Fix DPSJob.wait_for_completion() race condition with initial delay

[sujen1412]

Summary

Fixes the race condition in DPSJob.wait_for_completion() where the method would return immediately with an incorrect "Deleted" status if called right after job submission, before the job appears in the OpenSearch database.

Problem

• When jobs are submitted via DPS API, they take up to 5 seconds to appear in the OpenSearch database
• If wait_for_completion() is called immediately after submission, the API returns an unexpected status before the job appears in the database
• The method would return immediately with this incorrect status instead of retrying

Solution

Implemented an initial delay mechanism:

• Added initial_delay_seconds parameter (default 5 seconds) to wait_for_completion()
• The method now waits 5 seconds before the first status check, allowing the job to appear in the database
• Improved status check logic to properly handle all non-terminal states by retrying

Changes

• ✅ Added time module import
• ✅ Added initial_delay_seconds parameter with default value of 5 seconds
• ✅ Added initial sleep before first status check
• ✅ Changed status check logic from whitelist (accepted/running) to blacklist (terminal states)
• ✅ Updated docstring with parameter documentation

Testing

Users can now reliably call wait_for_completion() immediately after job submission:

job = maap.submit_job(...)
job.wait_for_completion()  # Waits 5 seconds then polls correctly
print(f"Status: {job.status}")  # Returns "Succeeded" (or other terminal state)

To disable the initial delay for existing jobs (where status is already available):

job.wait_for_completion(initial_delay_seconds=0)

Related Issues

Fixes #126

https://claude.ai/code/session_01KEDGhwSuoY9N5inWR9bk5u

---

Generated by Claude Code