Much longer running time when ran as package / cli than webapp / notebook

[williamz-homelight]

Hi all, I've played with a query in webapp and it seemed to reliably finish between 30-60s. (doesn't matter if its "quick" mode or "detailed} mode)

However when running it in a plain python script like below it seemed to be "stuck " at "Researching for relevant information..."
"Getting relevant content based on query" (more so).

• What are these steps doing and what would you think is the main factor causing the delay?
• What env settings / default configs could be relevant? I suspect there's some kind of timeout / parallelization or use local llm / embedding v.s apis (like openai) default setting that's different?
• How to configure a researcher in general , to be fast like in webapp?

Weirdly enough, when I tried running the same query and almost identical code in jupyter notebook (using same venv installation) and it also ran fast.

Any help or ideas is much appreciated, I'm stumped and not thrilled about debugging through the code 😅

import asyncio

async def async_handler(query):
    researcher = GPTResearcher(query=query, report_type="custom_report")
    research_result = await researcher.conduct_research()
    report_custom = await researcher.write_report()
    return report_custom

loop = asyncio.get_event_loop()
research_obj = loop.run_until_complete(async_handler(query))
print(research_obj)