Fixes #335

Description

This fixes an error where specifying a built-in provider by name when using ModelConfig would fail, because router.resolve_provider(config.provider) without first calling providers.load_builtins_once(), which was only called when the provider was being inferred.

The actual fix here is tiny, which is to lift the load_builtins_once/load_plugins_once calls outside the if/then. Since the loading is done only once, the impact should be insignificant.

Before:

langextract.core.exceptions.InferenceConfigError:
    No provider found matching: 'ollama'. Available providers can be listed with list_providers()

How Has This Been Tested?

Manually and new unit test added.

Checklist:

• I have read and acknowledged Google's Open Source Code of conduct.
• I have read the Contributing page, and I either signed the Google Individual CLA or am covered by my company's Corporate CLA.
• I have made any needed documentation changes, or noted in the linked issue(s) that documentation elsewhere needs updating. (N/A)
• I have added tests, or I have ensured existing tests cover the changes
• I have followed Google's Python Style Guide and ran pylint over the affected code.