feat: Add LangMem integration example with branch-based memory isolation

This commit introduces a comprehensive Python example demonstrating how to use
ProllyTree's versioned key-value store as a backend for LangMem's memory
management system, featuring Git-like branching for AI agent memory isolation.

## New Features

### ProllyTree + LangMem Integration
- **ProllyTreeLangMemStore**: Custom BaseStore implementation adapting ProllyTree's
  VersionedKvStore to work seamlessly with LangMem's memory management
- **Full BaseStore Interface**: Complete compatibility with LangGraph's storage
  interface (get, put, delete, list, search, batch operations)
- **Vector Embeddings**: Support for OpenAI embeddings with metadata storage
- **Versioned Operations**: All memory operations create Git commits for full audit trail

### Branch-Based Memory Isolation
- **WorktreeManager Integration**: Git-like branching for isolated memory contexts
- **Branch Management API**: create_branch(), switch_branch(), list_branches(), merge_branch()
- **Context Isolation**: Each branch maintains completely separate memory state
- **Intelligent Merging**: Support for conflict resolution strategies when merging branches

### Comprehensive Example
- **Graceful Degradation**: Works with or without LangMem dependencies installed
- **Interactive Demo**: Shows memory storage, search, versioning, and branching
- **Integration Ready**: Demonstrates LangMem memory tools and managers
- **Production Patterns**: Includes error handling, logging, and best practices

## Use Cases Enabled
- Multi-user memory isolation
- A/B testing of AI agent behaviors
- Experimental feature development
- Context-aware agent systems
- Parallel agent execution with memory isolation

## Files Added/Modified
- `python/examples/langmem_integration.py` - Main integration example (870+ lines)
- `python/examples/requirements.txt` - Added LangMem dependencies
- `python/examples/run_examples.sh` - Added new `langmem` example option

## Architecture Benefits
- **Versioned Memory**: Every memory operation creates a Git commit
- **Branch Isolation**: Use ProllyTree branches for different memory contexts
- **Cryptographic Verification**: All memory operations are Merkle-tree verified
- **Efficient Storage**: ProllyTree's probabilistic tree structure for optimal performance
- **Cross-Language**: Works with both Python and Rust ecosystems

The implementation provides a robust foundation for persistent, versioned AI agent
memories with full audit trails and Git-like workflow capabilities.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: zhangfengcdt