[EAGLE-7291] Use PATs for API calls that involve external models, part 2

[tiit-clarifai]

Search and custom predict operations with an image input use the embedder model from the base workflow of the application. The test applications use the general-image-embedding model from clarifai/main, so they need to use a Personal Access Token to access the model from another application.

• Update all search calls that rank by embeddings (image data) to use a PAT.
• Update predict calls to embedding-classifier models to use a PAT.

[Copilot]

Pull request overview

Updates the Python gRPC client test suite to authenticate with Personal Access Tokens (PATs) for requests that implicitly invoke external (cross-application) embedder models via image inputs, ensuring tests can access models from clarifai/main.

Changes:

• Switched image-embedding-based input search tests to use metadata(pat=True) and added user_app_id=user_app_id() to the affected requests.
• Updated a model predict call (with image input) in the model CRUD test to use PAT auth and explicit user_app_id.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
tests/client/test_model_crud.py	Uses PAT auth + user_app_id for image-based PostModelOutputs that triggers an external embedder.
tests/client/test_inputs_searches.py	Uses PAT auth + user_app_id for image-ranking PostInputsSearches queries that require external embeddings.
tests/client/test_input_search_legacy.py	Uses PAT auth + user_app_id for legacy image-based PostSearches embedding searches.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.