openedu-ai-workspace 0.1.26

A package for ai workspace (project) features.

openedu_ai_workspace

A package for AI workspace (project) features, enabling management of workspaces and documents, with integration for MongoDB, Qdrant, and Azure OpenAI embeddings.

Overview

openedu_ai_workspace provides a robust framework for managing AI-centric projects. It allows developers to create, organize, and interact with workspaces, store and retrieve documents, and leverage powerful AI capabilities through integrations with vector databases and embedding models.

Features

• Workspace Management:
  • Create, find, update, and delete workspaces.
  • Manage workspace-specific instructions and chat sessions.
• Document Management:
  • Upload documents to workspaces.
  • Retrieve document metadata.
• Database Integration:
  • Uses MongoDB for storing workspace and document metadata.
  • Integrates with Qdrant for efficient vector storage and similarity search of document embeddings.
• AI Embeddings:
  • Utilizes Azure OpenAI Embeddings for processing and embedding document content.

Prerequisites

• Python 3.10 or higher.
• Access to a running MongoDB instance.
• Access to a running Qdrant instance.
• Azure OpenAI Service:
  • API Key
  • Endpoint URL
  • Embedding Model Deployment Name

Installation

Install openedu_ai_workspace using pip:

pip install openedu_ai_workspace

Configuration

The package relies on environment variables for critical configurations, such as database connection strings and API credentials. Ensure these are set in your environment or a .env file:

• MONGODB_URL: The connection URI for your MongoDB instance (e.g., mongodb://localhost:27017/aicore).
• QDRANT_URL: The URL for your Qdrant instance (e.g., http://localhost:6333).
• AZURE_OPENAI_API_KEY: Your Azure OpenAI API key.
• AZURE_OPENAI_ENDPOINT: Your Azure OpenAI endpoint URL.
• AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME: The deployment name for your Azure OpenAI embedding model (e.g., embedding).

Usage Examples

First, ensure you have a .env file in your project root with the necessary credentials, or that the environment variables are set.

# .env example
MONGODB_URL="mongodb://localhost:27017/mydatabase"
QDRANT_URL="http://localhost:6333"
AZURE_OPENAI_API_KEY="your_azure_openai_api_key"
AZURE_OPENAI_ENDPOINT="your_azure_openai_endpoint"
AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME="your_embedding_deployment_name"

Initializing the Workspace Client

from dotenv import load_dotenv
from ai_workspace import Workspace, WorkspaceSchema, DocumentSchema
from ai_workspace.database import MongoDB, Qdrant
from langchain_openai import AzureOpenAIEmbeddings
import os

# Load environment variables from .env file
load_dotenv()

# Configuration (fetched from environment variables)
MONGODB_URL = os.getenv("MONGODB_URL", "mongodb://localhost:27017/aicore")
QDRANT_URL = os.getenv("QDRANT_URL", "http://localhost:6333")
AZURE_OPENAI_API_KEY = os.getenv("AZURE_OPENAI_API_KEY")
AZURE_OPENAI_ENDPOINT = os.getenv("AZURE_OPENAI_ENDPOINT")
AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME = os.getenv("AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME", "embedding")

# Initialize database connections and embeddings model
mongodb = MongoDB(uri=MONGODB_URL)
qdrant = Qdrant(uri=QDRANT_URL)
embedding_model = AzureOpenAIEmbeddings(
    azure_deployment=AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME,
    api_key=AZURE_OPENAI_API_KEY,
    azure_endpoint=AZURE_OPENAI_ENDPOINT
)

# Create Workspace instance
ai_workspace = Workspace(mongodb=mongodb, qdrant=qdrant, embedding=embedding_model)

print("Workspace client initialized.")

Creating a Workspace

from ai_workspace.packages.shared import IDGenerator # Assuming IDGenerator is part of your shared utilities

id_generator = IDGenerator()
new_workspace_id = str(next(id_generator))

workspace_data = WorkspaceSchema(
    workspace_id=new_workspace_id,
    title="My First AI Workspace",
    description="A dedicated space for exploring AI concepts.",
    user_id="user_alpha_001",
    instructions="Default instructions for tasks in this workspace."
    # chat_sessions can be omitted or initialized as an empty list
)

try:
    created_workspace_id = ai_workspace.create(workspace_data)
    print(f"Workspace created successfully with ID: {created_workspace_id}")
except Exception as e:
    print(f"Error creating workspace: {e}")

Adding Knowledge (Uploading a Document)

# Assuming 'created_workspace_id' is available from the previous step
if 'created_workspace_id' in locals():
    document_data = DocumentSchema(
        workspace_id=created_workspace_id,
        file_name="introduction_to_ai.txt",
        file_mime="text/plain",
        file_suffix=".txt",
    )
    
    # Content for the document
    document_content_texts = [
        "Artificial intelligence (AI) is intelligence demonstrated by machines.",
        "It encompasses various sub-fields like machine learning and natural language processing."
    ]
    
    # Optional metadata for the document
    document_metadata = {"source": "internal_notes", "version": "1.0"}

    try:
        document_id = ai_workspace.add_knowledge(
            document=document_data,
            raw_texts=document_content_texts,
            metadata=document_metadata,
            to_workspace=True  # This flag links the document to the workspace's document list
        )
        print(f"Document '{document_data.file_name}' added with ID: {document_id} to workspace {created_workspace_id}")

        # Optionally, retrieve the document to verify
        retrieved_doc = ai_workspace.document_client.get_document(document_id)
        if retrieved_doc:
            print(f"Successfully retrieved document: {retrieved_doc.file_name}")
    except Exception as e:
        print(f"Error adding document: {e}")
else:
    print("Workspace ID not found. Please create a workspace first.")

