fix: enable structured outputs for chat models

[markacianfrani]

resolves #16

[Copilot]

Pull Request Overview

This PR enables structured outputs support for GitHub Models chat models by adding the supportsStructuredOutputs: true configuration option to the base options.

Key Changes

• Added supportsStructuredOutputs: true to the base configuration for GitHub Models chat models
• Added test coverage to verify the structured outputs configuration is properly set

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/provider.ts	Added supportsStructuredOutputs: true to the base options configuration
src/provider.test.ts	Added test to verify structured outputs are enabled for chat models

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[maraisr]

Thank you so much 🙏 there may be some models that do not support structuredOutputs. I do wonder if there is a way to use something similar to https://github.com/github/models-ai-sdk/blob/main/script/sync-model-id.ts to pull this information too, and to default to true as youre doing.

[markacianfrani]

I had a version baked into that with a call like:

    const response = await fetch(`${baseURL}/v1/chat/completions`, {
      method: 'POST',
      headers: {
        Authorization: `Bearer ${TOKEN}`,
        'Content-Type': 'application/json',
      },
      body: JSON.stringify({
        model: modelId,
        messages: [{role: 'user', content: 'Generate a person with name and age.'}],
        response_format: {
          type: 'json_schema',
          json_schema: {
            name: 'person',
            schema: {
              type: 'object',
              properties: {
                name: {type: 'string'},
                age: {type: 'number'},
              },
              required: ['name', 'age'],
            },
          },
        },
      }),
    })

and just looped through the models. but it took forever and, to be completely honest, I actually have no idea what the rate limits and usage limits actually are for the free GitHub Models.

Another thought: we could make it configurable at least as a providerOption or something so a consumer can at least always pass it in themselves but if the model doesn't support it, there's not really anything you can do anyway.