Skip to content

Latest commit

 

History

History
213 lines (143 loc) · 7.63 KB

JET.md

File metadata and controls

213 lines (143 loc) · 7.63 KB
Raw

Enhance Engagement Using Content Generation with OCI Generative AI

License: UPL Quality gate

Introduction

Using Oracle JET, create a user-friendly prompt-led user interface (UI) to interact with Oracle's new Generative AI service. This toolkit will configure your Generative AI Service connection so you can begin your journey with AI, or migrate your existing (local or Cloud) LLMs to the Oracle AppDev ecosystem.

Oracle JET(Preact) allows you to craft pixel-perfect UIs which are fast, lightweight, and engaging. Your code takes centre stage with Oracle JET, while its powerful features enable you to create dynamic user experiences quickly and reliably.

Oracle's Generative AI service allows developers to unlock a better user experience for chat systems, question-and-answer solutions, and much more. This project provides a front end to that service so you can experiment and get a sense of the immense power of Oracle Generative AI. This is an excellent starting point on your AI journey, and experienced developers will be able to quickly port their LLMs to leverage this powerful service.

Check out demo here

alt text here

Getting Started

0. Prerequisites and setup

Follow the links below to generate a config file and a key pair in your ~/.oci directory

After completion, you should have the following 2 things in your ~/.oci directory

  • A config file(where key file point to private key:key_file=~/.oci/oci_api_key.pem)
  • A key pair named oci_api_key.pem and oci_api_key_public.pem
  • Now make sure you change the reference of the key file in the config file
  • Append OCI Generative-AI service compartment and endpoint URL
vim service/python/server.py
#TODO: Update this section with your tenancy details
compartment_id = "ocid1.compartment.oc1.."
CONFIG_PROFILE = "DEFAULT"
config = oci.config.from_file("~/.oci/config", CONFIG_PROFILE)
endpoint = "https://inference.generativeai.<REGION>.oci.oraclecloud.com"
generative_ai_inference_client = (
    oci.generative_ai_inference.GenerativeAiInferenceClient(
        config=config,
        service_endpoint=endpoint,
        retry_strategy=oci.retry.NoneRetryStrategy(),
        timeout=(10, 240),
    )
)

1. (Optional) Modify websocket ports

  • In the root of the project directory run to edit ports
vim app/src/components/content/index.tsx
const gateway = ws://${window.location.hostname}:1234;
  • Update default port in Python websocket server:
vim service/python/server.py
async def start_server():
    await websockets.serve(handle_websocket, "localhost", 1234 )

2. Upload Public Key

3. Install all dependencies

We suggest you install dependencies in a virtual environment to avoid conflicts on your system.

  • Navigate to the server root folder
cd service/python
  • Create a virtual environment:
python3 -m venv venv
  • Activate your virtual environment:
. venv/bin/activate
  • Upgrade pip:
pip3 install --upgrade pip
  • Install requirements:
pip3 install -r requirements.txt

4. Start the websocket server app

Once dependencies are installed and your service credentials are updated you can run server.py

python3 server.py

5. Start JET Client

  • Open app directory:
cd ../../app
  • Install dependencies:
npm install
  • Run local version:
npx ojet serve
  • Or package for web deployment
npx ojet build web

You can now ask question to generate LLM based response. alt text here

Note that sample app can generate markdown. alt text here

Appendix: Token-based Authentication

Check Token-based Authentication for the CLI

config = oci.config.from_file('~/.oci/config', profile_name="DEFAULT")

def make_security_token_signer(oci_config):
    pk = oci.signer.load_private_key_from_file(oci_config.get("key_file"), None)
    with open(oci_config.get("security_token_file")) as f:
        st_string = f.read()
    return oci.auth.signers.SecurityTokenSigner(st_string, pk)

signer = make_security_token_signer(oci_config=config)
# Service endpoint
endpoint = "https://generativeai.aiservice.<Region>.oci.oraclecloud.com"

generative_ai_client = oci.generative_ai.generative_ai_client.GenerativeAiClient(config=config, service_endpoint=endpoint, retry_strategy=oci.retry.NoneRetryStrategy(), signer=signer)

Notes/Issues

Additional Use Cases like summarization and embedding coming soon.

To change output parameters edit server.py

    cohere_generate_text_request.max_tokens = 500 # choose the number of tokens 1-4000
    cohere_generate_text_request.temperature = 0.75 # adjust temperature 0-1
    cohere_generate_text_request.top_p = 0.7 # adjust top_p 0-1
    cohere_generate_text_request.frequency_penalty = 1.0 # adjust frequency_penalty

URLs

Contributing

This project is open source. Please submit your contributions by forking this repository and submitting a pull request! Oracle appreciates any contributions that are made by the open-source community.

License

Copyright (c) 2024 Oracle and/or its affiliates.

Licensed under the Universal Permissive License (UPL), Version 1.0.

See LICENSE for more details.

ORACLE AND ITS AFFILIATES DO NOT PROVIDE ANY WARRANTY WHATSOEVER, EXPRESS OR IMPLIED, FOR ANY SOFTWARE, MATERIAL OR CONTENT OF ANY KIND CONTAINED OR PRODUCED WITHIN THIS REPOSITORY, AND IN PARTICULAR SPECIFICALLY DISCLAIM ANY AND ALL IMPLIED WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. FURTHERMORE, ORACLE AND ITS AFFILIATES DO NOT REPRESENT THAT ANY CUSTOMARY SECURITY REVIEW HAS BEEN PERFORMED WITH RESPECT TO ANY SOFTWARE, MATERIAL OR CONTENT CONTAINED OR PRODUCED WITHIN THIS REPOSITORY. IN ADDITION, AND WITHOUT LIMITING THE FOREGOING, THIRD PARTIES MAY HAVE POSTED SOFTWARE, MATERIAL OR CONTENT TO THIS REPOSITORY WITHOUT ANY REVIEW. USE AT YOUR OWN RISK.