def init_1figure9axes(self):
"""
同时显示9张数据表：
instructions， L1D_R_misses ， L1D_W_misses ,
cycles , L1D_R_access , L1D_W_access ,
IPC , L1D_R_missRate, L1D_W_missRate
:return:
"""
# reset status.
plt.close('all')
# 注：matplot里面所有中文字符都得是：u"中文"的raw格式；
matplotlib.rcParams["axes.unicode_minus"] = False
# figure 3x3
self.fig = plt.figure(figsize=(8, 6), dpi=80)
self.grid = Grid(self.fig, rect=111, nrows_ncols=(3, 3),
axes_pad=0.25, label_mode='L',)
self.Instruction_ax = self.grid[0]
self.Cycle_ax = self.grid[3]
self.IPC_ax = self.grid[6]
self.L1D_R_misses_ax = self.grid[1]
self.L1D_R_access_ax = self.grid[4]
self.L1D_R_missRate_ax = self.grid[7]
self.L1D_W_misses_ax = self.grid[2]
self.L1D_W_access_ax = self.grid[5]
self.L1D_W_missRate_ax = self.grid[8]
# 竖轴
self.grid[0].set_ylabel(u'分子', fontproperties=self.myfont, fontsize=12)
self.grid[3].set_ylabel(u'分母', fontproperties=self.myfont, fontsize=12)
self.grid[6].set_ylabel(u'除积', fontproperties=self.myfont, fontsize=12)
# 横轴
self.grid[6].set_xlabel(u'IPC(Instructions/cycles)' , fontproperties=self.myfont, fontsize=12)
self.grid[7].set_xlabel(u'cache(miss/Read=miss rate)' , fontproperties=self.myfont, fontsize=12)
self.grid[8].set_xlabel(u'cache(Read/Write=miss rate)', fontproperties=self.myfont, fontsize=12)
# figure1 4x1
self.fig1 = plt.figure(figsize=(8, 6), dpi=80)
self.grid1 = Grid(self.fig1, rect=111, nrows_ncols=(4, 1),
axes_pad=0.25, label_mode='L', )
self.IPC_ax_1 = self.grid1[0]
self.L1D_R_missRate_ax_1 = self.grid1[1]
self.L1D_W_missRate_ax_1 = self.grid1[2]
self.Time_ax_1 = self.grid1[3]
self.grid1[0].set_ylabel(u'L1D_W_missRate', fontproperties=self.myfont, fontsize=12)
self.grid1[1].set_ylabel(u'L1D_R_missRate', fontproperties=self.myfont, fontsize=12)
self.grid1[2].set_ylabel(u'IPC', fontproperties=self.myfont, fontsize=12)
self.grid1[3].set_ylabel(u'time{sys/perf}', fontproperties=self.myfont, fontsize=12)
# 紧致layout
plt.tight_layout()
def test_grid_axes_lists():
"""Test Grid axes_all, axes_row and axes_column relationship."""
fig = plt.figure()
grid = Grid(fig, 111, (2, 3), direction="row")
assert_array_equal(grid, grid.axes_all)
assert_array_equal(grid.axes_row, np.transpose(grid.axes_column))
assert_array_equal(grid, np.ravel(grid.axes_row), "row")
grid = Grid(fig, 111, (2, 3), direction="column")
assert_array_equal(grid, np.ravel(grid.axes_column), "column")
def compare_figures(images,
nrows_ncols,
show_axis=False,
fig_size=(10, 8),
show_fig=True):
"""
Show three figures in a row, link their axes
:param img_1: image to show on top left
:param img_2: image to show at top right
:param img_3: image to show on bottom left
:param img_4: image to show on bottom right
:param show_axis: if False, axes will be hide
:param fig_size: size of the figure
:param show_fig: show figure or not
:param color_bar: if True, add color bar to the last plot
:return:
"""
from mpl_toolkits.axes_grid1 import Grid
fig = plt.figure(figsize=fig_size)
grid = Grid(fig,
rect=111,
nrows_ncols=nrows_ncols,
axes_pad=0.25,
label_mode='L',
share_all=True)
for i, (ax, img) in enumerate(zip(grid, images)):
ax.imshow(img)
if not show_axis:
ax.axis('off')
plt.tight_layout()
if show_fig:
plt.show()
def prepare_order_trace_plot(s_list, row_col=(3, 2)):
import matplotlib.pyplot as plt
row, col = row_col
n_ax = len(s_list)
n_f, n_remain = divmod(n_ax, row * col)
if n_remain:
n_ax_list = [row * col] * n_f + [n_remain]
else:
n_ax_list = [row * col] * n_f
from mpl_toolkits.axes_grid1 import Grid
i_ax = 0
fig_list = []
ax_list = []
for n_ax in n_ax_list:
fig = plt.figure()
fig_list.append(fig)
grid = Grid(fig, 111, (row, col), ngrids=n_ax, share_x=True)
sl = slice(i_ax, i_ax + n_ax)