Retrieving a Workspace

# Assuming 'created_workspace_id' is available
if 'created_workspace_id' in locals():
    try:
        retrieved_workspace = ai_workspace.find_workspace_by_id(created_workspace_id)
        if retrieved_workspace:
            print(f"Retrieved workspace: '{retrieved_workspace.title}' (ID: {retrieved_workspace.workspace_id})")
            print(f"Instructions: {retrieved_workspace.instructions}")
        else:
            print(f"Workspace with ID {created_workspace_id} not found.")
    except Exception as e:
        print(f"Error retrieving workspace: {e}")
else:
    print("Workspace ID not found for retrieval.")

Updating Workspace Instructions

# Assuming 'created_workspace_id' is available
if 'created_workspace_id' in locals():
    new_instructions = "Focus on summarization tasks for all documents."
    try:
        success = ai_workspace.update_instructions(created_workspace_id, new_instructions)
        if success:
            print(f"Workspace instructions updated for ID: {created_workspace_id}")
            # Verify by fetching the workspace again
            updated_workspace = ai_workspace.find_workspace_by_id(created_workspace_id)
            if updated_workspace:
                print(f"New instructions: {updated_workspace.instructions}")
        else:
            print(f"Failed to update instructions for workspace ID: {created_workspace_id}")
    except Exception as e:
        print(f"Error updating instructions: {e}")
else:
    print("Workspace ID not found for updating instructions.")

Adding a Chat Session to a Workspace

# Assuming 'created_workspace_id' is available
if 'created_workspace_id' in locals():
    new_session_id = "chat_session_beta_002"
    try:
        success = ai_workspace.add_session(created_workspace_id, new_session_id)
        if success:
            print(f"Chat session '{new_session_id}' added to workspace ID: {created_workspace_id}")
            # Verify by fetching the workspace again
            workspace_with_session = ai_workspace.find_workspace_by_id(created_workspace_id)
            if workspace_with_session and new_session_id in workspace_with_session.chat_sessions:
                print(f"Verified: Session '{new_session_id}' is in chat_sessions: {workspace_with_session.chat_sessions}")
        else:
            print(f"Failed to add chat session to workspace ID: {created_workspace_id}")
    except Exception as e:
        print(f"Error adding chat session: {e}")
else:
    print("Workspace ID not found for adding a chat session.")

Deleting a Workspace

# Be cautious with this operation.
# Assuming 'created_workspace_id' is available and you want to delete it.
# if 'created_workspace_id' in locals():
#     try:
#         success = ai_workspace.delete_workspace(created_workspace_id)
#         if success:
#             print(f"Workspace with ID: {created_workspace_id} and its associated Qdrant collection deleted successfully.")
#             # Try to find it again, it should fail or return None
#             # verify_deleted_workspace = ai_workspace.find_workspace_by_id(created_workspace_id)
#             # if not verify_deleted_workspace:
#             #     print(f"Verified: Workspace {created_workspace_id} no longer exists.")
#         else:
#             print(f"Failed to delete workspace ID: {created_workspace_id}")
#     except Exception as e:
#         print(f"Error deleting workspace: {e}")
# else:
#     print("Workspace ID not found for deletion.")

(The delete operation is commented out by default in this example to prevent accidental data loss during example runs.)

Running Tests

To run the test suite, ensure you have pytest installed (pip install pytest) and that your MongoDB and Qdrant instances are running and accessible with the configurations specified in your test environment (or .env file loaded by tests).

Navigate to the project root directory and run:

pytest

The tests will cover various functionalities of the Workspace and Document clients, including creation, retrieval, updates, and deletions.

Project Structure

The project is organized as follows:

openedu_ai_workspace/
├── src/
│   ├── ai_workspace/
│   │   ├── __init__.py         # Makes 'ai_workspace' a package, exports main classes
│   │   ├── database/           # Database interaction modules (mongo.py, qdrant.py)
│   │   ├── exceptions/         # Custom exceptions and handlers
│   │   ├── packages/           # Core logic for Workspace, Document, Embeddings
│   │   ├── schemas/            # Pydantic models for data validation
│   │   └── utils/              # Utility functions (e.g., logger)
│   └── openedu_ai_workspace.egg-info/ # Packaging metadata
├── tests/                      # Unit and integration tests
│   ├── test_document.py
│   └── test_workspace.py
├── .gitignore
├── pyproject.toml              # Project metadata and build configuration
├── README.md                   # This file
└── ...                         # Other configuration files

Key Dependencies

This package relies on several key libraries:

• langchain-openai: For Azure OpenAI embeddings.
• langchain-qdrant: For Qdrant vector store integration with LangChain.
• pydantic: For data validation and settings management.
• pymongo: The official Python driver for MongoDB.
• qdrant-client: The official Python client for Qdrant.
• python-dotenv: For loading environment variables from .env files.
• sentence-transformers: (Though not directly in dependencies, it's a common underlying library for embeddings if not using Azure OpenAI directly for all embedding tasks, or if other embedding models were considered).
• snowflake-id: For generating unique IDs (as seen in IDGenerator usage in tests).

Contributing

Contributions are welcome! If you have suggestions for improvements or encounter any issues, please feel free to open an issue or submit a pull request on the project's repository.

License

This project is licensed under the MIT License. See the LICENSE file for more details (if a LICENSE file is present in the repository).