dkist-processing-common 11.8.1rc3__tar.gz → 11.9.0__tar.gz

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/CHANGELOG.rst

@@ -1,3 +1,24 @@
+v11.9.0 (2025-12-03)
+====================
+
+Misc
+----
+
+- Upgrade to the globus 4.x SDK. (`#274 <https://bitbucket.org/dkistdc/dkist-processing-common/pull-requests/274>`__)
+- Use a pool of direction aware globus accounts for transfers to mitigate contention between simultaneous inbound (TransferL0Data) and/or outbound (TransferL1Data) transfers. (`#274 <https://bitbucket.org/dkistdc/dkist-processing-common/pull-requests/274>`__)
+- Integrate dkist-processing-core 6.0.1 which adds additional attributes to metrics and tracing to facilitate discoverability and analysis. (`#275 <https://bitbucket.org/dkistdc/dkist-processing-common/pull-requests/275>`__)
+
+
+v11.8.1 (2025-12-02)
+====================
+
+Misc
+----
+
+- Bump minimum version of `pydantic` to 2.7.2 to avoid install failures for python versions >= 3.12. (`#271 <https://bitbucket.org/dkistdc/dkist-processing-common/pull-requests/271>`__)
+- Move `solar-wavelength-calibration` dep to 2.0.0 and make use of new helper properties in that release. (`#271 <https://bitbucket.org/dkistdc/dkist-processing-common/pull-requests/271>`__)
+
+
 v11.8.0 (2025-11-03)
 ====================
 

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dkist-processing-common
-Version: 11.8.1rc3
+Version: 11.9.0
 Summary: Common task classes used by the DKIST science data processing pipelines
 Author-email: NSO / AURA <dkistdc@nso.edu>
 License: BSD-3-Clause
@@ -17,12 +17,12 @@ Requires-Dist: asdf<4.0.0,>=3.5.0
 Requires-Dist: astropy>=7.0.0
 Requires-Dist: dkist-fits-specifications<5.0,>=4.0.0
 Requires-Dist: dkist-header-validator<6.0,>=5.0.0
-Requires-Dist: dkist-processing-core==6.0.0
+Requires-Dist: dkist-processing-core==6.0.1
 Requires-Dist: dkist-processing-pac<4.0,>=3.1
 Requires-Dist: dkist-service-configuration<5.0,>=4.1.7
 Requires-Dist: dkist-spectral-lines<4.0,>=3.0.0
-Requires-Dist: solar-wavelength-calibration<3.0,>=2.0.0rc4
-Requires-Dist: globus-sdk<4.0.0,>=3.12.0
+Requires-Dist: solar-wavelength-calibration<3.0,>=2.0.0
+Requires-Dist: globus-sdk<5.0.0,>=4.0.0
 Requires-Dist: gqlclient[pydantic]==1.2.3
 Requires-Dist: sqids==0.5.1
 Requires-Dist: matplotlib>=3.4
@@ -118,6 +118,10 @@ Environment Variables
      - annotation=str required=False default='unknown-service-version' alias_priority=2 validation_alias='DKIST_SERVICE_VERSION' description='Service version for OpenTelemetry'
    * - NOMAD_ALLOC_ID
      - annotation=str required=False default='unknown-allocation-id' alias_priority=2 validation_alias='NOMAD_ALLOC_ID' description='Nomad allocation ID for OpenTelemetry'
+   * - NOMAD_ALLOC_NAME
+     - annotation=str required=False default='unknown-allocation-name' alias='NOMAD_ALLOC_NAME' alias_priority=2 description='Allocation name for the deployed container the task is running on.'
+   * - NOMAD_GROUP_NAME
+     - annotation=str required=False default='unknown-allocation-group' alias='NOMAD_GROUP_NAME' alias_priority=2 description='Allocation group for the deployed container the task is running on'
    * - OTEL_EXPORTER_OTLP_TRACES_INSECURE
      - annotation=bool required=False default=True description='Use insecure connection for OTLP traces'
    * - OTEL_EXPORTER_OTLP_METRICS_INSECURE
@@ -158,12 +162,12 @@ Environment Variables
      - annotation=Union[dict, NoneType] required=False default=None description='S3 upload configuration for the object store.'
    * - S3_DOWNLOAD_CONFIG
      - annotation=Union[dict, NoneType] required=False default=None description='S3 download configuration for the object store.'
-   * - GLOBUS_TRANSPORT_PARAMS
-     - annotation=dict required=False default_factory=dict description='Globus transfer parameters.'
-   * - GLOBUS_CLIENT_ID
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client ID for inbound/outbound transfers.'
-   * - GLOBUS_CLIENT_SECRET
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client secret for inbound/outbound transfers.'
+   * - GLOBUS_MAX_RETRIES
+     - annotation=int required=False default=5 description='Max retries for transient errors on calls to the globus api.'
+   * - GLOBUS_INBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for inbound transfers.' examples=[[{'client_id': 'id1', 'client_secret': 'secret1'}, {'client_id': 'id2', 'client_secret': 'secret2'}]]
+   * - GLOBUS_OUTBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for outbound transfers.' examples=[[{'client_id': 'id3', 'client_secret': 'secret3'}, {'client_id': 'id4', 'client_secret': 'secret4'}]]
    * - OBJECT_STORE_ENDPOINT
      - annotation=Union[str, NoneType] required=False default=None description='Object store Globus Endpoint ID.'
    * - SCRATCH_ENDPOINT

