feat: Add a travel planner powered by a custom-configured model.

[Linux2010]

Description

feat: Add a travel planner powered by a custom-configured model (openAI Compatible).

Merge reasons

• When running the demo of the current SDK, there will be situations where the MCP tool or function call cannot be used, and it is difficult to run.
• The current SDK demo supports Google's large model, but lacks demo examples of customizing the configuration of large models.
• The demo I provided, based on the standard protocol of A2A, supports custom configuration of large models and provides standard code for the agent, client, and server.
• I sincerely hope to provide developers who are new to the A2A project with an easy-to-get-started demo example. Therefore, if there are any parts that need to be modified, please leave a message. I hope my code can be helpful for the example of the SDK project.

Subsequent plans

• When the official SDK is updated, I will update this example to make it compatible with the latest SDK.
• Enrich the functions of the travel assistant, add a Python file, and demonstrate examples of using MCP and function - call.

Running Screenshot

• server

• client

• Follow the CONTRIBUTING Guide.
• Make your Pull Request title in the https://www.conventionalcommits.org/ specification.
• Ensure the tests and linter pass (Run nox -s format from the repository root to format)
• Appropriate docs were updated (if necessary)

[holtskinner]

Should this API Key be added here?

[Linux2010]

You're right. I didn't think it through carefully. I will revise the logic here. Thank you for pointing it out.

[Linux2010]

description

• In this optimization, the real API key is placed in the environment variables.

PS: The configuration method of conf.json refers to the implementation approach of Codex.

• codex

[holtskinner]

Can you please move this PR to the new samples repo? https://github.com/google-a2a/a2a-samples