for s, ax in zip(s_list[sl], grid):
ax_list.append(ax)
i_ax += n_ax
return fig_list, ax_list
def view_encoded_patches(img_ids,
files,
patchDir,
model_name,
img_dir,
cust_name=''):
if model_name == 'unet':
input_size = 572
else:
input_size = 321
fig = plt.figure(figsize=(11, 2))
grid = Grid(fig,
rect=111,
nrows_ncols=(1, 6),
axes_pad=0.1,
label_mode='L')
for plt_cnt, iid in enumerate(img_ids):
img = np.zeros((input_size, input_size, 3), dtype=np.uint8)
for cnt in range(3):
img[:, :, cnt] = imageio.imread(
os.path.join(patchDir,
files[iid][:-1] + '_RGB{}.jpg'.format(cnt)))
grid[plt_cnt].imshow(crop_center(img, 224, 224))
grid[plt_cnt].set_axis_off()
plt.tight_layout()
if not os.path.exists(os.path.join(img_dir, cust_name)):
os.makedirs(os.path.join(img_dir, cust_name))
plt.savefig(
os.path.join(
img_dir, cust_name,
'{}_{}.png'.format(model_name,
'_'.join([str(a) for a in img_ids]))))
plt.show()
def compare_figures(images,
nrows_ncols,
fig_size=(10, 8),
show_axis=False,
show_fig=True,
title_list=None):
"""
Show images in grid pattern, link their x and y axis
:param images: list of images to be displayed
:param nrows_ncols: a tuple of (n_h, n_w) where n_h is #elements/row and n_w is #elements/col
:param fig_size: a tuple of figure size
:param show_axis: if True, each subplot will have its axis shown
:param show_fig: if True, plt.show() will be called
:param title_list: list of title names to be displayed on each sub images
:return:
"""
from mpl_toolkits.axes_grid1 import Grid
if title_list:
assert len(title_list) == len(images)
fig = plt.figure(figsize=fig_size)
grid = Grid(fig,
rect=111,
nrows_ncols=nrows_ncols,
axes_pad=0.25,
label_mode='L',
share_all=True)
for i, (ax, img) in enumerate(zip(grid, images)):
ax.imshow(img)
if not show_axis:
ax.axis('off')
if title_list:
ax.set_title(title_list[i])
plt.tight_layout()
if show_fig:
plt.show()
def check_readout_pattern(fig, hdu_list, axis=1):
from mpl_toolkits.axes_grid1 import Grid
grid = Grid(fig,
111, (len(hdu_list), 1),
share_x=True,
share_y=True,
share_all=True,
label_mode="1")
smin, smax = np.inf, -np.inf
for hdu, ax in zip(hdu_list, grid):
p_horizontal = np.median(hdu.data, axis=axis)
p_horizontal0 = np.median(p_horizontal)
p_horizontal -= p_horizontal0
smin = min(smin, p_horizontal[100:-100].min())
smax = max(smax, p_horizontal[100:-100].max())
ax.plot(p_horizontal)
ax.axhline(0, color="0.5")
ax.locator_params(axis="y", nbins=5)
ax = grid[-1]
ax.set_xlim(0, len(p_horizontal))
sminmax = max(-smin, smax)
ax.set_ylim(-sminmax, sminmax)
if axis == 1:
ax.set_xlabel("y-pixel")
elif axis == 0:
ax.set_xlabel("x-pixel")
ax.set_ylabel("ADUs")
def images_four(imgs, save, title):
from mpl_toolkits.axes_grid1 import Grid
f = plt.figure(figsize=(4,4))
grid = Grid(
f, rect=111, nrows_ncols=(2,2),
axes_pad=0.0, label_mode='none',)
for ax, im in zip(grid, imgs):
ax.imshow(im)
ax.spines['top'].set_visible(False)
ax.spines['bottom'].set_visible(False)
ax.spines['left'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.yaxis.set_ticks_position('none')
ax.yaxis.set_ticklabels([])
ax.xaxis.set_ticks_position('none')
ax.xaxis.set_ticklabels([])
ax.axis('off')
f.text(.08, .52, 'AP', ha='left', va='center')
f.text(.95, .52, 'AN', ha='right', va='center')
f.text(.50, .95, 'distintivo', ha='center', va='top',)
f.text(.50, .08, 'comum', ha='center', va='bottom',)
f.text(.15, .90, 'S+', ha='left', va='top')
f.text(.85, .90, 'S-', ha='right', va='top')
f.text(.90, .15, 'S+', ha='right', va='bottom',)
f.text(.15, .15, 'S-', ha='left', va='bottom',)
if save:
return save_figure(title)
else:
plt.show()
def test_grid_axes_position(direction):
"""Test positioning of the axes in Grid."""
