Bug: asyncio.run fails when synchronous method is called from already running async context

[Akash-Kumar-Sen]

Description

• Gemini structured_predict was failing when called inside a synchronous method that is inside an asyncio event loop.
• Use from llama_index.core.async_utils import asyncio_run instead of asyncio.run() to make sure the synchronous methods also work inside a asyncio event loop.

Regression caused in PR:

Type of Change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

• I believe this change is already covered by existing unit tests

Suggested Checklist:

• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added Google Colab support for the newly added notebooks.
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I ran uv run make format; uv run make lint to appease the lint gods

[logan-markewich]

I'm not actually sure if this will solve the issue or not.

I think Ideally, we remove the need for asyncio.run completely from the sync code-path

[Akash-Kumar-Sen]

I am writing in detail how we encountered this issue:

• We were initially using Llamaindex version llama-index-core==0.12.49 and recently upgraded that to llama-index-core==0.14.5

• We had one Async FastAPI function. That has a synchronous method which uses GoogleGenAI.structured_predict

• It started throwing asyncio event loop error after the upgrade.

• After some investingation we got to know that this is the line that started causing the error that was not there in earlier release: https://github.com/run-llama/llama_index/pull/20240/files#diff-b0287e7fc5241f21fe7ad711693b2b7d21e494ca700ac892a353e5de633a037bL659

• We solved this by just calling asyncio_run(GoogleGenAI.astructured_predict)

Steps to reproduce:

• Call the GoogleGenAI.structured_predict from an asynchronous method

Now there could be two possible solutions:

1. Completely unwrap the asyncio.run and create synchronous alternatives for whatever is inside.
2. Remove all synchronous methods and just replace synchronous method = asyncio_run(asynchronous method)