client/README.md

h2oGPT Client

A Python thin-client for h2oGPT.

Prerequisites

• Python 3.8+

If you don't have Python 3.8 in your system, you can use Conda.

conda create -n h2ogpt_client_build -y
conda activate h2ogpt_client_build
conda install python=3.8 -y

Download Client Wheel

Install the latest nightly wheel from S3.

pip install https://s3.amazonaws.com/artifacts.h2o.ai/snapshots/ai/h2o/h2ogpt_client/latest-nightly/h2ogpt_client-0.1.0-py3-none-any.whl

Nightly releases can also be found here

Build Client Wheel

If want to build fresh wheel from main branch instead of getting nightly, follow these instructions.

Setup

:information_source: Poetry is used as the build tool.

rm -rf client/.poetry/
make -C client setup

Build

make -C client build

Distribution wheel file can be found in the client/dist directory. This wheel can be installed in the primary h2oGPT environment or any other environment, e.g.

pip uninstall -y h2ogpt_client
pip install client/dist/h2ogpt_client-*-py3-none-any.whl

Usage

Based upon test code and test code test_readme_example:

def test_readme_example(local_server):
    import os
    import asyncio
    from h2ogpt_client import Client

    if local_server:
        client = Client("http://0.0.0.0:7860")
    else:
        h2ogpt_key = os.getenv("H2OGPT_KEY") or os.getenv("H2OGPT_H2OGPT_KEY")
        if h2ogpt_key is None:
            return
        # if you have API key for public instance:
        client = Client("https://gpt.h2o.ai", h2ogpt_key=h2ogpt_key)

    # Text completion
    text_completion = client.text_completion.create()
    response = asyncio.run(text_completion.complete("Hello world"))
    print("asyncio text completion response: %s" % response)
    # Text completion: synchronous
    response = text_completion.complete_sync("Hello world")
    print("sync text completion response: %s" % response)

    # Chat completion
    chat_completion = client.chat_completion.create()
    reply = asyncio.run(chat_completion.chat("Hey!"))
    print("asyncio text completion user: %s gpt: %s" % (reply["user"], reply["gpt"]))
    chat_history = chat_completion.chat_history()
    print("chat_history: %s" % chat_history)
    # Chat completion: synchronous
    reply = chat_completion.chat_sync("Hey!")
    print("sync chat completion gpt: %s" % reply["gpt"])

test_readme_example(local_server=True)

:warning: Note: Client APIs are still evolving. Hence, APIs can be changed without prior warnings.

Development Guide

Test

In an h2oGPT environment with the client installed, can run tests that test client and server.

Test with h2oGPT env

1. Install test dependencies of the Client into the h2oGPT Python environment.

make -C client setup_test

2. Run the tests with h2oGPT.

pytest client/tests/

Test with an existing h2oGPT server

If you already have a running h2oGPT server, then set the H2OGPT_SERVER environment variable to use it for testing.

make H2OGPT_SERVER="http://0.0.0.0:7860" -C client test