This is an automated email from the ASF dual-hosted git repository.
yasithdev pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airavata-portals.git
The following commit(s) were added to refs/heads/main by this push:
new e85b7e31a feat(portal): repoint experiment reads to gRPC (Track D, D2)
(#170)
e85b7e31a is described below
commit e85b7e31aae08c5205f5980762b944f64d6eb3bd
Author: Yasith Jayawardana <[email protected]>
AuthorDate: Mon Jun 8 20:03:06 2026 -0400
feat(portal): repoint experiment reads to gRPC (Track D, D2) (#170)
Migrate the core experiment reads from the Thrift client to the gRPC
research facade: ExperimentViewSet.get_instance (get_experiment),
ProjectViewSet.experiments (get_experiments_in_project), and
ExperimentViewSet.jobs (get_job_details). Write paths (create/update/
launch/clone) and FullExperimentViewSet/archive/statistics stay on Thrift
for follow-up.
ExperimentModel is the deepest read model. getExperiment returns the full
processes tree, so the new experiment adapter recursively adapts the
user configuration data (with computational resource scheduling), the
experiment input/output and status lists, the errors, and the
processes -> tasks -> jobs tree, each level with its status list. A new
job_model adapter backs both the jobs action and the nested task jobs.
The status and type enums render as raw integers and proto3 namespaces
their members with a per-enum prefix (proto EXPERIMENT_STATE_CREATED vs
Thrift CREATED), so a new _thrift_enum_prefixed helper strips the prefix
and bridges by name to the Thrift integer the frontend expects, mapping
the zero *_UNKNOWN sentinel to None. Covers ExperimentState,
ProcessState, TaskState, JobState, ExperimentType, and TaskTypes. The
ExperimentSerializer's userHasWriteAccess migrates to the gRPC sharing
helper. The legacy workflow-engine fields (experiment.workflow,
process.processWorkflows) are not adapted (rarely populated) and default
to None/[] as Thrift did for non-workflow experiments.
Verified: manage.py check clean; experiments-in-project returns 200 live
(proving the full path) and a missing experiment maps to an HTTP error;
offline serializer render of a full experiment tree (processes -> tasks
-> jobs with statuses) confirms every state/type enum bridges to the
correct Thrift integer, the scheduling/userConfig nesting renders, and
userHasWriteAccess resolves via sharing.
---
.../django_airavata/apps/api/grpc_adapters.py | 245 +++++++++++++++++++++
.../django_airavata/apps/api/serializers.py | 5 +-
.../django_airavata/apps/api/views.py | 17 +-
3 files changed, 257 insertions(+), 10 deletions(-)
diff --git a/airavata-django-portal/django_airavata/apps/api/grpc_adapters.py
b/airavata-django-portal/django_airavata/apps/api/grpc_adapters.py
index 57b7333ad..b708a1005 100644
--- a/airavata-django-portal/django_airavata/apps/api/grpc_adapters.py
+++ b/airavata-django-portal/django_airavata/apps/api/grpc_adapters.py
@@ -27,6 +27,16 @@ from airavata.model.credential.store.ttypes import
SummaryType as _ThriftSummary
from airavata.model.data.movement.ttypes import (
DataMovementProtocol as _ThriftDataMovementProtocol,
)
+from airavata.model.experiment.ttypes import (
+ ExperimentType as _ThriftExperimentType,
+)
+from airavata.model.status.ttypes import (
+ ExperimentState as _ThriftExperimentState,
+ JobState as _ThriftJobState,
+ ProcessState as _ThriftProcessState,
+ TaskState as _ThriftTaskState,
+)
+from airavata.model.task.ttypes import TaskTypes as _ThriftTaskTypes
def _thrift_enum(pb, field, thrift_enum):
@@ -42,6 +52,22 @@ def _thrift_enum(pb, field, thrift_enum):
return getattr(thrift_enum, name)
+def _thrift_enum_prefixed(pb, field, thrift_enum, proto_prefix):
+ """Bridge a proto enum field to a Thrift value by name after stripping a
+ proto-only prefix.
+
+ proto3 namespaces enum members that would otherwise collide in the file
+ (``EXPERIMENT_STATE_CREATED``) where the Thrift enum uses the bare name
+ (``CREATED``). Members absent from Thrift — notably the zero ``*_UNKNOWN``
+ sentinel — map to None (the serializer renders these as nullable ints).