fig = plt.figure()
grid = Grid(fig, 111, (2, 2), direction=direction)
loc = [ax.get_axes_locator() for ax in np.ravel(grid.axes_row)]
assert loc[1]._nx > loc[0]._nx and loc[2]._ny < loc[0]._ny
assert loc[0]._nx == loc[2]._nx and loc[0]._ny == loc[1]._ny
assert loc[3]._nx == loc[1]._nx and loc[3]._ny == loc[2]._ny
def draw_snapshot(sites, bamfiles, color="black", min_y=30, Nsite=5):
import metaseq
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import Grid
Nsites_use = min(Nsite, len(sites))
# take read counts from samples
ip_signals = [None]*len(bamfiles)
ip_arrays = [None]*len(bamfiles)
for i, bamfile in zip(range(len(bamfiles)), bamfiles):
ip_signals[i] = metaseq.genomic_signal(bamfile, 'bam')
ip_arrays[i] = [None]*Nsites_use
for k in range(Nsites_use):
ip_arrays[i][k] = ip_signals[i].local_coverage(sites[k])
# draw figure
fig = plt.figure(figsize=(100, 20))
grid = Grid(fig, 142, nrows_ncols=(len(bamfiles), Nsites_use),
axes_pad=0.05, direction="row",
add_all=True, share_all=False, label_mode="all")
ymaxs = [None]*len(bamfiles)
for i in range(len(bamfiles)):
for k in range(Nsites_use):
grid[i*Nsites_use+k].bar(ip_arrays[i][k][0],
ip_arrays[i][k][1], color=color, edgecolor=color)
xmin, xmax, ymin, ymax = grid[i*Nsites_use+k].axis()
ymaxs[i] = max(ymaxs[i], ymax)
for i in range(len(bamfiles)):
for k in range(Nsites_use):
xmin, xmax, ymin, ymax = grid[i*Nsites_use+k].axis()
grid[i*Nsites_use+k].axis([xmin,xmax,ymin, ymaxs[i]])
grid[i*Nsites_use+k].get_xaxis().set_visible(False)
grid[i*Nsites_use+k].get_yaxis().set_visible(False)
grid[i*Nsites_use+k].annotate(
bamfiles[i].split('/')[-1].split('.')[0]
+" [0-"+str(ymaxs[i])+"]",
xy=(0,1), xytext=(10, -10),
va='top', xycoords='axes fraction',
textcoords='offset points',
fontsize=25)
if i==0:
chrom = str(sites[k]).split('\t')[0]
start = str(sites[k]).split('\t')[1]
end = str(sites[k]).split('\t')[2].split('\n')[0]
grid[i*Nsites_use+k].set_title(
"Location: " + chrom + " " + start + "-" + end
)
return fig
def GridPlot(self):
numPlot = len(self.lc)
tmp = array(self.tBins)
maxCounts = max(map(max, self.lc)) + 5
maxTime = tmp.max()
minTime = tmp.min()
Xtxt = maxTime * .6
Ytxt = maxCounts * .7
timeAxis = tmp[:, 1]
fig = plt.figure(self.fignum)
#pltNum = numPlot*100 +11
grid = Grid(fig,
111,
nrows_ncols=(numPlot, 1),
axes_pad=0,
direction='column')
for i in xrange(numPlot):
txtString = str(self.eBins[i][0]) + ' - ' + str(
self.eBins[i][1]) + ' keV'
pl, = grid[i].step(timeAxis, self.lc[i], color='k')
ax = pl.get_axes()
#ax.set_hatch('/')
#plt.yticks(rotation=-25)
ax.set_ylim(top=maxCounts)
#ax.set_yticklables(ax.get_yticks(),rotation=-25)
yloc = plt.MaxNLocator(10, prune='lower')
ax.yaxis.set_major_locator(yloc)
#ax.text(Xtxt,Ytxt,txtString) #uncomment later
ax.set_ylabel(r"counts s$^{-1}$")
ax.set_xlabel("Time [s]")
ax.set_xlim(right=maxTime, left=minTime)
#plt.xlabel('Time (s)')
#plt.ylabel('Counts')
self.vlineLims = [0, maxCounts]
self.f2 = fig
self.grid = grid
#fig.tight_layout()
if self.save:
plt.savefig(self.fname + "_grid.pdf")
def plot_corner_latent(z, lim=3, nbins=25, show_title=True):
"""
Make a corner plot of standard gaussian distributed latent variables.
Parameters
----------
z : latent variables, array of size (n_samples, latent_dim)
lim : int, optional
[description], by default 3
nbins : int, optional
[description], by default 25
show_title : bool, optional
[description], by default True
Example
-------
z = np.random.normal(size=(1000,8))
plot_corner_latent(z)
"""
import matplotlib as mpl
from mpl_toolkits.axes_grid1 import Grid
latent_dim = z.shape[1]
fig = plt.figure(figsize=(latent_dim * 2, latent_dim * 2))
grid = Grid(fig,
rect=111,
nrows_ncols=(latent_dim, latent_dim),
axes_pad=0.25,
label_mode='L',
share_y=False)
colors = mpl.cm.jet(np.linspace(0, 1, latent_dim))
bins = nbins
for i in range(latent_dim):
for j in range(latent_dim):
ax = grid[i * latent_dim + j]
if i == j:
n, _, _ = ax.hist(z[:, i],
bins=bins,
normed=True,
color=colors[i])
ax.set_yticks([])
if show_title:
ax.set_title('$z_{}$'.format(i))
if i > j:
ax.hist2d(z[:, j], z[:, i], bins=bins, cmap=mpl.cm.gray)
if i < j:
ax.axis('off')
plt.tight_layout()
plt.xlim(-lim, lim)
def plotFunnelsStd(proteins, decoys, decoys_scores, outputFile):
from matplotlib import pylab as plt
import numpy as np
fig = plt.figure(figsize=(12, 12))
N = len(proteins)
sqrt_n = int(np.sqrt(N))
if N == sqrt_n * sqrt_n:
nrows = int(np.sqrt(N))
ncols = int(N / nrows)
else:
nrows = int(np.sqrt(N)) + 1
ncols = int(N / nrows)
if nrows * ncols < N: ncols += 1
from mpl_toolkits.axes_grid1 import Grid
grid = Grid(fig,
rect=111,
nrows_ncols=(nrows, ncols),
axes_pad=0.25,
label_mode='L',
share_x=False,
share_y=False)
num_proteins = [(s, int(s[1:])) for s in proteins]
num_proteins = sorted(num_proteins, key=lambda x: x[1])
proteins, num_proteins = zip(*num_proteins)
for n, protein in enumerate(proteins):
tmscores = []
scores = []
stds = []
for decoy in decoys[protein]:
tmscores.append(decoy[1])
scores.append(decoys_scores[protein][decoy[0]][0])
stds.append(decoys_scores[protein][decoy[0]][1])
grid[n].errorbar(tmscores, scores, yerr=stds, fmt='.', ecolor='r')
plt.xlim(-0.1, max(tmscores) + 0.1)
plt.ylim(min(scores) - 1, max(scores) + 1)
grid[n].set_title(protein)
#plt.tight_layout()
# plt.savefig(outputFile)
# plt.tick_params(axis='both', which='minor', labelsize=8)
# plt.savefig(outputFile, format='png', dpi=600)
plt.tight_layout()
plt.savefig(outputFile, format='png', dpi=600)
def PlotLC(self):
fig = plt.figure(1)
grid = Grid(fig,111,nrows_ncols = (3,1), axes_pad=0.,direction='column')
Step(grid[0],self.tBins,self.swift,'r',1.)
