Skip to content

Latest commit

 

History

History
55 lines (43 loc) · 1.52 KB

File metadata and controls

55 lines (43 loc) · 1.52 KB

AIris Architecture

System Overview

graph TD
    A[User Interface] --> B[Gradio Web App]
    B --> C[AIris Agent]
    C --> D[LLM Processing]
    C --> E[Image Analysis]
    
    D --> F[Groq API]
    D --> G[Offline Mode]
    
    E --> H[PIL Image Processing]
    
    C --> I[Knowledge Base]
    I --> J[Fashion Rules]
    I --> K[Style Guidelines]
    I --> L[Color Theory]
    
    B --> M[User Inputs]
    M --> N[Text Queries]
    M --> O[Image Upload]
    M --> P[Context Selection]
    P --> Q[Weather]
    P --> R[Occasion]
    P --> S[Style]
Loading

Component Description

Frontend Layer

  • Gradio Web App: Provides an intuitive interface for users to interact with AIris
  • User Inputs: Handles multiple input types (text, images, selections)

Processing Layer

  • AIris Agent: Core component that orchestrates all functionality
  • LLM Processing: Handles natural language understanding and generation
  • Image Analysis: Processes uploaded images for style analysis

Knowledge Layer

  • Fashion Rules: Core fashion guidelines and rules
  • Style Guidelines: Different style aesthetics and their characteristics
  • Color Theory: Color combination rules and recommendations

External Services

  • Groq API: For advanced language processing
  • Offline Mode: Fallback functionality when API is unavailable

Key Features

  1. Multi-modal input processing (text + images)
  2. Context-aware recommendations
  3. Flexible deployment (online/offline modes)
  4. Modular architecture for easy extensions