import numpy as np
from stereo.plots.decorator import plot_scale
from stereo.plots.plot_base import PlotBase
from stereo.plots.plot_vec.vec import Vec
class PlotVec(PlotBase):
[docs] def plot_vec(
self,
x_raw,
y_raw,
ty_raw,
ptime,
type='vec',
count_thresh=0,
tick_step=4000,
line_len_co=1,
vec_alpha=1,
line_width=0.0025,
density=2,
background=None,
background_alpha=0.5,
num_pix=50,
filter_type='gauss',
sigma_val=0.4,
radius_val=1,
scatter_s=1,
seed_val=1,
num_legend_per_col=12,
dpi_val=1000
):
"""
Plot vectors or streams of pseudo-time.
: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 ty_raw: 1d NdArray, involving cell types of all cells, or bin sets, can take the format of either string, int, or float
:param ptime: Array of pseudo-time, suggested being calculated by StereoPy dpt process
:param type: 'vec' or 'vector' plots vector plots, 'stream' or 'streamplot' plots stream plots
:param count_thresh: threshold of counts when filtering spots to plot
:param tick_step: step between tick labels
:param line_len_co: length coeeficient of vectors
:param vec_alpha: transparency of vectors, 0-1
:param line_width: width of vectors
:param density: density of streams in stream plot
:param background: 'field' plots fields-like background, with pixel color representing cell types, while 'scatter', 'cell', 'bin', or 'spot' plots each spot as a scatter
:param background_alpha: transparency of background
:param num_pix: number of pixel on shorter axis (x or y) when plotting background as fields
:param filter_type: type of kernel when smoothing vectors, if type is 'vec' or 'vector'. Pass 'gauss' to use Gaussian kernel, pass 'mean' to use Mean kernel.
:param sigma_val: sigma of kernel if passing 'gauss' to filter_type
:param radius_val: half of width of kernel array, if passing 'mean' to filter_type
:param scatter_s: size of scatter, if passing 'scatter', 'cell', 'bin', or 'spot' to background
:param seed_val: seed value to assign colors for different cell types when plotting background
:param num_legend_per_col: number of lines per column in legend
:param dpi_val: dpi value of figure
"""
assert len(x_raw.shape) == 1 and x_raw.shape[0] == y_raw.shape[0] == ptime.shape[0], \
"input has wrong array shape"
# 预处理，获得矩阵数据需要的输入
vec = Vec()
x_raw, y_raw, ty_raw, ptime = vec.filter_minority(ty_raw, count_thresh, x_raw, y_raw, ty_raw, ptime)
vec.preprocess(x_raw, y_raw, num_pix)
# print('preprocessed')
# 生成画图用的矩阵数据
plt_avg_ptime = vec.gen_arr_for_mean(ptime)
plt_avg_ptime_fil = vec.filter(plt_avg_ptime, filter_type, sigma_val, radius_val)
plt_common_ty = vec.gen_arr_for_common(ty_raw)
u, v = vec.cal_param(plt_avg_ptime_fil)
mask_nan = np.isnan(u) | np.isnan(v) | (u == 0) | (v == 0)
u[mask_nan] = np.nan
v[mask_nan] = np.nan
return vec.plot_line(x_raw, y_raw, ty_raw, plt_common_ty, u, v,
type, background, background_alpha, scatter_s,
seed_val, num_legend_per_col,
line_len_co, vec_alpha, line_width, density,
tick_step, dpi_val)
[docs] @plot_scale
def plot_time_scatter(
self,
group='leiden',
vmin: float = None,
vmax: float = None,
palette: str = 'stereo',
**kwargs
):
"""
Spatial distribution of pseudotime.
:param group: The key to get clustering result, now it will be ignored, defaults to 'leiden'.
:param vmin: Define the data range that the colormap covers, defaults to None.
:param vmax: Define the data range that the colormap covers, defaults to None.
:param palette: colormap, defaults to 'stereo'.
"""
data = self.stereo_exp_data
data.cells[group] = data.cells[group].astype('category')
ptime = data.tl.result['dpt_pseudotime']
from ..scatter import multi_scatter
fig = multi_scatter(
x=data.position[:, 0],
y=data.position[:, 1],
hue=[ptime],
x_label=['spatial1'],
y_label=['spatial2'],
title=['dpt_pseudotime'],
color_bar=True,
width=None,
height=None,
palette=palette,
vmin=vmin,
vmax=vmax,
**kwargs
)
return fig