Step(grid[1],self.tBins,self.nai,'b',1.)
Step(grid[2],self.tBins,self.bgo,'g',1.)
def __init__(self, fig=None, nrows_ncols=(8, 4)):
if fig is None:
import matplotlib.pyplot as plt
fig = plt.figure()
from mpl_toolkits.axes_grid1 import Grid
self.grid = Grid(
fig,
111, # similar to subplot(111)
nrows_ncols=nrows_ncols, # creates 2x2 grid of axes
axes_pad=0.1,
share_y=False,
direction="column") # pad between axes in inch.
def check_destriper(hdu, bias_mask, vmin, vmax):
destriper = Destriper()
hh1 = destriper.get_destriped(hdu.data, mask=bias_mask, pattern=2048)
hh2 = destriper.get_destriped(hdu.data, pattern=64, mask=bias_mask)
from matplotlib.colors import Normalize
from mpl_toolkits.axes_grid1 import ImageGrid, Grid
fig1 = plt.figure(figsize=(13, 5))
grid = ImageGrid(fig1,
111, (1, 3),
share_all=True,
label_mode="1",
axes_pad=0.1,
cbar_mode="single")
norm = Normalize(vmin=vmin, vmax=vmax)
for ax, dd in zip(grid, [hdu.data, hh1, hh2]):
im = ax.imshow(dd,
norm=norm,
origin="lower",
cmap="gray_r",
interpolation="none")
plt.colorbar(im, cax=grid[0].cax)
grid[0].set_xlabel("x-pixel")
grid[0].set_ylabel("y-pixel")
fig1.tight_layout()
fig2 = plt.figure()
grid = Grid(fig2,
111, (3, 1),
share_x=True,
share_y=True,
share_all=True,
label_mode="1",
axes_pad=0.1)
for ax, dd in zip(grid, [hdu.data, hh1, hh2]):
s = np.median(dd[:, 1024 - 128:1024 + 128], axis=1)
ax.plot(s)
ax.axhline(0, color="0.5")
grid[2].set_xlim(0, 2048)
grid[2].set_xlabel("y-pixel")
fig2.tight_layout()
return fig1, fig2
def show_features(self, shape, suptitle, count=-1):
maxw = np.amax(self.w.T)
minw = np.amin(self.w.T)
count = self._output_size if count == -1 or count > \
self._output_size else count
ncols = count if count < 14 else 14
nrows = count // ncols
nrows = nrows if nrows > 2 else 3
fig = plt.figure(figsize=(ncols, nrows), dpi=100)
grid = Grid(fig, rect=111, nrows_ncols=(nrows, ncols), axes_pad=0.01)
for i, ax in enumerate(grid):
x = self.w.T[i] if i < self._input_size else np.zeros(shape)
x = (x.reshape(1, -1) - minw) / maxw
ax.imshow(x.reshape(*shape), cmap="Greys")
ax.set_axis_off()
fig.text(0.5, 1, suptitle, fontsize=20, horizontalalignment='center')
fig.tight_layout()
plt.show()
return
def do_figure_stat_per_order(jo, calib, fig=None):
# smoothed_profiles = jo["smoothed_profiles"]
stats = jo["per_order"]["50%"]
slit_length_arcsec = calib.slit_length_arcsec
if fig is None:
fig = Figure(figsize=(8, 4))
grid = Grid(fig, 111, (3, 1), share_y=False, axes_pad=0.15)
ax31 = grid[0]
ax32 = grid[1]
ax33 = grid[2]
for o, k in stats:
ax31.plot(o, k["height"], "o")
bg_line, = ax31.plot([o for (o, _) in stats],
[_["bg"] for (o, _) in stats],
color="0.8",
ls="--")
ax31.legend([bg_line], ["background"], loc=1)
for o, k in stats:
ax32.plot(o, k["equivalent_width"] * slit_length_arcsec, "o")
for o, k in stats:
v = (k["height"] * k["equivalent_width"] * k["slit_length_in_pixel"] *
3.5)
ax33.plot(o, v**.5, "o")
ax33.set_xlabel("order number")
ax31.set_ylabel("peak count\n/ piexl")
ax32.set_ylabel("FWHM [\"]")
ax33.set_ylabel("(total counts\nper RE)^1/2")
return fig
def plotPairwiseClass(X, Y, labels=None, **kwargs):
X = np.array(X)
Y = np.array(Y)
nVariables = X.shape[1]
fig = pl.figure(1, (11, 8.5))
grid = Grid(
fig,
111, # similar to subplot(132)
nrows_ncols=(nVariables, nVariables),
axes_pad=0.0,
share_all=False,
label_mode="L",
direction="column",
)
if labels is None:
labels = ['var%d' % i for i in range(nVariables)]
for i in range(nVariables):
for j in range(nVariables):
nSub = i * (nVariables) + j
print nSub
if i == j:
n1, bins1 = np.histogram(X[:, i][np.where(np.array(Y) == 1)],
bins=20)
n0, bins1 = np.histogram(X[:, i][np.where(np.array(Y) == 0)],
bins=bins1)
scaledata = np.max(X[:, i]) / np.max((n0, n1))
grid[nSub].plot(.5 * (bins1[0:-1] + bins1[1:]), scaledata * n1,
'b-')
grid[nSub].plot(.5 * (bins1[0:-1] + bins1[1:]), scaledata * n0,
'r-')
grid[nSub].text(.5, .8, labels[i])
else:
grid[nSub].scatter(X[:, i],
X[:, j],
c=Y,
alpha=0.2,
linewidth=0,
cmap='seismic',
**kwargs)
return fig
