#!/usr/bin/env python
#
# pyLOM - Python Low Order Modeling.
#
# Plotting utilities.
#
# Last rev: 27/10/2021
from __future__ import print_function, division
import numpy as np
import matplotlib, matplotlib.pyplot as plt
from ..vmmath import vector_norm
from ..utils import cr_nvtx as cr, gpu_to_cpu, raiseWarning
DEFAULTSTYLE = {
'font' : {
'size' : 16,
'weight' : 'normal'
},
'legend' : {
'fontsize' : 12
},
'text' : {
'usetex' : False
},
'axes' : {
'linewidth' : 2.75
},
'savefig' : {
'bbox' : 'tight'
}
}
[docs]
def show_plots():
'''
Wrapper to matplotlib.pyplot.show()
'''
plt.show()
[docs]
def close_plots():
'''
Wrapper to matplotlib.pyplot.close()
'''
plt.close()
[docs]
def style_plots(styleDict=DEFAULTSTYLE):
'''
Define a common plot style in the scripts
'''
for key in styleDict.keys():
matplotlib.rc(key,**styleDict[key])
def plotFieldStruct2D(ax,nx,ny,ndim,xyz,field,dim,cmap,clear=False):
'''
Plot a 2D point or cell centered field on an structured mesh
'''
# Clear the axis if needed
if clear: ax.clear()
# Obtain the colormap if needed
if cmap is None: cmap = plt.get_cmap('coolwarm',256)
# Obtain X, Y matrices
X = xyz[:,0].reshape((nx,ny),order='c').T
Y = xyz[:,1].reshape((nx,ny),order='c').T
# Obtain data matrix
Z = field.reshape((nx,ny,ndim),order='c').T if ndim > 1 else field.reshape((nx,ny),order='c').T
levels = np.linspace(-1e-2, 1e-2, 11)
return ax.contourf(X,Y,Z[ndim,:,:] if dim >= 0 else np.linalg.norm(Z,axis=0) if ndim > 1 else Z,cmap=cmap)
[docs]
def plotResidual(S,fig=None,ax=None):
'''
Given the S matrix as a vector, plot
the accumulated residual.
'''
S = gpu_to_cpu(S)
# Build figure and axes
if fig is None:
fig = plt.figure(figsize=(8,6),dpi=100)
if ax is None:
ax = fig.add_subplot(1,1,1)
# Compute accumulated residual
accumulative_S = np.array([vector_norm(S,ii) for ii in range(S.shape[0])],np.double)
accumulative_S /= accumulative_S[0]
# Plot accumulated residual
ax.semilogy(np.arange(1,S.shape[0]+1),accumulative_S,'bo')
# Set labels
ax.set_ylabel(r'$\varepsilon_1$')
ax.set_xlabel(r'Truncation size')
ax.set_title(r'Tolerance')
# Return
return fig, ax
[docs]
def plotModalErrorBars(error:np.ndarray):
'''
Do a barplot of a 1D array of errors, where each element is the error associated in the prediction of a mode.
'''
indices = np.arange(len(error))+1
cmap = plt.cm.jet
colors = cmap(np.linspace(0.1, 0.9, len(error)))
fig, ax = plt.subplots(figsize=(20, 3))
bars = ax.bar(indices, error, capsize=5, color=colors, edgecolor='black')
ax.set_xlabel("Rank", fontsize=14)
ax.set_ylabel("Average Relative Error", fontsize=14)
ax.set_xticks(indices[24::25])
ax.set_xticklabels([f"{i}" for i in indices[24::25]], fontsize=12)
ax.tick_params(axis='both', labelsize=12)
fig.tight_layout()
return fig, ax
[docs]
def plotTimeSeries(time:np.ndarray, truth:np.ndarray, pred:np.ndarray, std:np.ndarray = None):
'''
Function to plot the comparison between the truth and predicted N temporal series.
