Add random seed access and callback methods to temporalio.workflow

[tconley1428]

Summary

• Add workflow.random_seed() to get current random seed from core
• Add workflow.register_random_seed_callback() to register for seed change notifications
• Add workflow.new_random() to create auto-reseeded Random instances

Changes

This PR adds three new methods to temporalio.workflow that provide access to and control over the workflow's random seed:

1. workflow.random_seed() - Returns the current random seed value from core
2. workflow.register_random_seed_callback(callback) - Registers callbacks that get invoked when the random seed changes (e.g., during workflow resets)
3. workflow.new_random() - Creates a new Random instance that automatically reseeds itself when the workflow seed changes

Implementation Details

• Added abstract methods to _Runtime class in workflow.py
• Implemented concrete methods in _WorkflowInstanceImpl class
• Enhanced _apply_update_random_seed to invoke registered callbacks
• Added seed change tracking with proper error handling
• Comprehensive test coverage including workflow reset scenarios

Use Cases

These methods enable workflows to:

• Access the deterministic random seed for custom randomization logic
• React to seed changes during workflow resets and replays
• Create multiple synchronized random number generators
• Maintain determinism while providing flexibility for complex random operations

Test Plan

• Added comprehensive test with workflow reset to verify seed change callbacks
• All existing tests pass
• Type checking passes
• Linting passes

Fixes #1319

🤖 Generated with Claude Code