def plotFunnels(proteins, decoys, decoys_scores, outputFile): from matplotlib import pylab as plt import numpy as np fig = plt.figure(figsize=(20,20)) N = len(proteins) sqrt_n = int(np.sqrt(N)) if N==sqrt_n*sqrt_n: nrows = int(np.sqrt(N)) ncols = int(N/nrows) else: nrows = int(np.sqrt(N))+1 ncols = int(N/nrows) if nrows*ncols<N: ncols+=1 from mpl_toolkits.axes_grid1 import Grid grid = Grid(fig, rect=111, nrows_ncols=(nrows,ncols), axes_pad=0.25, label_mode='L',share_x=False,share_y=False) num_proteins = [ (s,int(s[1:])) for s in proteins] num_proteins = sorted(num_proteins, key=lambda x: x[1]) proteins, num_proteins = zip(*num_proteins) for n,protein in enumerate(proteins): tmscores = [] scores = [] for decoy in decoys[protein]: tmscores.append(decoy[1]) scores.append(decoys_scores[protein][decoy[0]]) grid[n].plot(tmscores,scores,'.') plt.xlim(-0.1, max(tmscores)+0.1) plt.ylim(min(scores)-1, max(scores)+1) grid[n].set_title(protein) #plt.tight_layout() plt.savefig(outputFile)
def plot_exp(data, stat, plot_type, figname, savefig, col='abserr'):
level_name, xlabel = get_levelname_and_xlabel(plot_type)
grouped = data[col].dropna().groupby(level=(level_name, 'estimation',
'knode')).agg(stat)
n_params = len(grouped.groupby(level=('knode', )).groups.keys())
fig = plt.figure(figsize=(8, n_params * 3))
grid = Grid(fig,
111,
nrows_ncols=(n_params, 1),
add_all=True,
share_all=False,
label_mode='L',
share_x=True,
share_y=False,
axes_pad=.25)
for i, (param_name,
param_data) in enumerate(grouped.groupby(level=('knode', ))):
ax = grid[i]
ax.set_ylabel(param_name)
for est_name, est_data in param_data.groupby(level=['estimation']):
ax.errorbar(est_data.index.get_level_values(level_name),
est_data,
label=est_name,
lw=2.,
marker='o')
ax.set_xlabel(xlabel)
plt.legend(loc=0)
title = '%s_exp_%s' % (plot_type, figname)
plt.suptitle(title)
if savefig:
plt.savefig(title + '.png')
plt.savefig(title + '.svg')
def main():
args = parseargs()
if args.l:
# Load layout
font = load_attribute(args.l + '.font')
matplotlib.rc('font', **font)
graph = None
if args.show or args.s:
fig = plt.figure()
ax_main, ax_res = Grid(fig,
rect=111,
nrows_ncols=(2, 1),
axes_pad=0.2,
label_mode='L')
# ax_main.get_xaxis().set_visible(False)
# div = make_axes_locatable(ax_main)
# ax_res = div.append_axes('bottom', pad=0.05, label='Residuals',
# size='50%')
# ax_res = fig.add_subplot(212, label='Residuals')
graph = (fig, (ax_main, ax_res))
f_class = args.f()
control = (None, None, None)
colors = iter(cm.brg(np.linspace(0, 1, len(args.i) + bool(args.c))))
if args.c:
_, dat = unpack_data(args.c, args.delimiter).__next__()
data = [('control', dat)]
control = analysis_update(data,
f_class,
colors.__next__(),
graph,
report=False,
control=control,
notify=False)
for path in args.i:
color = colors.__next__()
try:
data = unpack_data(path,
args.delimiter,
filtr=args.filtr,
split_column=args.split)
analysis_update(data,
f_class,
color,
graph,
args.r,
control=control,
notify=True)
except Exception as e:
raise e
print('Could not process', path, os.linesep)
return None
if args.s:
plt.tight_layout()
bname = os.path.basename(path)
name = os.path.splitext(bname)[0]
fig.savefig(os.path.join(args.s, name + '.png'),
dpi=160,
bbox_inches='tight',
pad_inches=0)
if args.show:
plt.tight_layout()
plt.show()
# 5 center
init_mu,mu_index = initial_centers(image_vector,5)
final_c, final_mu = Kmeans(image_vector,init_mu,100)
f, ax = plt.subplots(1, 5,figsize=(12,3))
for i in range(5):
ax[i].imshow(final_mu[i].reshape(28,28))
ax[i].get_xaxis().set_visible(False)
ax[i].get_yaxis().set_visible(False)
plt.show()
# 10 center
init_mu,mu_index = initial_centers(image_vector,10)
final_c, final_mu = Kmeans(image_vector,init_mu,200)
fig = plt.figure()
grid = Grid(fig, rect=111, nrows_ncols=(2,5),axes_pad=0.1)
for i in range(10):
grid[i].imshow(final_mu[i].reshape(28,28))
grid[i].get_xaxis().set_visible(False)
grid[i].get_yaxis().set_visible(False)
plt.show()
# 20 center
init_mu,mu_index = initial_centers(image_vector,20)
final_c, final_mu = Kmeans(image_vector,init_mu,400)
fig = plt.figure()
grid = Grid(fig, rect=111, nrows_ncols=(4,5),axes_pad=0.1)
for i in range(20):
grid[i].imshow(final_mu[i].reshape(28,28))
grid[i].get_xaxis().set_visible(False)
grid[i].get_yaxis().set_visible(False)
def plot_t1t1_params(fig,
conn_and_dist,
assign_vect,
ss,
hps,
MAX_DIST=10,
model="LogisticDistance",
MAX_CLASSES=20):
"""
In the same order that we would plot the latent matrix, plot