+ """
+ name = pb.DESCRIPTOR.fields_by_name[field].enum_type.values_by_number[
+ getattr(pb, field)].name
+ if name.startswith(proto_prefix):
+ name = name[len(proto_prefix):]
+ return getattr(thrift_enum, name, None)
+
+
def _thrift_enum_mapped(pb, field, proto_name_to_thrift):
"""Bridge a proto enum field to a Thrift value via an EXPLICIT name map.
@@ -510,3 +536,222 @@ def group_resource_profile(pb):
updatedTime=pb.updated_time or None,
defaultCredentialStoreToken=pb.default_credential_store_token or None,
)
+
+
+# --- Experiment tree -------------------------------------------------------
+#
+# getExperiment returns the full ExperimentModel including the processes tree
+# (process -> tasks -> jobs, each with its status list). The status/type enums
+# render as raw integers and are bridged by name with the proto prefix stripped
+# (proto EXPERIMENT_STATE_CREATED -> Thrift CREATED). Status timeOfStateChange
+# stays an int (ExperimentStatusSerializer/ProcessStatusSerializer use a
+# non-nullable UTC field; the nested auto-generated ones render the int);
+# model creation/update times map proto-zero -> None (nullable).
+
+
+def _error_model(pb):
+ """gRPC ``ErrorModel`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ errorId=pb.error_id,
+ creationTime=pb.creation_time or None,
+ actualErrorMessage=pb.actual_error_message,
+ userFriendlyMessage=pb.user_friendly_message,
+ transientOrPersistent=pb.transient_or_persistent,
+ rootCauseErrorIdList=list(pb.root_cause_error_id_list),
+ )
+
+
+def _computational_resource_scheduling(pb):
+ """gRPC ``ComputationalResourceSchedulingModel`` -> auto-generated
shape."""
+ return SimpleNamespace(
+ resourceHostId=pb.resource_host_id,
+ totalCPUCount=pb.total_cpu_count,
+ nodeCount=pb.node_count,
+ numberOfThreads=pb.number_of_threads,
+ queueName=pb.queue_name,
+ wallTimeLimit=pb.wall_time_limit,
+ totalPhysicalMemory=pb.total_physical_memory,
+ chessisNumber=pb.chessis_number,
+ staticWorkingDir=pb.static_working_dir,
+ overrideLoginUserName=pb.override_login_user_name,
+ overrideScratchLocation=pb.override_scratch_location,
+ overrideAllocationProjectNumber=pb.override_allocation_project_number,
+ mGroupCount=pb.m_group_count,
+ )
+
+
+def _user_configuration_data(pb):
+ """gRPC ``UserConfigurationDataModel`` -> auto-generated serializer
shape."""
+ return SimpleNamespace(
+ airavataAutoSchedule=pb.airavata_auto_schedule,
+ overrideManualScheduledParams=pb.override_manual_scheduled_params,
+ shareExperimentPublicly=pb.share_experiment_publicly,
+ computationalResourceScheduling=(
+
_computational_resource_scheduling(pb.computational_resource_scheduling)
+ if pb.HasField('computational_resource_scheduling') else None),
+ throttleResources=pb.throttle_resources,
+ userDN=pb.user_dn,
+ generateCert=pb.generate_cert,
+ inputStorageResourceId=pb.input_storage_resource_id,
+ outputStorageResourceId=pb.output_storage_resource_id,
+ experimentDataDir=pb.experiment_data_dir,
+ useUserCRPref=pb.use_user_cr_pref,
+ groupResourceProfileId=pb.group_resource_profile_id,
+ autoScheduledCompResourceSchedulingList=[
+ _computational_resource_scheduling(s)
+ for s in pb.auto_scheduled_comp_resource_scheduling_list],
+ )
+
+
+def _experiment_status(pb):
+ """gRPC ``ExperimentStatus`` -> ``ExperimentStatusSerializer`` shape."""
+ return SimpleNamespace(
+ state=_thrift_enum_prefixed(
+ pb, 'state', _ThriftExperimentState, 'EXPERIMENT_STATE_'),
+ timeOfStateChange=pb.time_of_state_change,
+ reason=pb.reason,
+ statusId=pb.status_id,
+ )
+
+
+def _process_status(pb):
+ """gRPC ``ProcessStatus`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ state=_thrift_enum_prefixed(
+ pb, 'state', _ThriftProcessState, 'PROCESS_STATE_'),
+ timeOfStateChange=pb.time_of_state_change,
+ reason=pb.reason,
+ statusId=pb.status_id,
+ processId=pb.process_id,
+ )
+
+
+def _task_status(pb):
+ """gRPC ``TaskStatus`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ state=_thrift_enum_prefixed(
+ pb, 'state', _ThriftTaskState, 'TASK_STATE_'),
+ timeOfStateChange=pb.time_of_state_change,
+ reason=pb.reason,
+ statusId=pb.status_id,
+ )
+
+
+def _job_status(pb):
+ """gRPC ``JobStatus`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ jobState=_thrift_enum_prefixed(
+ pb, 'job_state', _ThriftJobState, 'JOB_STATE_'),
+ timeOfStateChange=pb.time_of_state_change,
+ reason=pb.reason,
+ statusId=pb.status_id,
+ )
+
+
+def job_model(pb):
+ """gRPC ``JobModel`` -> ``JobSerializer`` shape."""
+ return SimpleNamespace(
+ jobId=pb.job_id,
+ taskId=pb.task_id,
+ processId=pb.process_id,
+ jobDescription=pb.job_description,
+ creationTime=pb.creation_time or None,
+ jobStatuses=[_job_status(s) for s in pb.job_statuses],
+ computeResourceConsumed=pb.compute_resource_consumed,
+ jobName=pb.job_name,
+ workingDir=pb.working_dir,
+ stdOut=pb.std_out,
+ stdErr=pb.std_err,
+ exitCode=pb.exit_code,
+ )
+
+
+def _task_model(pb):
+ """gRPC ``TaskModel`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ taskId=pb.task_id,
+ taskType=_thrift_enum_prefixed(
+ pb, 'task_type', _ThriftTaskTypes, 'TASK_TYPES_'),
+ parentProcessId=pb.parent_process_id,
+ creationTime=pb.creation_time or None,
+ lastUpdateTime=pb.last_update_time or None,
+ taskStatuses=[_task_status(s) for s in pb.task_statuses],
+ taskDetail=pb.task_detail,
+ subTaskModel=pb.sub_task_model,
+ taskErrors=[_error_model(e) for e in pb.task_errors],
+ jobs=[job_model(j) for j in pb.jobs],
+ maxRetry=pb.max_retry,
+ currentRetry=pb.current_retry,
+ )
+
+
+def _process_model(pb):
+ """gRPC ``ProcessModel`` -> auto-generated serializer shape."""
+ return SimpleNamespace(
+ processId=pb.process_id,
+ experimentId=pb.experiment_id,
+ creationTime=pb.creation_time or None,
+ lastUpdateTime=pb.last_update_time or None,
+ processStatuses=[_process_status(s) for s in pb.process_statuses],
+ processDetail=pb.process_detail,
+ applicationInterfaceId=pb.application_interface_id,
+ applicationDeploymentId=pb.application_deployment_id,
+ computeResourceId=pb.compute_resource_id,
+ processInputs=[_input_data_object(i) for i in pb.process_inputs],
+ processOutputs=[_output_data_object(o) for o in pb.process_outputs],
+ processResourceSchedule=(
+ _computational_resource_scheduling(pb.process_resource_schedule)
+ if pb.HasField('process_resource_schedule') else None),
+ tasks=[_task_model(t) for t in pb.tasks],
+ taskDag=pb.task_dag,
+ processErrors=[_error_model(e) for e in pb.process_errors],
+ gatewayExecutionId=pb.gateway_execution_id,
+ enableEmailNotification=pb.enable_email_notification,
+ emailAddresses=list(pb.email_addresses),
+ inputStorageResourceId=pb.input_storage_resource_id,
+ outputStorageResourceId=pb.output_storage_resource_id,
+ userDn=pb.user_dn,
+ generateCert=pb.generate_cert,
+ experimentDataDir=pb.experiment_data_dir,
+ userName=pb.user_name,
+ useUserCRPref=pb.use_user_cr_pref,
+ groupResourceProfileId=pb.group_resource_profile_id,
+ # the legacy workflow-engine subsystem is not adapted (rarely
populated);
+ # an empty list matches the Thrift default for non-workflow processes.