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/README.rst

@@ -53,6 +53,10 @@ Environment Variables
      - annotation=str required=False default='unknown-service-version' alias_priority=2 validation_alias='DKIST_SERVICE_VERSION' description='Service version for OpenTelemetry'
    * - NOMAD_ALLOC_ID
      - annotation=str required=False default='unknown-allocation-id' alias_priority=2 validation_alias='NOMAD_ALLOC_ID' description='Nomad allocation ID for OpenTelemetry'
+   * - NOMAD_ALLOC_NAME
+     - annotation=str required=False default='unknown-allocation-name' alias='NOMAD_ALLOC_NAME' alias_priority=2 description='Allocation name for the deployed container the task is running on.'
+   * - NOMAD_GROUP_NAME
+     - annotation=str required=False default='unknown-allocation-group' alias='NOMAD_GROUP_NAME' alias_priority=2 description='Allocation group for the deployed container the task is running on'
    * - OTEL_EXPORTER_OTLP_TRACES_INSECURE
      - annotation=bool required=False default=True description='Use insecure connection for OTLP traces'
    * - OTEL_EXPORTER_OTLP_METRICS_INSECURE
@@ -93,12 +97,12 @@ Environment Variables
      - annotation=Union[dict, NoneType] required=False default=None description='S3 upload configuration for the object store.'
    * - S3_DOWNLOAD_CONFIG
      - annotation=Union[dict, NoneType] required=False default=None description='S3 download configuration for the object store.'
-   * - GLOBUS_TRANSPORT_PARAMS
-     - annotation=dict required=False default_factory=dict description='Globus transfer parameters.'
-   * - GLOBUS_CLIENT_ID
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client ID for inbound/outbound transfers.'
-   * - GLOBUS_CLIENT_SECRET
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client secret for inbound/outbound transfers.'
+   * - GLOBUS_MAX_RETRIES
+     - annotation=int required=False default=5 description='Max retries for transient errors on calls to the globus api.'
+   * - GLOBUS_INBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for inbound transfers.' examples=[[{'client_id': 'id1', 'client_secret': 'secret1'}, {'client_id': 'id2', 'client_secret': 'secret2'}]]
+   * - GLOBUS_OUTBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for outbound transfers.' examples=[[{'client_id': 'id3', 'client_secret': 'secret3'}, {'client_id': 'id4', 'client_secret': 'secret4'}]]
    * - OBJECT_STORE_ENDPOINT
      - annotation=Union[str, NoneType] required=False default=None description='Object store Globus Endpoint ID.'
    * - SCRATCH_ENDPOINT

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/dkist_processing_common/config.py

@@ -3,12 +3,20 @@
 from dkist_processing_core.config import DKISTProcessingCoreConfiguration
 from dkist_service_configuration.settings import DEFAULT_MESH_SERVICE
 from dkist_service_configuration.settings import MeshService
+from pydantic import BaseModel
 from pydantic import Field
 from talus import ConnectionRetryerFactory
 from talus import ConsumerConnectionParameterFactory
 from talus import ProducerConnectionParameterFactory
 
 
+class GlobusClientCredential(BaseModel):
+    """Globus client credential."""
+
+    client_id: str = Field(..., description="Globus client ID for transfers.")
+    client_secret: str = Field(..., description="Globus client secret for transfers.")
+
+
 class DKISTProcessingCommonConfiguration(DKISTProcessingCoreConfiguration):
     """Common configurations."""
 
