eLLM Studio Data Operations Package

Welcome to the eLLM Studio Data Operations package! 🎉 This package enables streaming chat functionality with your AI assistant in eLLM Studio via WebSocket and GraphQL. It's designed for both frontend and backend implementations.

🚀 Features

- Data Ingestion: Ingest text data.

📦 Installation

``bash
npm i @e-llm-studio/data-engine
`

🛠️ Data Ingestion

Here’s how you can use the startChatStream function to set up the AI chat stream:

`typescript
import { ContentManager } from "@e-llm-studio/data-engine";

// Example usage:

const baseUrl = "https://dev-egpt.techo.camp";
const chatbotId = "3e94587c-733d-4e92-8d98-d5150c63b8cd";
const chatbotName = "Assistant1";
const organizationName = "techolution";
const username = "your-username";
const password = "your-password";
const contentManager = new ContentManager(
baseUrl,
chatbotId,
chatbotName,
organizationName,
username,
password
);

// Save text content example
contentManager
.saveTextContent("Title1", "Text Content")
.then((saveResponse) => {
console.log("Save Response:", saveResponse);
})
.catch((error) => {
console.error("Save Error:", error);
});

// Save text file content example
contentManager
.ingestTextFiles("File", "Text File Content")
.then((saveResponse) => {
console.log("Save Response:", saveResponse);
})
.catch((error) => {
console.error("Save Error:", error);
});
// For AccessToken


// Delete text content example
contentManager
.deleteTextContent(["671a391fbcce6ddc80e86099"])
.then((deleteResponse) => {
console.log("Delete Response:", deleteResponse);
})
.catch((error) => {
console.error("Delete Error:", error);
});
// Get Context Id
contentManager
.getContextId("6d25ddb4-e375-4adf-a98d-4d14114c2840")
.then((contextResponse) => {
console.log("Context Response:", contextResponse);
})
.catch((error) => {
console.error("Context Error:", error);
});

// Seamantic Search api
contentManager
.getSemanticSearch(
"model-name", // AI Model to use for semantic search
"search query", // The query string for searching
"request123", // Unique request ID for tracking
"user456", // User ID performing the search
0, // Offset for pagination
10 // Limit for pagination
)
.then((searchResponse) => {
console.log("Search Response:", searchResponse);
})
.catch((error) => {
console.error("Search Error:", error);
});

contentManager.ingestTextFilesText(
{
"contexts": [
{
"contentType": "local_file",
"files": [
{
"id": uuid(),
"name": "file name.txt",
"size": "file size",
"content": "file text content",
"createdTime": "2025-02-05T15:44:06.024Z",
"modifiedTime": "2025-02-05T15:44:06.024Z",
"mimeType": "text/plain"
}
],
"id": uuid()
}
],
"chatbotId": "chatBotId / assistant id "
}).then((context) => {
console.log("Ingestion Successful", ingestedData);
})
.catch((error) => {
console.error("Failed to save text content:", error);
});
``
📝 API Documentation
saveTextContent(title: string, textContent: string): Promise
Saves the provided text content under the specified title.

Parameters:
title (string): The title of the text content.
textContent (string): The content text to be saved.

deleteTextContent(contextInfoIds: string[]): Promise
Deletes the specified content using the provided array of contextInfoIds.

Parameters:
contextInfoIds (string[]): An array of IDs representing the content items to delete.
deleteContextId(fileId: string): Promise
Deletes a specific content item by its fileId.

Parameters:
fileId (string): The ID of the content item to delete from the chatbot’s context.
🔐 Authentication
The package requires user authentication. The ContentManager class will automatically log in using the provided credentials and manage access tokens internally.

Semantic Search
The getSemanticSearch method performs a semantic search based on a provided query and model, returning results from the chatbot’s knowledge base.

async ingestTextFiles(files: File[]): Promise

Processes and uploads plain text files, extracts content, and saves it to the chatbot’s context.

Parameters:

files (File[]): An array of text files to be ingested.

Returns:

A promise that resolves with IContextInfo on success or void if no valid text files are found.

async ensureLoggedIn(): Promise

Ensures that the user is logged in by checking the access token. If no token is present, it attempts to log in.

Returns:

A promise that resolves with the access token if login is successful.

Parameters:

aiModel (string): The name of the AI model to use for the search.
searchQuery (string): The search query string.
requestId (string): A unique identifier for tracking the search request.
userId (string): The ID of the user initiating the search.
offset (number): The number of results to skip (for pagination).
limit (number): The maximum number of results to return.
Returns: A promise that resolves with search results matching the query or an error message if the search fails.

👥 Community & Support

For any questions or issues, feel free to reach out via our GitHub repository or join our community chat! We’re here to help. 😊