+ processWorkflows=[],
+ )
+
+
+def experiment(pb):
+ """gRPC ``ExperimentModel`` -> ``ExperimentSerializer`` shape.
+
+ The deepest read model: recursively adapts the user configuration, the
+ experiment input/output and status lists, the errors, and the full
+ processes -> tasks -> jobs tree.
+ """
+ return SimpleNamespace(
+ experimentId=pb.experiment_id,
+ projectId=pb.project_id,
+ gatewayId=pb.gateway_id,
+ experimentType=_thrift_enum_prefixed(
+ pb, 'experiment_type', _ThriftExperimentType, 'EXPERIMENT_TYPE_'),
+ userName=pb.user_name,
+ experimentName=pb.experiment_name,
+ creationTime=pb.creation_time or None,
+ description=pb.description,
+ executionId=pb.execution_id,
+ gatewayExecutionId=pb.gateway_execution_id,
+ gatewayInstanceId=pb.gateway_instance_id,
+ enableEmailNotification=pb.enable_email_notification,
+ emailAddresses=list(pb.email_addresses),
+ userConfigurationData=(
+ _user_configuration_data(pb.user_configuration_data)
+ if pb.HasField('user_configuration_data') else None),
+ experimentInputs=[_input_data_object(i) for i in pb.experiment_inputs],
+ experimentOutputs=[_output_data_object(o) for o in
pb.experiment_outputs],
+ experimentStatus=[_experiment_status(s) for s in pb.experiment_status],
+ errors=[_error_model(e) for e in pb.errors],
+ processes=[_process_model(p) for p in pb.processes],
+ # legacy workflow-engine subsystem not adapted (rarely populated).
+ workflow=None,
+ )
diff --git a/airavata-django-portal/django_airavata/apps/api/serializers.py
b/airavata-django-portal/django_airavata/apps/api/serializers.py
index 55179966a..c64471289 100644
--- a/airavata-django-portal/django_airavata/apps/api/serializers.py
+++ b/airavata-django-portal/django_airavata/apps/api/serializers.py
@@ -576,10 +576,7 @@ class ExperimentSerializer(
userHasWriteAccess = serializers.SerializerMethodField()
def get_userHasWriteAccess(self, experiment):
- request = self.context['request']
- return request.airavata_client.userHasAccess(
- request.authz_token, experiment.experimentId,
- ResourcePermissionType.WRITE)
+ return user_has_access(self.context['request'],
experiment.experimentId)
def to_representation(self, experiment):
result = super().to_representation(experiment)
diff --git a/airavata-django-portal/django_airavata/apps/api/views.py
b/airavata-django-portal/django_airavata/apps/api/views.py
index 13c365c08..cee66fdc4 100644
--- a/airavata-django-portal/django_airavata/apps/api/views.py
+++ b/airavata-django-portal/django_airavata/apps/api/views.py
@@ -197,8 +197,11 @@ class ProjectViewSet(APIBackedViewSet):
@action(detail=True)
def experiments(self, request, project_id=None):
- experiments = request.airavata_client.getExperimentsInProject(
- self.authz_token, project_id, -1, 0)
+ experiments = [
+ grpc_adapters.experiment(e)
+ for e in request.airavata.research.get_experiments_in_project(
+ project_id, -1, 0)
+ ]
serializer = serializers.ExperimentSerializer(
experiments, many=True, context={'request': request})
return Response(serializer.data)
@@ -217,8 +220,8 @@ class ExperimentViewSet(mixins.CreateModelMixin,
lookup_field = 'experiment_id'
def get_instance(self, lookup_value):
- return self.request.airavata_client.getExperiment(
- self.authz_token, lookup_value)
+ return grpc_adapters.experiment(
+ self.request.airavata.research.get_experiment(lookup_value))
def perform_create(self, serializer):
experiment = serializer.save(
@@ -260,8 +263,10 @@ class ExperimentViewSet(mixins.CreateModelMixin,
@action(methods=['get'], detail=True)
def jobs(self, request, experiment_id=None):
- jobs = request.airavata_client.getJobDetails(
- self.authz_token, experiment_id)
+ jobs = [
+ grpc_adapters.job_model(j)
+ for j in request.airavata.research.get_job_details(experiment_id)
+ ]
serializer = serializers.JobSerializer(
jobs, many=True, context={'request': request})
return Response(serializer.data)