'''
N, nt = truth.shape
std = std if std is not None else np.zeros((N,nt))
fig, axs = plt.subplots(N,1, figsize=(20, 3*N))
axs = axs.flatten()
for rr in range(len(axs)):
if rr == 0:
axs[rr].plot(time, pred[rr], 'r-.', label='Prediction')
axs[rr].plot(time, truth[rr], 'b--', label='Truth')
else:
axs[rr].plot(time, pred[rr], 'r-.')
axs[rr].plot(time, truth[rr], 'b--')
axs[rr].fill_between(time, pred[rr] - 1.96*std[rr], pred[rr] + 1.96*std[rr], color='r', alpha=0.25) #95% confidence interval of a Gaussian distribution
axs[rr].set_ylabel('Mode %i' % rr)
fig.legend()
fig.tight_layout()
return fig, axs
try:
import pyvista as pv
def _cells_and_offsets(conec):
'''
Build the offsets and cells array to create am
UnstructuredGrid
'''
# Compute points per cell
ppcell = np.sum(conec >= 0,axis=1)
# Compute cells for pyVista, with the number of points per cell on front
cells = np.c_[ppcell,conec]
# Now we get rid of any -1 entries for mixed meshes
cellsf = cells.flatten('c')
cellsf = cellsf[cellsf>=0]
# Now build the offsets vector
offset = np.zeros((ppcell.shape[0]+1,),np.int32)
offset[1:] = np.cumsum(ppcell)
return cellsf, offset
@cr('plots.pyvista_snap')
def plotSnapshot(mesh,dset,vars=[],idim=0,instant=0,**kwargs):
'''
Plot using pyVista
'''
# First create the unstructured grid
cells, offsets = _cells_and_offsets(mesh.connectivity)
# Create the unstructured grid
ugrid = pv.UnstructuredGrid(offsets,cells,mesh.eltype2VTK,mesh.xyz) if pv.vtk_version_info < (9,) else pv.UnstructuredGrid(cells,mesh.eltype2VTK,mesh.xyz)
# Load the variables inside the unstructured grid
for v in vars:
info = dset.info(v)
sliced = tuple([np.s_[:]] + [0 if i != idim else instant for i in range(len(dset[v].shape)-1)])
if info['point']:
ugrid.point_data[v] = mesh.reshape_var(dset[v][sliced],info)
else:
ugrid.cell_data[v] = mesh.reshape_var(dset[v][sliced],info)
# Launch plot
return ugrid.plot(**kwargs)
@cr('plots.pyvista_layout')
def plotLayout(mesh,dset,nrows,ncols,imode,vars=[],cmap='jet',title='',off_screen=False,**kwargs):
'''
Plot using pyVista
'''
# First create the unstructured grid
cells, offsets = _cells_and_offsets(mesh.connectivity)
# Create the unstructured grid
ugrid = pv.UnstructuredGrid(offsets,cells,mesh.eltype2VTK,mesh.xyz) if pv.vtk_version_info < (9,) else pv.UnstructuredGrid(cells,mesh.eltype2VTK,mesh.xyz)
# Load the variables inside the unstructured grid
plotter = pv.Plotter(shape=(nrows,ncols),off_screen=off_screen)
irow, icol = 0, 0
for ivar, v in enumerate(vars):
# Add variable to mesh
info = dset.info(v)
if info['point']:
ugrid.point_data[v] = mesh.reshape_var(dset[v],info)
else:
ugrid.cell_data[v] = mesh.reshape_var(dset[v],info)
# Plot
plotter.subplot(irow,icol)
if ivar == 0: plotter.add_text(title)
plotter.add_mesh(ugrid,scalars=v,cmap=cmap,component=imode,copy_mesh=True)
# Check irow, icol bounds
icol += 1
if icol == ncols:
icol = 0
irow += 1
# Launch plot
return plotter.show(**kwargs)
except:
[docs]
def plotSnapshot(mesh,dset,vars=[],idim=0,instant=0,**kwargs):
'''
Plot using pyVista
'''
raiseWarning('Import - Problems loading pyVista!',False)
def plotLayout(mesh,dset,nrows,ncols,imode,vars=[],cmap='jet',title='',off_screen=False,**kwargs):
'''
Plot using pyVista
'''
raiseWarning('Import - Problems loading pyVista!',False)
