Alita SDK, built on top of Langchain, enables the creation of intelligent agents within the Alita Platform using project-specific prompts and data sources. This SDK is designed for developers looking to integrate advanced AI capabilities into their projects with ease.
Before you begin, ensure you have the following requirements met:
- Python 3.10
- An active deployment of Project Alita
- Access to personal project
First, you need to install the Langchain library. Alita SDK depends on Langchain for its core functionalities. You can install Langchain using pip:
pip install langchainNext, clone the Alita SDK repository (assuming it's available on GitHub or another source):
git clone https://github.com/ProjectAlita/alita-sdk.git
cd alita-sdkInstall the SDK along with its dependencies:
pip install -r requirements.txtBefore running your Alita agents, set up your environment variables. Create a .env file in the root directory of your project and include your Project Alita credentials:
AUTH_TOKEN=<your_api_token>
PROJECT_ID=<your_project_id>
INTEGRATION_UID=<your_integration_uid>Ensure you load these variables in your application:
from dotenv import load_dotenv
load_dotenv('.env')The Alita SDK allows you to create and execute agents with ease. Here's a simple example to get you started:
import logging
from os import environ
from dotenv import load_dotenv
logging.basicConfig(level=logging.INFO)
load_dotenv('.env')
logger = logging.getLogger(__name__)
from alita_sdk.utils.streamlit import run_streamlit
try:
import streamlit as st
except ImportError:
logger.error("Streamlit not found, please install it using `pip install streamlit`")
exit(1)
from alita_sdk.llms.alita import AlitaChatModel
# Minimal set of setting for AlitaChatModel
settings = {
"deployment": "https://eye.projectalita.ai",
"model": "gpt-4-0125-preview",
"api_key": environ.get("AUTH_TOKEN"),
"project_id": environ.get("PROJECT_ID"),
"integration_uid": environ.get("INTEGRATION_UID"),
}
print(settings)
if 'messages' not in st.session_state:
llm = AlitaChatModel(**settings)
st.session_state.messages = []
st.session_state.agent_executor = llm.client.application(llm, <application_id>, <application_version_id>)
run_streamlit(st)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent