AI Project Management Assistant

This project implements an AI assistant designed to help manage project information using a knowledge base. It leverages Large Language Models (LLMs) via the LangChain framework to understand user queries, retrieve relevant project details, and update the knowledge base with new information.

Core Components

1. agent.py: This is the central script containing the AI logic. It defines two LangChain agents:
   • Answering Agent: Queries the project_kb.json file to answer user questions about the project based on the stored information.
   • Updating Agent: Analyzes conversations (user input and AI responses) to identify and extract new project details, updating the project_kb.json accordingly.
2. project_kb.json: A JSON file acting as the project's knowledge base. It stores information categorized under standard project management domains (e.g., scope, schedule, cost, risk, stakeholders).
3. UI Files (chatUI-wx.py, PYSide6, PYSimpleGUI, etc.): Various Python scripts implementing graphical user interfaces (GUIs) using different libraries (wxPython, PySide6, PySimpleGUI). These provide front-ends for interacting with the AI agent. Note: It appears multiple UI frameworks have been explored.

Functionality

• Query Project Information: Users can ask questions about the project (e.g., "What is the project schedule?", "Who are the key stakeholders?"). The AI consults the knowledge base to provide answers.
• Update Knowledge Base: As users interact with the AI, the system can automatically identify relevant new information (e.g., changes in scope, new risks identified) and update the knowledge base.
• Multiple UI Options: The project includes experiments with different GUI toolkits, offering potential choices for the user interface.

Setup

1. Clone the repository:

   git clone <repository_url>
   cd <repository_directory>

2. Install dependencies:

   pip install langchain openai python-dotenv # Add other UI-specific dependencies (e.g., wxPython, PySide6, PySimpleGUI) if needed

3. Set up Environment Variables: Create a .env file in the root directory and add your OpenAI API key:

   OPENAI_API_KEY='your_openai_api_key_here'
   

4. Initialize Knowledge Base: Ensure project_kb.json exists. If not, running agent.py might create a default empty one, or you can create it manually with the basic structure seen in the agent.py read_knowledge_base function.

Usage

1. Run a UI script: Choose one of the UI scripts (e.g., chatUI-wx.py) and run it:

   python chatUI-wx.py

   (Replace chatUI-wx.py with your preferred UI file)
2. Interact with the AI: Use the GUI to ask questions or provide information about the project.

Future Development / Considerations

• Consolidate UI: Choose a primary UI framework and remove or refactor the others for clarity.
• Error Handling: Enhance error handling, especially around knowledge base updates and API interactions.
• Chat History: Implement persistent chat history within the agents for more contextual conversations. The current implementation seems to reset history for each input.
• Knowledge Base Schema: Potentially define a more rigid schema for project_kb.json for better validation and consistency.
• Tool Refinement: Explore adding more LangChain tools for enhanced capabilities (e.g., web search, document loading).