Getting Started

ktch allows you to conduct model-based morphometrics via scikit-learn compatible API efficiently.

Install

From PyPI or conda-forge

ktch is currently available on PyPI and conda-forge. You can install it with pip:

$ pip install ktch

or with conda:

$ conda install -c conda-forge ktch

Quick start

This example loads the mosquito wing outline dataset (Rohlf and Archie 1984), and calculates elliptic Fourier descriptors (EFDs). Finally, it performs principal component analysis (PCA) on the EFDs.

from sklearn.decomposition import PCA

from ktch.datasets import load_outline_mosquito_wings
from ktch.outline import EllipticFourierAnalysis

# load data
data_outline_mosquito_wings = load_outline_mosquito_wings()
X = data_outline_mosquito_wings.coords.to_numpy().reshape(-1,100,2)

# EFD
efa = EllipticFourierAnalysis(n_components=20)
coef = efa.transform(X)

# PCA
pca = PCA(n_components=3)
pcscores = pca.fit_transform(coef)