobQueueSDK,
+    doc_py_example=_TAGS_ADD_EXAMPLE,
+    arg_docstrings=_TAGS_ADD_ARG_DOCSTRINGS,
+)
+def add_tags(
+    *,
+    job_queue_id: Optional[str] = None,
+    name: Optional[str] = None,
+    tags: Dict[str, str],
+    _private_sdk: Optional[PrivateJobQueueSDK] = None,
+) -> None:
+    """Upsert tags for a job queue."""
+    return _private_sdk.add_tags(job_queue_id=job_queue_id, name=name, tags=tags)  # type: ignore
+
+
+@sdk_command(
+    _JOB_QUEUE_SDK_SINGLETON_KEY,
+    PrivateJobQueueSDK,
+    doc_py_example=_TAGS_REMOVE_EXAMPLE,
+    arg_docstrings=_TAGS_REMOVE_ARG_DOCSTRINGS,
+)
+def remove_tags(
+    *,
+    job_queue_id: Optional[str] = None,
+    name: Optional[str] = None,
+    keys: List[str],
+    _private_sdk: Optional[PrivateJobQueueSDK] = None,
+) -> None:
+    """Remove tags by key from a job queue."""
+    return _private_sdk.remove_tags(job_queue_id=job_queue_id, name=name, keys=keys)  # type: ignore

anyscale/sdk/anyscale_client/__init__.py

@@ -133,6 +133,7 @@ from anyscale_client.models.organization import Organization
 from anyscale_client.models.organization_response import OrganizationResponse
 from anyscale_client.models.page_query import PageQuery
 from anyscale_client.models.pause_schedule import PauseSchedule
+from anyscale_client.models.physical_resources import PhysicalResources
 from anyscale_client.models.production_job import ProductionJob
 from anyscale_client.models.production_job_config import ProductionJobConfig
 from anyscale_client.models.production_job_state_transition import ProductionJobStateTransition

anyscale/sdk/anyscale_client/api/default_api.py

@@ -2696,6 +2696,7 @@ class DefaultApi(object):
 
         :param async_req bool: execute request asynchronously
         :param str cloud_id: The cloud id whose default cluster compute you want to fetch. If None, will use the default cloud.
+        :param str cloud_resource_id: The cloud resource ID to fetch the default cluster compute for. If None, will use the primary cloud resource of the cloud.
         :param _preload_content: if False, the urllib3.HTTPResponse object will
                                  be returned without reading/decoding response
                                  data. Default is True.
@@ -2721,6 +2722,7 @@ class DefaultApi(object):
 
         :param async_req bool: execute request asynchronously
         :param str cloud_id: The cloud id whose default cluster compute you want to fetch. If None, will use the default cloud.
+        :param str cloud_resource_id: The cloud resource ID to fetch the default cluster compute for. If None, will use the primary cloud resource of the cloud.
         :param _return_http_data_only: response data without head status code
                                        and headers
         :param _preload_content: if False, the urllib3.HTTPResponse object will
@@ -2738,7 +2740,8 @@ class DefaultApi(object):
         local_var_params = locals()
 
         all_params = [
-            'cloud_id'
+            'cloud_id',
+            'cloud_resource_id'
         ]
         all_params.extend(
             [
@@ -2765,6 +2768,8 @@ class DefaultApi(object):
         query_params = []
         if 'cloud_id' in local_var_params and local_var_params['cloud_id'] is not None:  # noqa: E501
             query_params.append(('cloud_id', local_var_params['cloud_id']))  # noqa: E501
+        if 'cloud_resource_id' in local_var_params and local_var_params['cloud_resource_id'] is not None:  # noqa: E501
+            query_params.append(('cloud_resource_id', local_var_params['cloud_resource_id']))  # noqa: E501
 
         header_params = {}
 
@@ -2929,6 +2934,7 @@ class DefaultApi(object):
 
         :param async_req bool: execute request asynchronously
         :param str cloud_id: (required)
+        :param str cloud_resource_id:
         :param _preload_content: if False, the urllib3.HTTPResponse object will
                                  be returned without reading/decoding response
                                  data. Default is True.
@@ -2954,6 +2960,7 @@ class DefaultApi(object):
 
         :param async_req bool: execute request asynchronously
         :param str cloud_id: (required)
+        :param str cloud_resource_id:
         :param _return_http_data_only: response data without head status code
                                        and headers
         :param _preload_content: if False, the urllib3.HTTPResponse object will
@@ -2971,7 +2978,8 @@ class DefaultApi(object):
         local_var_params = locals()
 
         all_params = [
-            'cloud_id'
+            'cloud_id',
+            'cloud_resource_id'
         ]
         all_params.extend(
             [
@@ -3002,6 +3010,8 @@ class DefaultApi(object):
             path_params['cloud_id'] = local_var_params['cloud_id']  # noqa: E501
 
         query_params = []
+        if 'cloud_resource_id' in local_var_params and local_var_params['cloud_resource_id'] is not None:  # noqa: E501
+            query_params.append(('cloud_resource_id', local_var_params['cloud_resource_id']))  # noqa: E501
 
         header_params = {}
 

anyscale/sdk/anyscale_client/models/__init__.py

@@ -119,6 +119,7 @@ from anyscale_client.models.organization import Organization
 from anyscale_client.models.organization_response import OrganizationResponse
 from anyscale_client.models.page_query import PageQuery
 from anyscale_client.models.pause_schedule import PauseSchedule
+from anyscale_client.models.physical_resources import PhysicalResources
 from anyscale_client.models.production_job import ProductionJob
 from anyscale_client.models.production_job_config import ProductionJobConfig
 from anyscale_client.models.production_job_state_transition import ProductionJobStateTransition