feat(autogen): upgrade to v0.7.5, fix all examples, and add 4 new examples

[martimfasantos]

Summary

• Upgrade AutoGen from v0.7.2 to v0.7.5 and rewrite all 8 existing examples to follow repo conventions, fix bugs, and add 4 new examples covering GraphFlow, termination conditions, state management, and custom agents — bringing the total to 12 hands-on examples
• All examples follow repo conventions (zero-padded filenames, dashed-border docstrings, section headers, settings-based config)
• Includes verified OUTPUTS.md with captured output and feature highlight annotations for all 12 examples

Examples

#	File	Feature
00	00_hello_world.py	Basic AssistantAgent, streaming output with Console
01	01_tools.py	FunctionTool, tool schema generation, tool call execution
02	02_streaming_and_metrics.py	Streaming responses, token usage statistics
03	03_structured_outputs.py	Pydantic model responses, chain-of-thought reasoning
04	04_human_in_the_loop.py	HandoffTermination, pausing and resuming with user input
05	05_multi_agent_teams.py	RoundRobin, Selector, MagenticOne, Swarm team presets
06	06_agents_as_tool.py	AgentTool, TeamTool, disabling parallel tool calls
07	07_memory.py	ListMemory, user preferences, memory-aware tool usage
08	08_graph_flow.py	DiGraphBuilder, sequential/parallel/conditional workflows
09	09_termination_conditions.py	MaxMessage, TextMention, Timeout, OR/AND combinators
10	10_state_management.py	save_state/load_state for agents and teams, checkpointing
11	11_custom_agents.py	Subclassing BaseChatAgent (CountDownAgent, EchoAgent)

Changes to Existing Files

File	Change
00–07 examples	Renamed from 0_–7_ to zero-padded 00_–07_, added convention-compliant docstrings with verified docs URLs, added await model_client.close(), wrapped in async def main()
02_streaming_and_metrics.py	Fixed bug: renamed tool function from sum to add_numbers to avoid shadowing Python builtin (caused infinite recursion)
04_human_in_the_loop.py	Rewrote to use HandoffTermination pattern instead of input() calls (violates "No Interactive Input" rule)
07_memory.py	Removed access to private _model_context attribute
pyproject.toml	Cleaned up from 40+ pinned dependencies to 5 clean >= specifiers
.env.example	Fixed default model from gpt-4o to gpt-4o-mini to match settings.py
README.md	Complete rewrite with features list, 12-example table, uv setup instructions
autogen-project/	Deleted unused subdirectory (leftover from earlier experiments)

Configuration

• All examples use OpenAI (gpt-4o-mini) via settings.py
• Dependencies: autogen-agentchat>=0.7.5, autogen-core>=0.7.5, autogen-ext[openai]>=0.7.5, pydantic>=2.10, pydantic-settings>=2.10
• Managed with uv package manager

Testing

All 12 examples ran successfully with exit code 0 and produced feature-showcasing output (see OUTPUTS.md for full details).