AI Code Assistant

A sophisticated Streamlit-based application offering AI-powered code assistance and security scanning through multiple language models and advanced vector search capabilities. The system leverages various LLMs for different code-related tasks while maintaining context through vector-based document indexing.

Features

Core Capabilities

• Multi-Model AI Processing

  • Mistral (Ollama)
  • CodeLlama (Ollama)
  • LLaMA 3 (Groq)
  • GPT-4 (Optional OpenAI integration)

• Vector-Based Document Management

  • Persistent index storage
  • Real-time index updates
  • Context-aware querying
  • Efficient document retrieval

• Task Processing

  • Production code generation
  • Code review and analysis
  • Documentation generation
  • Test case creation
  • Contextual querying

Security Scanner

• Advanced Code Security Scanning
  • Quick, Deep, and Custom scan options
  • Comprehensive vulnerability detection
  • Git repository scanning with authentication support
  • Detailed vulnerability reports with severity filters
  • Export scan results in JSON or PDF formats

Technical Implementation

• Streamlit-based user interface
• Hugging Face embedding model integration
• Vector store indexing for document management
• Multi-model task routing system
• Persistent storage management

Installation

System Requirements

• Python 3.8+
• Ollama installation
• Groq API access
• Optional: OpenAI API access

Setup Process

1. Repository Clone:

git clone https://github.com/AKKI0511/AI-Code-Generator.git
cd AI-Code-Generator

2. Dependency Installation:

pip install -r requirements.txt

3. Environment Configuration:

GROQ_API_KEY=your_groq_api_key
OPENAI_API_KEY=your_openai_api_key  # Optional

Execution

Launch the application:

streamlit run main.py

Architecture

Directory Structure

project/
├── main.py                 # Application entry point
├── app.py                  # Core application logic
├── constants.py            # System constants
├── requirements.txt        # Dependencies
├── components/            # UI components
├── models/               # AI models and core logic
├── services/            # Business logic
└── utils/              # Utility functions

Component Details

• main.py: Application initialization and configuration
• app.py: Streamlit interface and user interaction handling
• code_assistant.py: Core AI processing and task management
• task_service.py: Task-specific business logic
• session_state.py: Application state management

Task Specifications

Code Generation

Input: Natural language description
Output: Production-ready Python code
Features:

• Error handling implementation
• Type hint integration
• Documentation generation
• PEP standard compliance
• Performance optimization

Code Review

Analysis Components:

• Code quality assessment
• Bug identification
• Performance analysis
• Security evaluation
• Optimization recommendations

Documentation Generation

Output Components:

• System overview
• Implementation details
• API documentation
• Usage examples
• Parameter specifications

Test Generation

Capabilities:

• Pytest framework integration
• Edge case coverage
• Error scenario testing
• Assertion implementation
• Test documentation

Query Processing

Features:

• Context-aware responses
• Source citation
• Code comprehension
• Implementation guidance
• Best practice recommendations

Security Scanner

Scan Options

• Quick Scan: Basic security checks
• Deep Scan: Comprehensive analysis
• Custom Scan: Configure specific checks

Input Methods

• Upload Files: Scan uploaded code files
• Scan Code Snippet: Direct code input for scanning
• Git Repository: Scan code from a Git repository with authentication support

Scan Results

• Detailed vulnerability reports with severity filters
• Export options in JSON or PDF formats
• Historical scan data with trend analysis

Configuration

Environment Variables

GROQ_API_KEY          # Required for LLaMA 3
OPENAI_API_KEY        # Optional for GPT-4
LLAMA_CLOUD_API_KEY   # Optional for cloud services

Model Configuration

• Embedding Model: BAAI/bge-small-en-v1.5
• Vector Store: LlamaIndex implementation
• Node Parser: SentenceSplitter configuration

Usage Guidelines

Task Selection

1. Choose task type from available options
2. Input task-specific requirements
3. Review generated output
4. Access saved files in output directory

Document Management

1. Upload relevant files via UI
2. System automatically indexes content
3. Access indexed content via queries
4. Refresh index for updates

Development

Extension Points

• Model integration interface
• Task type implementation
• UI component development
• Service layer modification

Best Practices

• Follow PEP standards
• Implement comprehensive error handling
• Maintain type hints
• Document new features
• Include unit tests