Skip to content

CLAUDE.md

Latest commit

 

History

History
129 lines (93 loc) · 4.29 KB

CLAUDE.md

File metadata and controls

129 lines (93 loc) · 4.29 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

Python Machine Learning (ML) Plot is a library providing reusable functions for machine learning visualization plotting. The primary focus is on classification model visualization with decision boundaries and classified regions.

Architecture

Module Structure

  • src/opengood/py_ml_plot/ - Main package implementing ML plotting utilities
    • classification_plot.py - Contains setup_classification_plot() function for creating 2D classification visualizations
    • __init__.py - Exports public API (setup_classification_plot)

Classification Plot Architecture

The setup_classification_plot() function visualizes classification models through several key steps:

  1. Meshgrid Generation: Creates a grid of points across the feature space using configurable padding and step sizes for each axis
  2. Feature Scaling (optional): Applies inverse transformations via feature_scale lambda to handle models trained on scaled data
  3. Prediction: Uses the predict lambda to generate class predictions across the entire meshgrid
  4. Visualization: Combines filled contour plots (decision regions) with scatter plots (actual data points)

The function is designed to work with any scikit-learn classifier by accepting prediction and feature scaling logic as lambda functions.

Development Commands

Environment Setup

# Create and activate virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install dependencies
python3 -m pip install matplotlib numpy pandas scikit-learn

Testing

# Run all tests with coverage
python -m pytest tests/

# Run a specific test
python -m pytest tests/py_ml_plot/test_classification_plot.py::TestClassificationPlot::test_logistic_regression_setup_classification_plot_with_shaded_regions

# Run tests with verbose output
python -m pytest tests/ -v

# Run tests with coverage report
python -m pytest tests/ --cov=src.opengood.py_ml_plot --cov-report=term-missing

Note: Some tests are marked with @skip decorator for long-running tests (e.g., K-NN visualization). These are intended for local verification only.

Test Configuration

Test configuration is in pyproject.toml under [tool.pytest.ini_options]:

  • Coverage threshold: 70%
  • Test markers: slow, smoke, unit
  • Coverage report outputs to coverage.xml

Project Structure

src/opengood/py_ml_plot/
  - classification_plot.py  # Core plotting function
  - __init__.py             # Public API exports
tests/
  - py_ml_plot/
    - test_classification_plot.py  # Tests for classification plotting
  - resources/
    - data.csv                     # Test dataset

Key Implementation Details

Meshgrid Parameter Structure

The meshgrid parameter controls the visualization bounds and resolution:

meshgrid = {
    0: {"min": 10, "max": 10, "step": 0.25},    # x-axis padding and step
    1: {"min": 1000, "max": 1000, "step": 0.25}  # y-axis padding and step
}
  • Keys 0 and 1 represent x-axis and y-axis respectively
  • min/max define padding subtracted/added from data min/max values
  • step controls the density of the meshgrid (smaller = higher resolution but slower)

Lambda Function Patterns

Feature Scaling Lambda:

feature_scale = lambda x_set, y_set: (sc.inverse_transform(x_set), y_set)
  • Inverts feature scaling applied during training to show original data ranges
  • Returns tuple of (transformed_x, original_y)

Prediction Lambda:

predict = lambda x1, x2: classifier.predict(
    sc.transform(np.array([x1.ravel(), x2.ravel()]).T)
).reshape(x1.shape)
  • Takes meshgrid coordinates (x1, x2)
  • Applies feature scaling if the classifier was trained on scaled data
  • Uses ravel() to flatten, combines into 2D array, transforms, predicts, then reshapes

Package Management

  • Build system: setuptools
  • Python version: >=3.12
  • Package name: opengood.py_ml_plot
  • Current version tracked in pyproject.toml
  • Version bumping configured in .bumpversion.toml

Dependencies

Core runtime dependencies:

  • matplotlib >= 3.10.3 (visualization)
  • numpy >= 2.3.0rc1 (numerical operations)
  • pandas >= 2.2.3 (data handling)
  • scikit-learn >= 1.7.0rc1 (ML models in examples/tests)