Agentic Workflow for ViewMo Platform

Introduction

Agentic Workflow is a research-focused project designed to enhance the ViewMo platform, a comprehensive research tool that empowers users to interact with their documents in innovative ways. The project introduces intelligent agents to simplify complex concepts, generate high-quality assessments, and provide actionable insights for users. Additionally, it offers a CLI tool and Python SDK for developers to integrate these features into their own products seamlessly.

Key Features

1. Document Interaction:

   • Users can upload documents (PDFs, Docs) and ask questions based on the content.
   • The agent simplifies explanations to make complex concepts easier to understand.
     
     
     

2. Graphical Notation for Complex Concepts:

   • Users can highlight sections of the document they find difficult to understand.
   • Based on the selected difficulty level (Easy, Medium, Hard), the system generates graphical notations to explain the concepts.
     
     
     

3. Knowledge Evaluation:

   • The system generates high-standard multiple-choice questions (MCQs) based on the uploaded documents.
   • Users can select the number of questions, and the agent will handle the rest.
   • Tracks the time taken to answer questions and provides insights into areas where the user needs improvement (future implementation).
     
     
     

4. Developer-Friendly CLI Tool:

   • Provides the Intellibot Python package, available on PyPI.
   • Developers can customize agentic workflows using JSON schemas.
   • Embed chatbots into their products by training them with documents uploaded to the ViewMo platform.
   • Advanced security features like service keys ensure secure integration.
     
     
     

5. Secure and Scalable Infrastructure:

   • Each user has a secure, private vectorDB and cloud bucket for their documents.
   • Uses JWT tokens and service accounts to secure the platform from unauthorized access.

---

Technologies Used

• Frameworks:
  • CrewAI
  • Langchain
  • FastAPI (for API development)
• Database:
  • GCP-based PostgreSQL
• Vector Database:
  • FAISS
• Cloud Platform:
  • Google Cloud Platform (GCP)
• Security:
  • JWT Tokens
  • Service Accounts
• Package Management:
  • PyPI (for Intellibot Python package)

---

How It Works

1. User Interaction:
   • Users log in to the ViewMo platform and upload documents.
   • They can ask questions, highlight difficult sections, or request MCQ generation.
     
     
     
2. Agentic Workflow:
   • Agents process the documents, simplify explanations, generate graphical notations, and create MCQs.
   • Insights are provided based on user interactions and performance.
     
     
     
3. Developer Integration:
   • Developers use the Intellibot Python package to integrate agentic workflows into their products.
   • JSON schemas define custom workflows, and service accounts manage permissions.

---

Setup Instructions

For Users:

1. Log in to the ViewMo platform.
2. Upload your documents (PDFs, Docs).
3. Use the features like question answering, graphical notation, and MCQ generation.

For Developers:

1. Install the Intellibot package:

   pip install intellibot

2. Use the provided Python SDK to integrate agentic workflows into your product.

3. Upload documents and define workflows using JSON schemas.

4. Secure your integration using service keys.

---

Usage

For Users:

• Upload documents and interact with them using the platform's features.
• Evaluate your knowledge using the MCQ generator and track your progress.

For Developers:

• Use the Intellibot package to embed chatbots and agentic workflows into your applications.
• Train chatbots by uploading documents to the ViewMo platform.
• Manage permissions and security using service accounts.

---

API and SDK Links

• Agents API: GitHub Repository
• Intellibot PyPI: PyPI Page

---

Contributors

• Pawan Perera
  • Email: 1pawanpinsara@gmail.com
  • LinkedIn: Pawan Perera
  • Website: pawanperera.com

---

License

This project is licensed under the MIT License. See the LICENSE file for details.

---

Future Scope

• Implement insights for MCQ performance to identify areas where users need improvement.
• Expand the graphical notation system to support more document formats.
• Add support for additional third-party integrations.

---

Contact Information

For questions, collaborations, or feedback, feel free to reach out:

• Email: 1pawanpinsara@gmail.com
• GitHub: PinsaraPerera
• LinkedIn: Pawan Perera
• Website: pawanperera.com