Skip to content

Latest commit

 

History

History
104 lines (88 loc) · 4.52 KB

README.md

File metadata and controls

104 lines (88 loc) · 4.52 KB

CodeNova

Overview

CodeNova is an AI-powered code assistant designed to streamline software development by automating code generation, debugging, documentation, and code explanations. It integrates multiple AI models to provide a seamless and intelligent coding experience.integrating IBM's Granite LLMs to enhance software development workflows. This project was built for the Generative AI Hackathon with IBM GRANITE and provides business-valued solutions by automating various software engineering tasks.

To access CodeNova through streamlit click CodeNova (open in dark mode for better user experience)

Features

The application includes six main tools:

  1. Code Wizard - Automatically generate code based on user input prompt.
  2. Fix-it - Identify and fix issues in the provided code.
  3. Code Whisperer - Provide detailed text & voice-based explanations of code functionality.
  4. Code Optimizer - Improve code structure and efficiency.
  5. DocuBot - Generate structured documentation in TXT, PDF and DOCX formats.
  6. AI Pair Coder - Assist developers in real-time coding collaborations.

Technology Used

CodeNova leverages IBM Granite AI models for core functionalities, ensuring high-performance AI-driven code assistance. Specific features are powered by industry-leading AI services:

  • Python
  • Streamlit - Used for Front-End UI and deployment
  • IBM Granite-34B-Code-Instruct: – Used for code generation, debugging, optimization, and pair programming.
  • IBM Granite-20B-Code-Instruct: - Powers text-based code explanations.
  • Google Gemini Pro – Enables comprehensive document generation to streamline documentation tasks.
  • Google Text-to-Speech – Provides text and voice-based code explanations, enhancing developer understanding.

Installation and Setup (For running it locally)

  1. Clone the repository:

    git clone https://github.com/allen-reji/CodeNova.git
cd CodeNova

  2. Create a virtual environment and activate it:

    python -m venv venv
source venv/bin/activate  # On Windows use `venv\Scripts\activate`

  3. Install dependencies:

    pip install -r requirements.txt

  4. Set Up API Keys: CodeNova uses IBM Granite, Google Gemini Pro, and Google Text-to-Speech. Ensure you have the necessary API keys configured in a .env file.

    • Create a .env file in the project root directory:
    touch .env
    • Add your API keys:
    IBM_API_KEY="your-ibm-api-key"
GOOGLE_API_KEY="your-google-api-key"

  5. In config.py, replace Gemini_API_key=st.secrets["GEMINI_API_KEY"] and API_KEY = os.getenv("IBM_CLOUD_API_KEY") with direct retrieval from .env file by doing the below.

    from dotenv import load_dotenv

# Load environment variables from .env file
load_dotenv()

#Retrieve API Keys from .env file
Gemini_API_key = os.getenv("GEMINI_API_KEY")
API_KEY = os.getenv("IBM_CLOUD_API_KEY")

  6. In tts_helper.py replace lines 9-19 with the following.

    os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "{Enter file path for Google Text-to-Speech API JSON file}"

  7. Run the Application:

    streamlit run app.py

    The app will be available at http://localhost:8501

How to Use

  1. Navigate through the sidebar to select a tool.
  2. Input your code or query and click 'Generate'.
  3. View, copy, or download the output.
  4. Use the documentation generator for structured reports.

Business Value

  • Accelerated Development: AI-driven code generation and refactoring reduce development time.
  • Enhanced Code Quality: Intelligent debugging and optimization ensure clean, maintainable code.
  • Accessibility: Provides multiple output formats, including audio explanations.
  • AI-Powered Collaboration: Acts as a virtual pair programmer, improving productivity and reducing bottlenecks in software development.

Future Enhancements

  • Enhanced AI Models
  • Integration with IDE plugins.
  • Advanced AI-driven pair programming.
  • Enterprise Adoption
  • Global Developer Reach

License

This project is licensed under the MIT License.

Contact

Developed for Generative AI Hackathon with IBM GRANITE