fix: Remove hasattr check for every invocation

[gavin-aguiar]

Description

This PR has three key optimizations in the dispatcher:

1. Eliminated Repeated Attribute Checks (hasattr):
   Change: Introduced a global flag _library_worker_has_cv that is set once when the worker library is loaded.
   Impact: get_current_invocation_id runs on every log message. Previously, it called hasattr(_library_worker, 'invocation_id_cv') thousands of times during startup. This is now a fast boolean check.

2. Lock-Free Fast Path for Invocation ID:
   Change: Reordered get_current_invocation_id to check the ContextVar before checking the thread-local registry.
   Impact: The thread registry requires acquiring a threading.Lock. By checking the context variable first, we avoid lock acquisition overhead and contention for the vast majority of log calls in modern Python versions.

3. Cached Protobuf Constants in Hot Path:
   Change: Created module-level constants for RpcLog levels (e.g., _LOG_LEVEL_INFO) and categories (_LOG_CATEGORY_SYSTEM).
   Impact: The on_logging method previously performed expensive attribute lookups (e.g., protos.RpcLog.Information) and function calls (.Value('System')) for every log record. These are now instant constant lookups.

Fixes #

---

Pull Request Checklist

Host-Worker Contract

• Does this PR impact the host-worker contract (e.g., gRPC messages, shared interfaces)?
  • If yes, have the changes been applied to:
    • azure_functions_worker (Python <= 3.12)
    • proxy_worker (Python >= 3.13)
  • If no, please explain why:

Worker Execution Logic

• Does this PR affect worker execution logic (e.g., function invocation, bindings, lifecycle)?
  If yes, please answer the following:

Python Version Coverage

• Does this change apply to both Python <=3.12 and 3.13+?
• If yes, have the changes been made to:
  • azure_functions_worker (Python <= 3.12)
  • runtimes/v1 / runtimes/v2 (Python >= 3.13)
• If no, please explain why:

Programming Model Compatibility (for Python 3.13+)

• Does this change apply to both:
  • V1 programming model (runtimes/v1)?
  • V2 programming model (runtimes/v2)?
• Explanation (if limited to one model):