Извлечение Цветовой Палитры из Изображения

# Extract dominant colors using colorthief
# pip install colorthief Pillow

from colorthief import ColorThief
from PIL import Image

def get_palette(image_path: str, num_colors: int = 6, quality: int = 1):
    """
    Extract dominant colors from an image.
    quality: 1=best (slow), 10=fast (less accurate)
    Returns list of (r, g, b) tuples.
    """
    ct = ColorThief(image_path)

    # Most dominant single color
    dominant = ct.get_color(quality=quality)
    print(f"Dominant color: RGB{dominant} -> #{dominant[0]:02X}{dominant[1]:02X}{dominant[2]:02X}")

    # Full palette
    palette = ct.get_palette(color_count=num_colors, quality=quality)
    print("\nPalette:")
    for i, (r, g, b) in enumerate(palette):
        hex_color = f"#{r:02X}{g:02X}{b:02X}"
        print(f"  {i+1}. RGB({r}, {g}, {b}) -> {hex_color}")

    return palette

# Alternative: k-means clustering with scikit-learn
# pip install scikit-learn numpy Pillow
import numpy as np
from sklearn.cluster import KMeans

def kmeans_palette(image_path: str, n_colors: int = 6) -> list:
    img = Image.open(image_path).convert("RGB")
    img = img.resize((150, 150))  # Downsample for speed
    pixels = np.array(img).reshape(-1, 3)

    kmeans = KMeans(n_clusters=n_colors, random_state=42, n_init="auto")
    kmeans.fit(pixels)

    centers = kmeans.cluster_centers_.astype(int)
    counts = np.bincount(kmeans.labels_)
    sorted_colors = centers[np.argsort(-counts)]  # Sort by frequency

    result = []
    for r, g, b in sorted_colors:
        hex_color = f"#{r:02X}{g:02X}{b:02X}"
        result.append({"rgb": (r, g, b), "hex": hex_color})
        print(f"RGB({r}, {g}, {b}) -> {hex_color}")

    return result

palette = get_palette("photo.jpg", num_colors=6)
print("\n--- k-means ---")
kmeans_palette("photo.jpg", n_colors=6)