Use Replicate streaming API

[RomainGehrig]

Description

The Replicate integration did not use streaming text response as the integration predates v0.21.0.

Version Bump?

Did I bump the version in the pyproject.toml file of the package I am updating? (Except for the llama-index-core package)

• Yes
• No

Type of Change

• Small breaking change as setups with replicate library < 0.25.2 may not work (and definitively when < 0.21.0).

How Has This Been Tested?

• I tested the modified integration with versions 0.25.2 (april 2024), 0.34.2 (last version in 0.x), 1.0.2 (newest) and they all stream the text response. Here is the testing code:

import replicate
from llama_index.llms.replicate import Replicate

replicate.default_client._api_token = "<YOUR_API_KEY>"

llm = Replicate(model="meta/meta-llama-3-70b-instruct")
for ev in llm.stream_complete("How are you doing ?"):
    print(ev.delta, end="", flush=True)

[anoopshrma]

This is great, thanks for the contribution!