Cache results of allowed_workflows

[mbthornton-lbl]

This PR addresses a stale file-handle error, which in turn cause the Watcher daemon to exit:

2024-11-19 12:40:33,557 INFO: Restored 0 jobs from state
2024-11-19 12:40:33,561 ERROR: Error occurred during cycle: [Errno 116] Stale file handle: '/global/homes/n/nmdcda/nmdc_automation/dev/nmdc_automation/nmdc_automation/config/workflows/workflows.yaml'
Traceback (most recent call last):
  File "/global/homes/n/nmdcda/nmdc_automation/dev/nmdc_automation/nmdc_automation/workflow_automation/watch_nmdc.py", line 360, in watch
  File "/global/homes/n/nmdcda/nmdc_automation/dev/nmdc_automation/nmdc_automation/workflow_automation/watch_nmdc.py", line 318, in cycle
  File "/global/homes/n/nmdcda/nmdc_automation/dev/nmdc_automation/nmdc_automation/config/siteconfig.py", line 104, in allowed_workflows
OSError: [Errno 116] Stale file handle: '/global/homes/n/nmdcda/nmdc_automation/dev/nmdc_automation/nmdc_automation/config/workflows/workflows.yaml'

The fix here is to cache the results of reading in workflows.yaml in the allowed_workflows function