How darkstarstrix/auto_api works
Main Components
a) Core AutoML Engine Layer:
• It to Base_model.py and Model_Library.py for model definitions
• Model_Types.py for type definitions and model categorization
• Link to Optimization_Suite.py for hyperparameter tuning
b) Model Layer:
• Decision Tree
• KMeans
• LightGBM
• Linear Models
• Naive Bayes
• Connection to model persistence (.pt files)
c) API/Interface Layer:
• Endpoints directory components:
• Main API (Main.py)
• Authentication (Auth.py)
• Dashboard (Dashboard.py)
• Monitoring (API_Monitoring.py)
• CLI Interface (CLI.py)
d) Training Infrastructure:
• Training Files component
• To Training Plots output
• Lightning Logs storage
• Link to Neptune.py for experiment tracking
Specific Implementation Details
[Component Layout Example]
┌─────────────────────────────────────────────┐
• API Layer │
• ┌─────────┐ ┌─────────┐ ┌────────────┐ │
• Main │ │ Auth │ │ Dashboard │ │
• ┘ └─────────┘ └────────────┘ │
• ┬─────────────────────────┘
┌───────────────────▼─────────────────────────┐
• AutoML Core Engine │
• ┌─────────────┐ ┌──────────────────┐ │
• Base Model │◄────►│Model Library │ │
• ┘ └──────────────────┘ │
• ▲ ▲ │
• ┌─────────────┐ ┌──────────────────┐ │
• Optimization │◄────►│Training Pipeline │ │
• ┘ └──────────────────┘ │
• ┬─────────────────────────┘
┌───────────────────▼─────────────────────────┐
• Model Layer │
• ┌─────────┐ ┌─────────┐ ┌─────────────┐ │
• Decision │ │ Linear │ │ LightGBM │ │
• Tree │ │ Models │ │ │ │
• ┘ └─────────┘ └─────────────┘ │
• ┘
Additional Elements
• A monitoring/logging section showing connection to Neptune.py
• Visualization pipeline connecting to Training_plots
• Configuration management flow
• External interfaces (CLI, API) entry points
Documentation References
• Link to Writerside documentation components
• Reference to configuration templates
• Connection to example notebooks
Remember to:
• Both synchronous and asynchronous interactions
• Scalability points and potential bottlenecks
• Version control and deployment aspects