from __future__ import annotations
from typing import TYPE_CHECKING, Any
import numpy as np
import numpy.typing as npt
from numba import njit # type: ignore [attr-defined]
from numba.typed import List
from py3dtiles.exceptions import TilerException
from py3dtiles.utils import SubdivisionType, aabb_size_to_subdivision_type
from .distance import is_point_far_enough, xyz_to_key
if TYPE_CHECKING:
from .node import Node
@njit(fastmath=True, cache=True) # type: ignore [misc]
def _insert(
cells_xyz: List[npt.NDArray[np.float32]],
cells_rgb: List[npt.NDArray[np.uint8]],
cells_classification: List[npt.NDArray[np.uint8]],
cells_intensity: List[npt.NDArray[np.uint8]],
aabmin: npt.NDArray[np.float32],
inv_aabb_size: npt.NDArray[np.float32],
cell_count: npt.NDArray[np.int32],
xyz: npt.NDArray[np.float32],
rgb: npt.NDArray[np.uint8],
classification: npt.NDArray[np.uint8],
intensity: npt.NDArray[np.uint8],
spacing: float,
shift: int,
force: bool = False,
) -> tuple[
npt.NDArray[np.float32],
npt.NDArray[np.uint8],
npt.NDArray[np.uint8],
npt.NDArray[np.uint8],
bool,
] | None:
keys = xyz_to_key(xyz, cell_count, aabmin, inv_aabb_size, shift)
if force:
# allocate this one once and for all
for k in np.unique(keys):
idx = np.where(keys - k == 0)
cells_xyz[k] = np.concatenate((cells_xyz[k], xyz[idx]))
cells_rgb[k] = np.concatenate((cells_rgb[k], rgb[idx]))
cells_classification[k] = np.concatenate(
(cells_classification[k], classification[idx])
)
cells_intensity[k] = np.concatenate((cells_intensity[k], intensity[idx]))
return None
else:
notinserted = np.full(len(xyz), False)
needs_balance = False
for i in range(len(xyz)):
k = keys[i]
if cells_xyz[k].shape[0] == 0 or is_point_far_enough(
cells_xyz[k], xyz[i], spacing
):
cells_xyz[k] = np.concatenate((cells_xyz[k], xyz[i].reshape(1, 3)))
cells_rgb[k] = np.concatenate((cells_rgb[k], rgb[i].reshape(1, 3)))
cells_classification[k] = np.concatenate(
(cells_classification[k], classification[i].reshape(1, 1))
)
cells_intensity[k] = np.concatenate(
(cells_intensity[k], intensity[i].reshape(1, 1))
)
if cell_count[0] < 8:
needs_balance = needs_balance or cells_xyz[k].shape[0] > 200000
else:
notinserted[i] = True
return (
xyz[notinserted],
rgb[notinserted],
classification[notinserted],
intensity[notinserted],
needs_balance,
)
[docs]
class Grid:
"""docstring for Grid"""
__slots__ = (
"cell_count",
"cells_xyz",
"cells_rgb",
"cells_classification",
"cells_intensity",
"spacing",
)
def __init__(self, node: Node, initial_count: int = 3) -> None:
self.cell_count = np.array(
[initial_count, initial_count, initial_count], dtype=np.int32
)
self.spacing = node.spacing * node.spacing
self.cells_xyz = List()
self.cells_rgb = List()
self.cells_classification = List()
self.cells_intensity = List()
for _ in range(self.max_key_value):
self.cells_xyz.append(np.zeros((0, 3), dtype=np.float32))
self.cells_rgb.append(np.zeros((0, 3), dtype=np.uint8))
self.cells_classification.append(np.zeros((0, 1), dtype=np.uint8))
self.cells_intensity.append(np.zeros((0, 1), dtype=np.uint8))
def __getstate__(self) -> dict[str, Any]:
return {
"cell_count": self.cell_count,
"spacing": self.spacing,
"cells_xyz": list(self.cells_xyz),
"cells_rgb": list(self.cells_rgb),
"cells_classification": list(self.cells_classification),
