🚀 Comprehensive Chatbot Desktop Application (GPT-4o Assistant)

📖 Project Overview

The GPT-4o Assistant is a sophisticated desktop chatbot application leveraging advanced GPT-based AI technology, structured agent interactions, embedding services, semantic search with vector databases, and a rich user-interface built with Python’s Flet framework. It is designed for robust, contextually-rich user interactions involving document parsing, dataset interrogation, web scraping, and general-purpose conversational agents.

🎯 Key Objectives

The GPT-4o Assistant aims to:

• Provide dynamic interactions through intelligent agents: Document Agent (DocAgent), Data Agent (DataAgent), Web Agent (WebAgent), and a General Agent (GeneralAgent).
• Integrate AI language models (GPT-4-turbo, GPT-3.5-turbo) for advanced conversational understanding.
• Utilize semantic embeddings (OpenAI’s ADA model) and vector services (Chroma DB) for efficient context retrieval.
• Enable intuitive interaction via a responsive GUI built on Python's Flet framework.
• Offer comprehensive file support (PDF, DOCX, CSV, Excel, TXT) and robust PDF-exporting capabilities for conversation archiving.

🗃️ Project Architecture and Core Components

📁 Directory Structure Overview

chatbot_desktop/
├── config/
│   └── config.py                     # API client configuration (OpenAI)
├── core/
│   ├── agents/
│   │   ├── base_agent.py             # Base class defining agent behaviors
│   │   ├── doc_agent.py              # Agent specialized in document interrogation
│   │   ├── data_agent.py             # Agent handling structured data queries
│   │   ├── web_agent.py              # Agent for fetching and interpreting web pages
│   │   └── general_agent.py          # General-purpose fallback conversational agent
│   └── agent_manager.py              # Central agent management & routing
│
├── services/
│   ├── ai_service.py                 # GPT model API interactions and prompts
│   ├── embedding_service.py          # Text embedding generation with OpenAI
│   └── vector_service.py             # Vector database (Chroma DB) semantic searching
│
├── storage/
│   ├── blackboard.py                 # Shared memory structure for agent coordination
│   └── file_handler.py               # Multi-format file parsing (TXT, PDF, DOCX, CSV, XLSX)
│
├── utils/
│   ├── chunker.py                    # Document segmentation for embeddings
│   └── hybrid_numeric_semantic.py    # Advanced semantic numeric analysis utilities
│
├── ui/
│   └── flet_app.py                   # User interface application built with Flet
│
├── temp/                             # Temporary storage for exports (PDF history, cache)
├── vector_db/                        # Persistent storage location for Chroma DB embeddings
├── requirements.txt                  # Python dependencies
└── README.md                         # Project documentation (this file)

🤖 Detailed Agent System

The project articulates a sophisticated agent-driven architecture, supported by:

1. Agent Manager (agent_manager.py)

• Central routing of queries based on current active context (Document, Data, Web).
• Agents can be automatically switched depending on content loaded (document or dataframe), or explicitly via UI interaction.

2. Intelligent Agents

Agent	Responsibility & Handling
DocAgent	Parses and manages textual documents; uses embeddings & vector search (Chroma DB) for semantic queries (vector_service.py).
DataAgent	Manages structured data (DataFrames), offers summaries, statistics & visualizations generated dynamically through pandas and matplotlib.
WebAgent	Fetches and analyzes web content, intelligently summarizing the fetched resources via GPT-based interpretation (ai_service.py).
GeneralAgent	Provides a flexible conversational fallback capable of general-purpose interactions, ensuring fluid conversations when specialized agents are inactive.

3. Agent Communication & Memory (Blackboard)

• A centralized global state (blackboard.py) enabling effective inter-agent memory sharing:
  • Stores history of interactions, file contexts, structured data, website contents, and intermediate analysis results.

📐 Embedding & Vector Services

• Text Embeddings:

  • Powered via OpenAI (embedding_service.py), using 'text-embedding-ada-002'.

• Semantic Search (Chroma DB):

  • Managed through vector_service.py, enabling rapid semantic search and similarity detection on document chunks stored persistently.

🗄️ File Handling & Document Management

• Robust multi-format file parsing (file_handler.py):

  • Text files (.txt)
  • PDF files parsing with pdfplumber
  • DOCX documents parsing with python-docx
  • CSV & Excel files parsing with pandas (dataframes)

• Chunking documents into segments appropriate for embedding and semantic search (chunker.py).

🖥️ Front-end Application (flet_app.py - Flet framework)

• Interactive GUI application powered by the modern, lightweight Python GUI toolkit "Flet".
• Offers conversation management, input/output display, PDF export functionality, file loading, agent switching, and app-wide settings management like theme toggling.

🔧 Configuration (settings.py & config.py)

• Manages OpenAI API keys and default embedding models securely through environment variables.

📊 Advanced Utility Modules

• Hybrid Semantic & Numeric Analysis (hybrid_numeric_semantic.py):
  • Allows for more sophisticated numerical data operations or hybrid semantic/numerical analytical capabilities.

✅ Key Functional Highlights

• AI-driven conversational agents with multi-domain specificity.
• Rich embedding-derived semantic search capabilities.
• Document and structured data integration for context enrichment.
• Elegant user-interface with comprehensive interactive controls.
• Persistent vector DB for scalable information retrieval.
• Robust conversation archiving & exporting as PDF.

🛠️ Setup & Installation (brief example)

git clone https://github.com/lechuck714/chatbot_desktop.git
cd chatbot_desktop
pip install -r requirements.txt

Ensure OPENAI_API_KEY environment variable is set correctly before running:

export OPENAI_API_KEY='your-key-here'

Launch with:

python ui/flet_app.py

🚧 Roadmap & Future Enhancements

1. Enhanced file-type support and richer preprocessing pipelines.
2. Extended UI features (real-time visualization, richer interactive widgets).
3. Scalable cloud deployment options.
4. Expanded agent capabilities and context-handling enhancements.

📋 Conclusion

This GPT-4o Assistant combines powerful GPT-based conversational AI, structured intelligent agents, and state-of-the-art semantic search techniques into a cohesive desktop tool, offering users intuitive and insightful interactions across document analyses, structured datasets interrogation, real-time web data parsing, and general-purpose conversations.