image_array_and_histogram

Utilities to convert images to NumPy arrays, compute grayscale histograms, and reconstruct images from arrays.

Version 1.1.0 introduces PEP-8 function names and fixes the historical axis ordering bug. Arrays are now always shaped (height, width). The old camelCase names are still available but deprecated.

Installation

$ pip install image-array-and-histogram

Functions (current API)

• get_image_array(image, ensure_grayscale=True) – Return a 2D uint8 NumPy array (height, width) from a PIL image. Converts to grayscale by default.
• get_hist(image_or_array, as_density=False) – Return a 256-length list of counts (or probabilities if as_density=True). Accepts either a PIL image or a NumPy/list array.
• array_to_image(arr, width=None, height=None) – Build a grayscale PIL image from a 1D or 2D array.

Deprecated aliases (will emit DeprecationWarning): getImageArray, getHist, getImageFromArray.

Quick Start

from PIL import Image
import numpy as np
from image_array_and_histogram import get_image_array, get_hist, array_to_image

# Load image and get array
img = Image.open('photo.jpg')
arr = get_image_array(img)  # shape (H, W)

# Compute histogram
hist = get_hist(arr)  # list of 256 counts

# Normalize histogram
hist_density = get_hist(arr, as_density=True)

# Create an image from a NumPy array
gradient = np.linspace(0, 255, 256, dtype=np.uint8).reshape(16, 16)
gradient_img = array_to_image(gradient)
gradient_img.save('gradient.png')

Notes

• If you pass a color image to get_image_array or get_hist, it will be converted to grayscale (mode 'L').
• Histogram computation is vectorized with NumPy (numpy.bincount) for speed.
• For legacy behavior (<=1.0.x) the array shape used (width, height). Adjust any downstream code if it relied on that ordering.

Testing

After cloning the repository:

pip install -e .[dev]
pytest -q

License

MIT