"cells_intensity": list(self.cells_intensity),
}
def __setstate__(self, state: dict[str, Any]) -> None:
self.cell_count = state["cell_count"]
self.spacing = state["spacing"]
self.cells_xyz = List(state["cells_xyz"])
self.cells_rgb = List(state["cells_rgb"])
self.cells_classification = List(state["cells_classification"])
self.cells_intensity = List(state["cells_intensity"])
@property
def max_key_value(self) -> int:
return 1 << (
2 * int(self.cell_count[0]).bit_length()
+ int(self.cell_count[2]).bit_length()
)
[docs]
def insert(
self,
aabmin: npt.NDArray[np.float32],
inv_aabb_size: npt.NDArray[np.float32],
xyz: npt.NDArray[np.float32],
rgb: npt.NDArray[np.uint8],
classification: npt.NDArray[np.uint8],
intensity: npt.NDArray[np.uint8],
force: bool = False,
) -> tuple[
npt.NDArray[np.float32],
npt.NDArray[np.uint8],
npt.NDArray[np.uint8],
npt.NDArray[np.uint8],
bool,
]:
return _insert( # type: ignore [no-any-return]
self.cells_xyz,
self.cells_rgb,
self.cells_classification,
self.cells_intensity,
aabmin,
inv_aabb_size,
self.cell_count,
xyz,
rgb,
classification,
intensity,
self.spacing,
int(self.cell_count[0] - 1).bit_length(),
force,
)
[docs]
def needs_balance(self) -> bool:
if self.cell_count[0] < 8:
for cell in self.cells_xyz:
if cell.shape[0] > 100000:
return True
return False
[docs]
def balance(
self,
aabb_size: npt.NDArray[np.float32],
aabmin: npt.NDArray[np.float32],
inv_aabb_size: npt.NDArray[np.float32],
) -> None:
t = aabb_size_to_subdivision_type(aabb_size)
self.cell_count[0] += 1
self.cell_count[1] += 1
if t != SubdivisionType.QUADTREE:
self.cell_count[2] += 1
if self.cell_count[0] > 8:
raise TilerException(
f"The first value of the attribute cell count must be lower or equal to 8, "
f"currently, it is {self.cell_count[0]}"
)
old_cells_xyz = self.cells_xyz
old_cells_rgb = self.cells_rgb
old_cells_classification = self.cells_classification
old_cells_intensity = self.cells_intensity
self.cells_xyz = List()
self.cells_rgb = List()
self.cells_classification = List()
self.cells_intensity = List()
for _ in range(self.max_key_value):
self.cells_xyz.append(np.zeros((0, 3), dtype=np.float32))
self.cells_rgb.append(np.zeros((0, 3), dtype=np.uint8))
self.cells_classification.append(np.zeros((0, 1), dtype=np.uint8))
self.cells_intensity.append(np.zeros((0, 1), dtype=np.uint8))
for cellxyz, cellrgb, cellclassification, cellintensity in zip(
old_cells_xyz, old_cells_rgb, old_cells_classification, old_cells_intensity
):
self.insert(
aabmin,
inv_aabb_size,
cellxyz,
cellrgb,
cellclassification,
cellintensity,
True,
)
[docs]
def get_points(
self, include_rgb: bool, include_classification: bool, include_intensity: bool
) -> npt.NDArray[np.uint8]:
xyz = []
rgb = []
classification = []
intensity = []
for i in range(len(self.cells_xyz)):
xyz.append(self.cells_xyz[i].view(np.uint8).ravel())
if include_rgb:
rgb.append(self.cells_rgb[i].ravel())
if include_classification:
classification.append(self.cells_classification[i].ravel())
if include_intensity:
intensity.append(self.cells_intensity[i].ravel())
return np.concatenate((*xyz, *rgb, *classification, *intensity))
[docs]
def get_point_count(self) -> int:
pt = 0
for i in range(len(self.cells_xyz)):
pt += self.cells_xyz[i].shape[0]
return pt