feat: implement resilient timeout handling for model switcher component

[prasiddhi-105]

Description

This PR resolves the issue regarding model switching delays by implementing comprehensive timeout handling mechanisms across both the client-side and server-side components. This prevents the application from hanging indefinitely when loading large models or waiting on a slow local Ollama instance.

Changes Implemented

🔹 Backend (backend/routes/models.py)

• Added a dedicated POST /api/models/{model_name}/switch endpoint.
• Wrapped the model loading logic inside an explicit asyncio.timeout block bounded to 15.0 seconds.
• Gracefully handles asyncio.TimeoutError and converts it into a 504 Gateway Timeout HTTP response.

🔹 Frontend (frontend/src/services/api.js)

• Created switchModelWithTimeout using a client-side AbortController as a backup boundary.
• Dispatches an explicit client-side fallback cancellation if the endpoint request takes longer than 15 seconds.

🔹 Testing (backend/tests/test_model_cache.py)

• Added a robust test_switch_model_timeout async unit test case using unittest.mock.patch to simulate deep runtime hangs and verify the correct 504 error codes are thrown.

Verification Results

All local backend unit tests ran successfully within the execution container environment:

Closes #814

[vercel]

@prasiddhi-105 is attempting to deploy a commit to the Darshan's projects Team on Vercel.

A member of the Team first needs to authorize it.