the per-parameter properties
hps are per-relation hps
note, tragically, this wants the whole figure
"""
from mpl_toolkits.axes_grid1 import Grid
from matplotlib import pylab
assign_vect = np.array(assign_vect)
canon_assign_vect = util.canonicalize_assignment(assign_vect)
# create the mapping between existing and new
canon_to_old = {}
for i, v in enumerate(canon_assign_vect):
canon_to_old[v] = assign_vect[i]
CLASSES = np.sort(np.unique(canon_assign_vect))
CLASSN = len(CLASSES)
if CLASSN > MAX_CLASSES:
print "WARNING, TOO MANY CLASSES"
CLASSN = MAX_CLASSES
img_grid = Grid(
fig,
111, # similar to subplot(111)
nrows_ncols=(CLASSN, CLASSN),
axes_pad=0.1,
add_all=True,
share_all=True,
label_mode='L',
)
if "istance" not in model:
return
for c1i, c1_canon in enumerate(CLASSES[:MAX_CLASSES]):
for c2i, c2_canon in enumerate(CLASSES[:MAX_CLASSES]):
c1 = canon_to_old[c1_canon]
c2 = canon_to_old[c2_canon]
ax_pos = c1i * CLASSN + c2i
ax = img_grid[ax_pos]
nodes_1 = np.argwhere(assign_vect == c1).flatten()
nodes_2 = np.argwhere(assign_vect == c2).flatten()
conn_dist_hist = []
noconn_dist_hist = []
flatten_dist_val = []
assert len(nodes_1) > 0
assert len(nodes_2) > 0
for n1 in nodes_1:
for n2 in nodes_2:
d = conn_and_dist[n1, n2]['distance']
if conn_and_dist[n1, n2]['link']:
conn_dist_hist.append(d)
else:
noconn_dist_hist.append(d)
flatten_dist_val.append((d, conn_and_dist[n1, n2]['link']))
flatten_dist_val = np.array(flatten_dist_val)
bins = np.linspace(0, MAX_DIST, 20)
fine_bins = np.linspace(0, MAX_DIST, 100)
if model == "LogisticDistance" or model == "LogisticDistanceFixedLambda":
# compute prob as a function of distance for this class
htrue, _ = np.histogram(conn_dist_hist, bins)
hfalse, _ = np.histogram(noconn_dist_hist, bins)
p = htrue.astype(float) / (hfalse + htrue)
ax.plot(bins[:-1], p, c='b', linewidth=3)
if model == "LogisticDistance":
c = ss[(c1, c2)]
print "MAX_DISTANCE=", MAX_DIST, np.max(fine_bins), np.max(
bins), c
y = util.logistic(fine_bins, c['mu'], c['lambda'])
y = y * (hps['p_max'] - hps['p_min']) + hps['p_min']
ax.plot(fine_bins, y, c='r', linewidth=2)
ax.text(0, 0.2, r"mu: %3.2f" % c['mu'], fontsize=4)
ax.text(0, 0.6, r"lamb: %3.2f" % c['lambda'], fontsize=4)
ax.axvline(c['mu'], c='k')
elif model == "LogisticDistanceFixedLambda":
c = ss[(c1, c2)]
y = util.logistic(fine_bins, c['mu'], hps['lambda'])
y = y * (c['p_scale'] - hps['p_min']) + hps['p_min']
ax.plot(fine_bins, y, c='r')
ax.text(0, 0.2, r"mu: %3.2f" % c['mu'], fontsize=4)
ax.text(0, 0.6, r"lamb: %3.2f" % hps['lambda'], fontsize=4)
ax.axvline(c['mu'], c='k')
elif model == "ExponentialDistancePoisson":
if len(flatten_dist_val) > 0:
x_jitter = np.random.normal(0, 0.01, len(flatten_dist_val))
y_jitter = np.random.normal(0, 0.05, len(flatten_dist_val))
ax.scatter(flatten_dist_val[:, 0] + x_jitter,
flatten_dist_val[:, 1] + y_jitter,
edgecolor='none',
s=2)
c = ss[(c1, c2)]
mu = c['mu']
rate_scale = c['rate_scale']
y = np.exp(-fine_bins / mu)
y = y * rate_scale
ax.plot(fine_bins, y, c='r')
ax.text(0, 0.2, r"mu: %3.2f" % c['mu'], fontsize=4)
ax.text(0,
0.6,
r"rate_scale: %3.2f" % c['rate_scale'],
fontsize=4)
ax.set_ylim(-1, 20.0)
ax.axvline(c['mu'], c='k')
elif model == "LogisticDistancePoisson":
if len(flatten_dist_val) > 0:
x_jitter = np.random.normal(0, 0.01, len(flatten_dist_val))
y_jitter = np.random.normal(0, 0.05, len(flatten_dist_val))
ax.scatter(flatten_dist_val[:, 0] + x_jitter,
flatten_dist_val[:, 1] + y_jitter,
edgecolor='none',
s=2)
c = ss[(c1, c2)]
y = util.logistic(fine_bins, c['mu'], hps['lambda'])
y = y * (c['rate_scale'] - hps['rate_min']) + hps['rate_min']
ax.plot(fine_bins, y, c='r')
ax.text(0, 1, r"mu: %3.2f" % c['mu'], fontsize=4)
ax.text(0,
3,
r"rate_scale: %3.2f" % c['rate_scale'],
fontsize=4)
ax.set_ylim(-1, 20.0)
ax.axvline(c['mu'], c='k')
elif model == "LinearDistance":
print "MAX_DISTANCE=", MAX_DIST, np.max(fine_bins), np.max(
bins)
c = ss[(c1, c2)]
y = util.linear_dist(fine_bins, c['p'], c['mu'])
y += hps['p_min']
ax.plot(fine_bins, y, c='r')
ax.set_xlim(0, MAX_DIST)
def __init__(self):
self.figure = plt.figure(figsize=(6 * 0.9, 10 * 0.9))
self.ax = Grid(self.figure, rect=111, nrows_ncols=(5, 1), axes_pad = 0.0)
leg.set_in_layout(False)
fig.tight_layout()
plt.show()