@@ -40,14 +48,28 @@ class DKISTProcessingCommonConfiguration(DKISTProcessingCoreConfiguration):
         default=None, description="S3 download configuration for the object store."
     )
     # globus
-    globus_transport_params: dict = Field(
-        default_factory=dict, description="Globus transfer parameters."
+    globus_max_retries: int = Field(
+        default=5, description="Max retries for transient errors on calls to the globus api."
     )
-    globus_client_id: str | None = Field(
-        default=None, description="Globus client ID for inbound/outbound transfers."
+    globus_inbound_client_credentials: list[GlobusClientCredential] = Field(
+        default_factory=list,
+        description="Globus client credentials for inbound transfers.",
+        examples=[
+            [
+                {"client_id": "id1", "client_secret": "secret1"},
+                {"client_id": "id2", "client_secret": "secret2"},
+            ],
+        ],
     )
-    globus_client_secret: str | None = Field(
-        default=None, description="Globus client secret for inbound/outbound transfers."
+    globus_outbound_client_credentials: list[GlobusClientCredential] = Field(
+        default_factory=list,
+        description="Globus client credentials for outbound transfers.",
+        examples=[
+            [
+                {"client_id": "id3", "client_secret": "secret3"},
+                {"client_id": "id4", "client_secret": "secret4"},
+            ],
+        ],
     )
     object_store_endpoint: str | None = Field(
         default=None, description="Object store Globus Endpoint ID."

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/dkist_processing_common/tasks/mixin/globus.py

@@ -9,6 +9,8 @@ from globus_sdk import ConfidentialAppAuthClient
 from globus_sdk import GlobusError
 from globus_sdk import TransferClient
 from globus_sdk import TransferData
+from globus_sdk.scopes import TransferScopes
+from globus_sdk.transport import RetryConfig
 
 from dkist_processing_common.config import common_configurations
 
@@ -31,27 +33,32 @@ class GlobusTransferItem:
 class GlobusMixin:
     """Mixin to add methods to a Task to support globus transfers."""
 
-    @property
-    def globus_transfer_client(self) -> TransferClient:
-        """Get the globus transfer client, creating it if it doesn't exist."""
-        if getattr(self, "_globus_transfer_client", False):
-            return self._globus_transfer_client
+    def globus_transfer_client_factory(self, transfer_data: TransferData) -> TransferClient:
+        """Create a globus transfer client based on the direction of transfer and round-robin the available application credentials."""
+        if (
+            transfer_data["source_endpoint"] == common_configurations.object_store_endpoint
+        ):  # inbound
+            client_credentials = common_configurations.globus_inbound_client_credentials
+        else:  # outbound
+            client_credentials = common_configurations.globus_outbound_client_credentials
+
+        # Round-robin the client credentials based on the recipe run id
+        index = self.recipe_run_id % len(client_credentials)
+        selected_credential = client_credentials[index]
+
         confidential_client = ConfidentialAppAuthClient(
-            client_id=common_configurations.globus_client_id,
-            client_secret=common_configurations.globus_client_secret,
-            transport_params=common_configurations.globus_transport_params,
+            client_id=selected_credential.client_id,
+            client_secret=selected_credential.client_secret,
         )
-        authorizer = ClientCredentialsAuthorizer(
-            confidential_client, scopes="urn:globus:auth:scope:transfer.api.globus.org:all"
-        )
-        self._globus_transfer_client = TransferClient(authorizer=authorizer)
-        return self._globus_transfer_client
+        authorizer = ClientCredentialsAuthorizer(confidential_client, scopes=TransferScopes)
+        retry_config = RetryConfig(max_retries=common_configurations.globus_max_retries)
+
+        return TransferClient(authorizer=authorizer, retry_config=retry_config)
 
     def globus_transfer_scratch_to_object_store(
         self,
         transfer_items: list[GlobusTransferItem],
         label: str = None,
-        sync_level: str = None,
         verify_checksum: bool = True,
     ) -> None:
         """Transfer data from scratch to the object store."""
@@ -60,7 +67,6 @@ class GlobusMixin:
             destination_endpoint=common_configurations.object_store_endpoint,
             transfer_items=transfer_items,
             label=label,
-            sync_level=sync_level,
             verify_checksum=verify_checksum,
         )
 
@@ -68,7 +74,6 @@ class GlobusMixin:
         self,
         transfer_items: list[GlobusTransferItem],
         label: str = None,
-        sync_level: str = None,
         verify_checksum: bool = True,
     ) -> None:
         """Transfer data from the object store to scratch."""
@@ -77,7 +82,6 @@ class GlobusMixin:
             destination_endpoint=common_configurations.scratch_endpoint,
             transfer_items=transfer_items,
             label=label,
-            sync_level=sync_level,
             verify_checksum=verify_checksum,
         )
 
