Add timeout for each workflow steps

[jackiehanyang]

12/26 revision:

• Change the system default NODE_TIMEOUT_DEFAULT_VALUE to 10s

• Removed all 10s timeout values in the workflow-step.json file, only leaving the timeout field for those workflow steps that require an overwrite timeout value other than 10s.
  For example, register_local_model has default timeout value as 60s, and deploy_model has default timeout value as 15s

• Do try catch at where we read workflow-step.json file instead of passing it up.

• Add a new unit test case to cover all possible ways of how we parse timeout value

---

Description

This pr does two things:

• Having a specific timeout for all the workflow steps in workflow-steps.json
• rename indices:
  .plugins-ai-global-context -> .plugins-flow-framework-templates
.plugins-workflow-state -> .plugins-flow-framework-state

Issues Resolved

#249

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[dbwiddis]

Nice code, and I learned something! Two general comments:

• Let's use TimeValue rather than String as early as we can in parsing
• Let's handle exceptions as early as we can in parsing

[jackiehanyang]

Nice code, and I learned something! Two general comments:

• Let's use TimeValue rather than String as early as we can in parsing
• Let's handle exceptions as early as we can in parsing

Thanks for the quick review! Yeah, I immediately regretted the decision not to handle exceptions in parsing right after I sent out this PR and closed my laptop. But since I had already closed my laptop, I wanted to wait until Tuesday to address this. I will send out a new revision soon to address it. :)

[jackiehanyang]

Nice code, and I learned something! Two general comments:

• Let's use TimeValue rather than String as early as we can in parsing
• Let's handle exceptions as early as we can in parsing

I prefer to leave timeout as a String value, and leave it as timeout = parser.text() for the following reasons:

1. TimeValue.parseTimeValue(...) throws NPE if parser.text() is null - https://github.com/opensearch-project/OpenSearch/blob/71d7e4daf7dc05c9598d447e4424072e76de79d1/libs/common/src/main/java/org/opensearch/common/unit/TimeValue.java#L374-L375. But a String type allows it to be null.
2. The reason we allow it to be null is that we want to handle parsing errors when we parse it, rather than in the actual parser. If we handle it in the actual parser, we would need to catch every switch case. However, if we handle it when we call the parser, we can catch all parsing errors at once. For example, we don't throw exception here, instead we do try and catch here - https://github.com/opensearch-project/flow-framework/blob/main/src/main/java/org/opensearch/flowframework/workflow/WorkflowProcessSorter.java#L143-L144
3. Since the TIMEOUT in the workflow-step.json file is part of our own configuration, it shouldn't be null. However, just in case it is, we assign a default value in the code to prevent failing the entire workflow. As it's not a user-provided field, it makes no sense for the workflow to fail just because it's null. Therefore, we catch it and assign a default value in every step where possible.

[dbwiddis]

LGTM with a nit and few outstanding comments from prior review.