Improve README and add comprehensive test suite

[cpfiffer]

Overview

This PR significantly improves the README documentation and adds a comprehensive test suite to ensure all code examples work correctly.

README Improvements

• Restructured for clarity: More informative and concise structure
• Added quicklinks section: Easy navigation to docs, API reference, ADE, and Letta Cloud
• Better Hello World example: Replaced basic example with compelling memory self-correction demo
• Core Concepts section: Added concise explanation of MemGPT foundations (Memory Hierarchy, Memory Blocks, Agentic Context Engineering, Perpetual Agents)
• Added missing major features:
  • Sleep-time agents (background agents with shared memory)
  • Agent File import/export (.af format)
  • MCP Tools (Model Context Protocol integration)
  • Filesystem (agent access to files and folders)
  • Long-running agents with resumable streaming
  • Message Types (discriminated union handling)
• Fixed code examples: Corrected API signatures throughout (using block_label parameter, modify instead of update, retrieve method)
• Improved Contributing section: Updated with correct social links

Test Suite

Created comprehensive test suite covering all README examples:

• 12 test files (01_quick_start.py through 12_custom_http_client.py)
• run_all.sh script to execute all tests
• run_all.py programmatic test runner
• All tests validate that README examples work as documented