import numpy as np

import math

sourcecolor = np.array([4, 55, 242]) / 255

vectorcolor = np.array([54, 105, 242]) / 255

targetcolor = np.array([0, 255, 0]) / 255

constraintcolor = np.array([255, 0, 0]) / 255

defaultcolor = np.array([137,137,137]) / 255

bluecolor = np.array([54,94,137]) / 255

radius = 3.5

start_elev = math.radians(0)

end_elev = math.radians(30)

elev_samples = 2

start_azim = math.radians(0)

azim_samples = 8

end_azim = math.radians(360 - 360/azim_samples)

pointradius = 0.03

curveradius = 0.015

imagenet_mean = [0.485, 0.456, 0.406]

imagenet_std = [0.229, 0.224, 0.225]

patch_sizes = {

'diff3f': 14,

'dino2': 14,

'dino': 14, # backward-compatible alias

'dino3': 16,

'clip': 14,

'radio': 16,

# Architecture-name aliases (so callers can key by actual model architecture/id)

'dinov2_vitg14_reg': 14,

'dinov3_vit7b16': 16,

}

# NOTE: Replace below with your path to the SAM2 checkpoint

SAM2_CHECKPOINT = "/net/projects/ranalab/guanzhi/DFD/sam2_repo/checkpoints/sam2.1_hiera_large.pt"