###############################################################################
# Use with AxesGrid1
# ==================
#
# While limited, :mod:`mpl_toolkits.axes_grid1` is also supported.
from mpl_toolkits.axes_grid1 import Grid
plt.close('all')
fig = plt.figure()
grid = Grid(fig, rect=111, nrows_ncols=(2, 2),
axes_pad=0.25, label_mode='L',
)
for ax in grid:
example_plot(ax)
ax.title.set_visible(False)
plt.tight_layout()
###############################################################################
# Colorbar
# ========
#
# If you create a colorbar with `.Figure.colorbar`, the created colorbar is
# an instance of Axes, *not* Subplot, so tight_layout does not work. With
# Matplotlib v1.1, you may create a colorbar as a subplot using the gridspec.
def plot_comparisons(plogs, plog_filenames, order, rows_cols = (3, 6)):
# theoretical optimal mean and rate
dH_mean_opt = 0.41
acc_opt = 0.65
interval = 100
xlim = (0, 20000)
# determine figure size by sizes of individual subplots
figsize = (rows_cols[1] * 20./6, rows_cols[0] * 12.5/3)
# total acceptance rate
fig_acc = pl.figure(figsize = figsize)
fig_acc.suptitle("Total accumulated acceptance rate after # iteration attempts. Only showing one in each %i iterations." % interval)
axgrid_acc = Grid(fig_acc, rect=111, nrows_ncols=rows_cols, axes_pad=0.25, label_mode='L')
for i in range(rows_cols[0]*rows_cols[1]):
j = order[i]
accepted = np.cumsum(plogs[j]['acc'])
accrate_full = accepted / np.arange(1, len(plogs[j]['acc'])+1)
name = plog_filenames[j][43:plog_filenames[j].rfind("/")]
axgrid_acc[i].set_title(name)
accrate = accrate_full[::interval]
iteration_nr = np.arange(0, len(accrate_full), interval)
axgrid_acc[i].plot(iteration_nr, accrate, '.', label='acc.rate')
axgrid_acc[i].axhline(y=acc_opt, c='brown', ls='--', label='optimal')
acc200 = accepted.searchsorted(200)
axgrid_acc[i].axvline(x=acc200, c='magenta', ls=':', label="200 acc'ed")
axgrid_acc[i].axhline(y=accrate_full[acc200:].mean(), c='magenta', ls=':', label="mean, acc'ed > 200")
axgrid_acc[i].set_ylim(0, 1)
axgrid_acc[i].set_xlim(xlim)
axgrid_acc[i].set_xlabel("iteration attempt #")
axgrid_acc[i].set_ylabel("acceptance rate till then")
pl.tight_layout()
pl.subplots_adjust(top=0.9) # to make sure doesn't overlap with suptitle
# windowed acceptance rate
window_size = 500
fig_accwin = pl.figure(figsize = figsize)
fig_accwin.suptitle("Windowed acceptance rate (accumulated over %i previous iterations) after # iteration attempts. Only showing one in each %i iterations." % (window_size, interval))
axgrid_accwin = Grid(fig_accwin, rect=111, nrows_ncols=rows_cols, axes_pad=0.25, label_mode='L')
for i in range(rows_cols[0]*rows_cols[1]):
j = order[i]
accepted = plogs[j]['acc']
name = plog_filenames[j][43:plog_filenames[j].rfind("/")]
ax = axgrid_accwin[i]
plot_windowed_accrate(accepted, window_size, interval, name, ax, acc_opt = acc_opt, xlim = xlim)
pl.tight_layout()
pl.subplots_adjust(top=0.9) # to make sure doesn't overlap with suptitle
# dH
fig_dH = pl.figure(figsize = figsize)
fig_dH.suptitle("dH for each iteration attempt. Only showing one in each %i iterations." % interval)
axgrid_dH = Grid(fig_dH, rect=111, nrows_ncols=rows_cols, axes_pad=0.25, label_mode='L')
for i in range(rows_cols[0]*rows_cols[1]):
j = order[i]
name = plog_filenames[j][43:plog_filenames[j].rfind("/")]
axgrid_dH[i].set_title(name)
axgrid_dH[i].set_yscale('log')
dH_j_full = plogs[j]['dH']
dH_j = dH_j_full[::interval]
pos = np.argwhere(dH_j >= 0)
neg = np.argwhere(dH_j < 0)
iteration_nr = np.arange(0, len(dH_j_full), interval)
axgrid_dH[i].plot(iteration_nr[pos], dH_j[pos], '.b', label='dH > 0')
axgrid_dH[i].plot(iteration_nr[neg], -dH_j[neg], '.r', label='-dH > 0')
axgrid_dH[i].axhline(y=dH_mean_opt, c='brown', ls='--', label='optimal')
