from __future__ import annotations
from dataclasses import dataclass
from enum import IntEnum
from typing import Any, Dict, Optional
[docs]
class PixelCategory(IntEnum):
INACTIVE = 0
CHARGE_COLLECTION = 1
CHARGE_NEIGHBOR = 2
INDUCTION_ONLY = 3
[docs]
@dataclass
class PixelClassificationResult:
"""Classification output with pixel IDs for each category.
Attributes:
charge_pixels: Pixel IDs for CHARGE_COLLECTION pixels
neighbor_pixels: Pixel IDs for CHARGE_NEIGHBOR pixels
induction_pixels: Pixel IDs for INDUCTION_ONLY pixels
induction_pixels_x: X coordinates (cm) for INDUCTION_ONLY pixels
induction_pixels_y: Y coordinates (cm) for INDUCTION_ONLY pixels
Note:
- "charge_pixels" means pixels in the CHARGE_COLLECTION category
- The \*_pixels fields contain global pixel IDs (not indices)
- Types are Optional[Any] to avoid hard dependency on CuPy at import time.
"""
charge_pixels: Optional[Any] # (n_charge,) pixel IDs in CHARGE_COLLECTION category
neighbor_pixels: Optional[Any] # (n_neighbor,) pixel IDs in CHARGE_NEIGHBOR category
induction_pixels: Optional[Any] # (n_induction,) pixel IDs in INDUCTION_ONLY category
induction_pixels_x: Optional[Any] = None # (n_induction,) x-coordinates in cm
induction_pixels_y: Optional[Any] = None # (n_induction,) y-coordinates in cm
[docs]
def summary(self, n_total_pixels: int):
n_charge = len(self.charge_pixels) if self.charge_pixels is not None else 0
n_neighbor = len(self.neighbor_pixels) if self.neighbor_pixels is not None else 0
n_induction = len(self.induction_pixels) if self.induction_pixels is not None else 0
n_active = n_charge + n_neighbor + n_induction
lines = [
"Pixel classification:",
f" Total pixels: {n_total_pixels}",
f" Charge collection: {n_charge} ({(100*n_charge/max(n_total_pixels,1)):.2f}%)",
f" Charge neighbors: {n_neighbor} ({(100*n_neighbor/max(n_total_pixels,1)):.2f}%)",
f" Induction only: {n_induction} ({(100*n_induction/max(n_total_pixels,1)):.2f}%)",
f" Active total: {n_active} ({(100*n_active/max(n_total_pixels,1)):.2f}%)",
]
return "\n".join(lines)
from . import voxelization
from . import pixel_classifier
__all__ = [
'PixelCategory',
'PixelClassificationResult',
'voxelization',
'pixel_classifier',
]
