Hyphae Network

Math essence

For each tip i and attractor a_j with |p_i - a_j| < d_attract, accumulate diff = (a_j - p_i). The new direction d = normalize(sum diffs). New tip p_i' = p_i + d * L_segment. Attractor a_j is removed when |p_i - a_j| < d_kill. With probability p_branch a tip spawns two rotated clones, yielding the rhizome-like hierarchy.

Parameters

Reference

First 60 lines of reference.py. Full source: library_external/nervous_system_hansmeyer/HyphaeNetwork/reference.py

from __future__ import annotations

from typing import ClassVar

import imageio.v2 as imageio
import numpy as np


class HyphaeNetwork:
    family: ClassVar[str] = "nervous_system_hansmeyer"
    sketch: ClassVar[str] = "HyphaeNetwork"

    def __init__(
        self,
        dim: int = 600,
        n_attractors: int = 600,
        attractor_radius: float = 0.42,
        n_roots: int = 4,
        root_radius: float = 0.05,
        segment_length: float = 0.0085,
        attract_distance: float = 0.18,
        kill_distance: float = 0.012,
        branch_prob: float = 0.10,
        branch_angle: float = 0.65,
        thickness_growth: float = 0.018,
        max_tips: int = 1200,
        max_steps: int = 240,
        bg_color: tuple[int, int, int] = (12, 14, 22),
        vein_color: tuple[int, int, int] = (200, 230, 220),
        core_color: tuple[int, int, int] = (255, 200, 130),
        seed: int | None = None,
    ) -> None:
        self.dim = dim
        self.n_attractors = n_attractors
        self.attractor_radius = attractor_radius
        self.n_roots = n_roots
        self.root_radius = root_radius
        self.segment_length = segment_length
        self.attract_distance = attract_distance
        self.kill_distance = kill_distance
        self.branch_prob = branch_prob
        self.branch_angle = branch_angle
        self.thickness_growth = thickness_growth
        self.max_tips = max_tips
        self.max_steps = max_steps
        self.bg_color = np.array(bg_color, dtype=np.uint8)
        self.vein_color = np.array(vein_color, dtype=np.uint8)
        self.core_color = np.array(core_color, dtype=np.uint8)
        self._seed = seed
        self.rng = np.random.default_rng(seed)
        self.canvas = np.full((dim, dim, 3), self.bg_color, dtype=np.uint8)
        self._stage = 0
        self.time = 0
        self._nodes: list[np.ndarray] = []
        self._parents: list[int] = []
        self._thickness: list[float] = []
        self._tips: list[int] = []
        self._attractors: np.ndarray = np.zeros((0, 2), dtype=np.float32)
        self.reset()