@@ -87,7 +91,6 @@ class GlobusMixin:
         destination_endpoint: str,
         transfer_items: list[GlobusTransferItem],
         label: str = None,
-        sync_level: str = None,
         verify_checksum: bool = True,
     ) -> TransferData:
         """Format a globus TransferData instance."""
@@ -95,7 +98,6 @@ class GlobusMixin:
             source_endpoint=source_endpoint,
             destination_endpoint=destination_endpoint,
             label=label,
-            sync_level=sync_level,
             verify_checksum=verify_checksum,
         )
         for item in transfer_items:
@@ -112,7 +114,6 @@ class GlobusMixin:
         destination_endpoint: str,
         transfer_items: list[GlobusTransferItem],
         label: str = None,
-        sync_level: str = None,
         verify_checksum: bool = True,
     ) -> None:
         """Perform a transfer of data using globus."""
@@ -121,7 +122,6 @@ class GlobusMixin:
             destination_endpoint=destination_endpoint,
             transfer_items=transfer_items,
             label=label,
-            sync_level=sync_level,
             verify_checksum=verify_checksum,
         )
         self._blocking_globus_transfer(transfer_data=transfer_data)
@@ -131,24 +131,21 @@ class GlobusMixin:
         source_endpoint: str,
         destination_endpoint: str,
         label: str = None,
-        sync_level: str = None,
         verify_checksum: bool = True,
     ) -> TransferData:
         label = label or "Data Processing Transfer"
         return TransferData(
-            transfer_client=self.globus_transfer_client,
             source_endpoint=source_endpoint,
             destination_endpoint=destination_endpoint,
             label=label,
-            sync_level=sync_level,
             verify_checksum=verify_checksum,
         )
 
     def _blocking_globus_transfer(self, transfer_data: TransferData) -> None:
-        tc = self.globus_transfer_client
-        logger.info(f"Starting globus transfer: label={transfer_data.get('label')}")
+        tc = self.globus_transfer_client_factory(transfer_data=transfer_data)
         transfer_result = tc.submit_transfer(transfer_data)
         task_id = transfer_result["task_id"]
+        logger.info(f"Starting globus transfer: label={transfer_data.get('label')}, {task_id=}, ")
         polling_interval = 60
         while not tc.task_wait(
             task_id=task_id, timeout=polling_interval, polling_interval=polling_interval

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/dkist_processing_common.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dkist-processing-common
-Version: 11.8.1rc3
+Version: 11.9.0
 Summary: Common task classes used by the DKIST science data processing pipelines
 Author-email: NSO / AURA <dkistdc@nso.edu>
 License: BSD-3-Clause
@@ -17,12 +17,12 @@ Requires-Dist: asdf<4.0.0,>=3.5.0
 Requires-Dist: astropy>=7.0.0
 Requires-Dist: dkist-fits-specifications<5.0,>=4.0.0
 Requires-Dist: dkist-header-validator<6.0,>=5.0.0
-Requires-Dist: dkist-processing-core==6.0.0
+Requires-Dist: dkist-processing-core==6.0.1
 Requires-Dist: dkist-processing-pac<4.0,>=3.1
 Requires-Dist: dkist-service-configuration<5.0,>=4.1.7
 Requires-Dist: dkist-spectral-lines<4.0,>=3.0.0
-Requires-Dist: solar-wavelength-calibration<3.0,>=2.0.0rc4
-Requires-Dist: globus-sdk<4.0.0,>=3.12.0
+Requires-Dist: solar-wavelength-calibration<3.0,>=2.0.0
+Requires-Dist: globus-sdk<5.0.0,>=4.0.0
 Requires-Dist: gqlclient[pydantic]==1.2.3
 Requires-Dist: sqids==0.5.1
 Requires-Dist: matplotlib>=3.4
@@ -118,6 +118,10 @@ Environment Variables
      - annotation=str required=False default='unknown-service-version' alias_priority=2 validation_alias='DKIST_SERVICE_VERSION' description='Service version for OpenTelemetry'
    * - NOMAD_ALLOC_ID
      - annotation=str required=False default='unknown-allocation-id' alias_priority=2 validation_alias='NOMAD_ALLOC_ID' description='Nomad allocation ID for OpenTelemetry'
+   * - NOMAD_ALLOC_NAME
+     - annotation=str required=False default='unknown-allocation-name' alias='NOMAD_ALLOC_NAME' alias_priority=2 description='Allocation name for the deployed container the task is running on.'
+   * - NOMAD_GROUP_NAME
+     - annotation=str required=False default='unknown-allocation-group' alias='NOMAD_GROUP_NAME' alias_priority=2 description='Allocation group for the deployed container the task is running on'
    * - OTEL_EXPORTER_OTLP_TRACES_INSECURE
      - annotation=bool required=False default=True description='Use insecure connection for OTLP traces'
    * - OTEL_EXPORTER_OTLP_METRICS_INSECURE
@@ -158,12 +162,12 @@ Environment Variables
      - annotation=Union[dict, NoneType] required=False default=None description='S3 upload configuration for the object store.'
    * - S3_DOWNLOAD_CONFIG
      - annotation=Union[dict, NoneType] required=False default=None description='S3 download configuration for the object store.'
-   * - GLOBUS_TRANSPORT_PARAMS
-     - annotation=dict required=False default_factory=dict description='Globus transfer parameters.'
-   * - GLOBUS_CLIENT_ID
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client ID for inbound/outbound transfers.'
-   * - GLOBUS_CLIENT_SECRET
-     - annotation=Union[str, NoneType] required=False default=None description='Globus client secret for inbound/outbound transfers.'
+   * - GLOBUS_MAX_RETRIES
+     - annotation=int required=False default=5 description='Max retries for transient errors on calls to the globus api.'
+   * - GLOBUS_INBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for inbound transfers.' examples=[[{'client_id': 'id1', 'client_secret': 'secret1'}, {'client_id': 'id2', 'client_secret': 'secret2'}]]
+   * - GLOBUS_OUTBOUND_CLIENT_CREDENTIALS
+     - annotation=list[GlobusClientCredential] required=False default_factory=list description='Globus client credentials for outbound transfers.' examples=[[{'client_id': 'id3', 'client_secret': 'secret3'}, {'client_id': 'id4', 'client_secret': 'secret4'}]]
    * - OBJECT_STORE_ENDPOINT
      - annotation=Union[str, NoneType] required=False default=None description='Object store Globus Endpoint ID.'
    * - SCRATCH_ENDPOINT