accepted = np.cumsum(plogs[j]['acc'])
acc200 = accepted.searchsorted(200)
axgrid_dH[i].axvline(x=acc200, c='magenta', ls=':', label="200 acc'ed")
axgrid_dH[i].axhline(y=plogs[j]['dH'][acc200:].mean(), c='magenta', ls=':', label="mean, acc'ed > 200")
axgrid_dH[i].set_xlim(xlim)
axgrid_dH[i].set_xlabel("iteration attempt #")
axgrid_dH[i].set_ylabel("dH")
pl.tight_layout()
pl.subplots_adjust(top=0.9) # to make sure doesn't overlap with suptitle
# convergence: likelihood (chi^2 for gaussian likelihood)
fig_chisq = pl.figure(figsize = figsize)
fig_chisq.suptitle("Convergence measure: likelihood (is chi-squared for gaussian likelihood) at beginning of iteration. Only showing one in each %i iterations." % interval)
axgrid_chisq = Grid(fig_chisq, rect=111, nrows_ncols=rows_cols, axes_pad=0.25, label_mode='L')
for i in range(rows_cols[0]*rows_cols[1]):
j = order[i]
likeli_full = plogs[j]['psi_likeli_i']
likeli = likeli_full[::interval]
name = plog_filenames[j][43:plog_filenames[j].rfind("/")]
axgrid_chisq[i].set_title(name)
iteration_nr = np.arange(0, len(likeli_full), interval)
axgrid_chisq[i].plot(iteration_nr, likeli, '.', label='likelihood')
axgrid_chisq[i].set_xlim(xlim)
axgrid_chisq[i].set_xlabel("iteration attempt #")
axgrid_chisq[i].set_ylabel("likelihood")
pl.tight_layout()
pl.subplots_adjust(top=0.9) # to make sure doesn't overlap with suptitle
axgrid_chisq[rows_cols[1]-1].legend()
axgrid_dH[rows_cols[1]-1].legend()
axgrid_acc[rows_cols[1]-1].legend()
axgrid_accwin[rows_cols[1]-1].legend()
return fig_acc, fig_accwin, fig_dH, fig_chisq
'ytick.labelsize': fontScale,
#'text.usetex': True,
'figure.figsize': fig_size
}
plt.rcdefaults()
plt.rcParams.update(params)
plt.figure(1)
#plt.tight_layout(pad=0.0)
plt.clf()
plt.subplots_adjust(bottom=0.2)
grid = Grid(fig, (1, 1, 1),
nrows_ncols=(1, 1),
ngrids=1,
direction='column',
axes_pad=0,
label_mode='L',
share_all=True)
rNum = 5
pressureComp,error = calc_pressure(radii[0:rNum],hi16[0:rNum],8.4,\
hiErr=hiErradii[0:rNum],vdispErr=1.4)
markerSize = 5
# Pressure Component
if True:
plt.errorbar(radii[0:rNum - 1],
pressureComp[0:rNum],
gmm_models = GaussianMixture(n_components=n_comp, covariance_type='diag')
gmm_models.fit(feature_train[train_idx, :])
# compute llh for each patch
city_dict = {'aus': 0, 'chi': 1, 'kit': 2, 'tyr': 3, 'vie': 4}
feature_valid = feature[idx < 6, :]
patch_valid = [
patch_names[i] for i in range(len(patch_names)) if idx[i] < 6
]
city_name_list = [a[:3] for a in patch_valid]
city_id_list = [city_dict[a] for a in city_name_list]
fig = plt.figure(figsize=(3, 8))
grid = Grid(fig,
rect=111,
nrows_ncols=(5, 1),
axes_pad=0.25,
label_mode='L',
share_all=True)
for test_city in range(5):
test_city_feature = feature_valid[[
i for i in range(len(city_id_list)) if city_id_list[i] == test_city
], :]
llh = gmm_models.score_samples(test_city_feature)
ax = grid[test_city]
ax.hist(llh, bins=100)
ax.set_xlabel('LLH')
ax.set_ylabel(city_list[test_city])
plt.suptitle(city_list[city_1])
plt.tight_layout()
fig = plt.figure()
mxx = rmsdPerAtom.max() * 1.1
#figure out the geometry of the plotting window
natom = rmsdPerAtom.shape[1]
if natom < 49:
nr, nc = 6, 8
elif natom < 97:
nr, nc = 12, 8
elif natom < 151:
nr, nc = 15, 10
elif natom < 193:
nr, nc = 16, 12
grid = Grid(fig, rect=111, nrows_ncols=(nr, nc), axes_pad=0.05, label_mode='L')
def coarseGrain(x, cg=10):
return [sum(x[i * cg:i * cg + cg]) / cg for i in range(int(len(x) / cg))]
def colorLine(x, y, c=None, cmap=plt.get_cmap('cool'), alpha=0.4):
if c == None:
c = linspace(0.0, 1.0, len(x))
c = array(c)
points = array([x, y]).T.reshape(-1, 1, 2)
segments = concatenate([points[:-1], points[1:]], axis=1)
lc = mc.LineCollection(segments, array=c, cmap=cmap, alpha=alpha)
return lc
