import numpy as np
from .vec import Vec
from ..plot_base import PlotBase
class PlotVec3D(PlotBase):
[docs] def plot_vec_3d(
self,
x_raw,
y_raw,
z_raw,
ptime,
num_cell=20,
fil_type='gauss',
sigma_val=1,
radius_val=3,
):
"""
Plot vector field of cell trajectories embodied in 3D space.
:param x_raw: Array of x coordinates, taken from the first column in adata spatial axis array
:param y_raw: Array of y coordinates, taken from the second column in adata spatial axis array
:param z_raw: Array of z coordinates, taken from the third column in adata spatial axis array, or other
equivalent methods
:param ptime: Array of pseudo-time, suggested being calculated by StereoPy dpt process
:param num_cell: Number of cells along the shortest coordinates among x, y and z. Number of cells among other
coordinates will be calculated proportionally to their respective axis length
:param fil_type: Filter type when smoothing the ptime voxelized result. Allowed values are:
'gaussian': using Gaussian filter
'mean': using Mean filter
:param sigma_val: Sigma of filter when smoothing the voxelized ptime result using Gaussian filter, by passing
'gaussian' to fil_type. Use this website to help understand kernel results when assigning different
sigma values: http://demofox.org/gauss.html. Value neglected when passing 'mean' to fil_type
:param radius_val: Size of 'extended radius' of kernel, when smoothing the voxelized ptime result using Mean
filter, by passing 'mean' to fil_type. The diameter of kernel d equals to (2 * radius_val + 1)
Value neglected when passing 'gaussian' to fil_type
:return:
""" # noqa
# todo: 1. assert
# 预处理，获得矩阵数据需要的输入
vec = Vec()
vec.preprocess(x_raw, y_raw, z_raw, num_pix=num_cell)
print('\npreprocessed')
# 生成画图用的矩阵数据
plt_avg_ptime = vec.gen_arr_for_mean(ptime)
print('\nplt_avg_ptime', plt_avg_ptime.shape)
plt_avg_ptime_fil = vec.filter(plt_avg_ptime, sigma_val=sigma_val, radius_val=radius_val, type_val=fil_type)
print('\nptime voxelized and smoothed')
print('plt_avg_ptime_fil', plt_avg_ptime_fil.shape)
ux, uy, uz = vec.cal_param(plt_avg_ptime_fil)
print('\nux, uy, uz calculated.')
print('ux', ux.shape)
print('uy', uy.shape)
print('uz', uz.shape)
x = np.linspace(0, ux.shape[0], ux.shape[0])
y = np.linspace(0, ux.shape[1], ux.shape[1])
z = np.linspace(0, ux.shape[2], ux.shape[2])
xm, ym, zm = np.meshgrid(y, x, z)
# todo：继续在这里写画图代码
return ux, uy, uz, xm, ym, zm