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/dkist_processing_common.egg-info/SOURCES.txt

@@ -7,8 +7,6 @@ README.rst
 bitbucket-pipelines.yml
 pyproject.toml
 changelog/.gitempty
-changelog/271.misc.2.rst
-changelog/271.misc.rst
 dkist_processing_common/__init__.py
 dkist_processing_common/config.py
 dkist_processing_common/manual.py

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/dkist_processing_common.egg-info/requires.txt

@@ -2,12 +2,12 @@ asdf<4.0.0,>=3.5.0
 astropy>=7.0.0
 dkist-fits-specifications<5.0,>=4.0.0
 dkist-header-validator<6.0,>=5.0.0
-dkist-processing-core==6.0.0
+dkist-processing-core==6.0.1
 dkist-processing-pac<4.0,>=3.1
 dkist-service-configuration<5.0,>=4.1.7
 dkist-spectral-lines<4.0,>=3.0.0
-solar-wavelength-calibration<3.0,>=2.0.0rc4
-globus-sdk<4.0.0,>=3.12.0
+solar-wavelength-calibration<3.0,>=2.0.0
+globus-sdk<5.0.0,>=4.0.0
 gqlclient[pydantic]==1.2.3
 sqids==0.5.1
 matplotlib>=3.4

{dkist_processing_common-11.8.1rc3 → dkist_processing_common-11.9.0}/pyproject.toml

@@ -25,12 +25,12 @@ dependencies = [
   "astropy  >= 7.0.0",
   "dkist-fits-specifications >= 4.0.0, <5.0",
   "dkist-header-validator >= 5.0.0, <6.0",
-  "dkist-processing-core == 6.0.0",
+  "dkist-processing-core == 6.0.1",
   "dkist-processing-pac >= 3.1, <4.0",
   "dkist-service-configuration >= 4.1.7, <5.0",
   "dkist-spectral-lines >= 3.0.0, <4.0",
-  "solar-wavelength-calibration >= 2.0.0rc4, <3.0",
-  "globus-sdk >= 3.12.0, <4.0.0",
+  "solar-wavelength-calibration >= 2.0.0, <3.0",
+  "globus-sdk >= 4.0.0, <5.0.0",
   "gqlclient[pydantic] == 1.2.3",
   "sqids == 0.5.1",
   "matplotlib >= 3.4",

dkist_processing_common-11.8.1rc3/changelog/271.misc.2.rst

@@ -1 +0,0 @@
-Bump minimum version of `pydantic` to 2.7.2 to avoid install failures for python versions >= 3.12.

dkist_processing_common-11.8.1rc3/changelog/271.misc.rst

@@ -1 +0,0 @@
-Move `solar-wavelength-calibration` dep to 2.0.0 and make use of new helper properties in that release.