Merge pull request #21 from NREL/perf-improvements

Perf improvements

Author: daniel-thom

api/openapi.yaml

@@ -880,6 +880,15 @@ paths:
           default: false
           type: boolean
         style: form
+      - explode: true
+        in: query
+        name: include_relationships
+        required: false
+        schema:
+          default: false
+          type: boolean
+        style: form
+        description: Include job relationships (blocked_by_job_ids, input_file_ids, output_file_ids, input_user_data_ids, output_user_data_ids). Default is false for performance.
       responses:
         "200":
           content:

docs/src/explanation/workflow-actions.md

@@ -238,7 +238,6 @@ Dynamically allocate compute resources from a Slurm scheduler.
 - `scheduler_type` (required) - Must be "slurm"
 - `num_allocations` (required) - Number of Slurm allocation requests to submit
 - `start_one_worker_per_node` (optional) - Start one job runner per node (default: false)
-- `start_server_on_head_node` (optional) - Start torc-server on head node (default: false)
 - `max_parallel_jobs` (optional) - Maximum concurrent jobs per runner
 
 **Use cases**:

docs/src/how-to/slurm.md

@@ -10,10 +10,7 @@
   be minimal. However, if you allocate hundreds of nodes with many thousands of
   short jobs, the Torc server process may exceed allowed resource limits. Check
   with your operations team if you have doubts.
-- **Slurm head node**: Create the workflow while running the Torc server on the
-  login node. Start the workflow with `start_server_on_head_node` set to true.
-  Shut down the server. Torc will configure the job to start the server upon
-  being granted the allocation.
+
 
 ## Basic Slurm Configuration
 

docs/src/how-to/torc-server-args.md

@@ -20,15 +20,16 @@ pip install -e .
 Let's suppose that your code is in a module called `simulation.py` and looks something like this:
 
 ```python
-def run(input_params: dict) -> dict:
+def run(job_name: str, input_params: dict) -> dict:
     """Runs one simulation on a set of input parameters.
 
     Returns
     -------
+    job_name: str
+        Name of the job.
     dict
         Result of the simulation.
     """
-    job_name = input_params["job_name"]
     return {
         "inputs": input_params,
         "result": 5,