Allow caching generators and async generators

[abidlabs]

Previously, we couldn't cache examples if the function was a generator or an async generator. This fixes that by getting the last yielded value and using that as the output. I also add a print statement to print out the status of caching. If there's a generator that takes a long time to finish iterating, users can notice that something went wrong and terminate (as suggested by @hysts)

Example:

import gradio as gr
import time

def test_generator(x):
    for y in range(len(x)):
        time.sleep(0.3)
        yield x[: y + 1]

gr.Interface(test_generator, "textbox", "textbox", examples=[["abcdef"]], cache_examples=True).queue().launch()

This will be useful in gr.ChatInterface so that users can cache examples for streaming chatbots.

Closes: #3570

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
gradio	✅ Ready (Inspect)	Visit Preview	Jul 14, 2023 11:33pm

[gradio-pr-bot]

All the demos for this PR have been deployed at https://huggingface.co/spaces/gradio-pr-deploys/pr-4927-all-demos

---

You can install the changes in this PR by running:

pip install https://gradio-builds.s3.amazonaws.com/9137f1caa015ba361ab2dacdd9b0071645a692ef/gradio-3.36.1-py3-none-any.whl

[freddyaboulton]

Looks good to me @abidlabs - nice we can finally support this

[abidlabs]

Thanks for the awesome fast review @freddyaboulton