def fig(self):
fig, axes = plt.subplots(nrows=self.nrows, ncols=self.ncols, figsize=self.figsize)
divider = Divider(fig, (0.0, 0.0, 1.0, 1.0), self.h, self.v, aspect=False)
if (self.nrows == 1) & (self.ncols == 1):
axes = [axes]
axes[0].set_axes_locator(divider.new_locator(nx=1, ny=1))
if self.n_axes > 1:
for i in range(self.n_axes-1):
axes.append(fig.add_subplot(111))
axes[i+1].set_axes_locator(divider.new_locator(nx=1, ny=1))
elif self.nrows > 1:
for j in range(self.nrows):
axes[j].set_axes_locator(divider.new_locator(nx=1, ny=j*2+1))
elif self.ncols > 1:
for i in range(self.ncols):
axes[i].set_axes_locator(divider.new_locator(nx=i*2+1, ny=1))
if (self.nrows > 1) & (self.ncols > 1):
for j in range(self.nrows):
for i in range(self.ncols):
axes[i, j].set_axes_locator(divider.new_locator(nx=i*2+1, ny=j*2+1))
return fig, axes
def demo_fixed_pad_axes():
fig = plt.figure(2, (6, 6))
# The first & third items are for padding and the second items are for the axes.
# sizes are in inch.
h = [
Size.Fixed(1.0),
Size.Scaled(1.),
Size.Fixed(.2),
]
v = [
Size.Fixed(0.7),
Size.Scaled(1.),
Size.Fixed(.5),
]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
# the width and height of the rectangle is ignored.
ax = LocatableAxes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
ax.plot([1, 2, 3])
def metric_svg(self,
host=None,
title="undefined",
ylabel="undefined",
metrics=[],
b64encode=True):
fig = pyplot.figure(figsize=self.FIGSIZE)
h = [Size.Fixed(1.2), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
plt = Axes(fig, divider.get_position())
plt.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(plt)
result = self.result
runinfo = result.runinfo
# ----
# The X limits are -rampupMins, runMins
# ----
plt.set_xlim(-int(runinfo['rampupMins']), int(runinfo['runMins']))
plt.axvspan(-int(runinfo['rampupMins']), 0, facecolor='0.2', alpha=0.1)
# ----
# Draw all the requested graphs
# ----
x = None
for m in metrics:
if x is None:
x, y = self._get_metric_tree(m)
else:
_, y = self._get_metric_tree(m, x)
plt.plot(x, y, m['color'], label=m['label'])
# ----
# Title and labels
# ----
plt.set_title(title)
plt.set_xlabel("Elapsed Minutes")
plt.set_ylabel(ylabel)
plt.legend(loc='upper left')
plt.grid()
# ----
# Now turn this into an in-memory SVG and return it as requested
# (raw or b64-encoded)
# ----
buf = io.StringIO()
pyplot.savefig(buf, format='svg')
if not b64encode:
return buf.getvalue()
return base64.b64encode(buf.getvalue().encode('utf-8')).decode('utf-8')
def plot_it():
# fig = plt.figure(1, figsize=(8, 4.5))
fig = plt.figure(1, figsize=(16, 12))
# h = [Size.Fixed(0.), Size.Fixed(6.5)]
# v = [Size.Fixed(0.5), Size.Fixed(3.25)]
h = [Size.Fixed(0.), Size.Fixed(13)]
v = [Size.Fixed(0.5), Size.Fixed(7)]
win = Divider(fig, (0.1, 0.08, 0.8, 0.8), h, v, aspect=False)
ax = Axes(fig, win.get_position())
ax.set_axes_locator(win.new_locator(nx=1, ny=1))
fig.add_axes(ax)
# fig = plt.figure(figsize=(16, 12))
# ax = fig.add_subplot(111)
for ii, ellipse in enumerate(ellipses):
ax.fill(ellipse[0], ellipse[1], color=cmap(norm_vals[ii]), zorder=0)
ax.plot(ellipse[0], ellipse[1], 'k-', linewidth=0.2)
ax.invert_yaxis()
plt.xlabel('Northing (km)', fontsize=16)
plt.ylabel(r'$\log_{10}$ Period (s)', fontsize=16)
# ax.set_aspect(1)
ax.tick_params(axis='both', labelsize=14)
# locs, labels = plt.xticks()
# plt.xticks(locs, [int(x * 10) for x in locs])
fake_vals = np.linspace(lower, upper, len(fill_vals))
fake_im = ax.scatter(loc, periods, c=fake_vals, cmap=cmap)
fake_im.set_visible(False)
# ax.set_ylim([-2.6, 3.25])
# ax.set_xlim([526.5, 538.2])
# ax.invert_yaxis()
# cb = plt.colorbar(mappable=fake_im)
#############################################
# Colour bar and site labelling
# cbaxes = fig.add_axes([0.925, 0.1351, 0.015, 0.72])
# cb = plt.colorbar(fake_im)
# if 'phi' in fill_param[:3]:
# label = r'${}{}(\degree)$'.format('\phi', fill_param[-2:])
# else:
# label = r'$\{}(\degree)$'.format(fill_param)
# cb.set_label(label,
# rotation=270,
# labelpad=20,
# fontsize=18)
# ax.tick_params(axis='both', labelsize=14)
# ax.set_xlim(x_lim)
# for ii, site in enumerate(main_sites):
# txt = site[-4:-1]
# if linear_xaxis:
# ax.text(linear_site[ii],
# label_offset, site, rotation=45) # 3.6
# else:
# ax.text(main_transect.sites[site].locations['X'],
# label_offset, site, rotation=45) # 3.6
plt.show()
return fig
def _set_ax_height_to_cm(fig: Fig, ax: Ax, height: float) -> None:
from mpl_toolkits.axes_grid1 import Size, Divider
height /= 2.54 # cm to inches
bbox = _get_ax_bbox(fig, ax)
hori = [Size.Fixed(bbox.x0), Size.Fixed(bbox.width), Size.Fixed(bbox.x1)]
vert = [Size.Fixed(bbox.y0), Size.Fixed(height), Size.Fixed(bbox.y1)]
divider = Divider(fig, (0.0, 0.0, 1.0, 1.0), hori, vert, aspect=False)
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
def demo_fixed_size_axes():
fig1 = plt.figure(1, (10, 5))
# The first items are for padding and the second items are for the axes.
# sizes are in inch.
h = [Size.Fixed(1.0), Size.Fixed(7.5)]
v = [Size.Fixed(0.7), Size.Fixed(4.5)]
divider = Divider(fig1, (0.0, 0.0, 1., 1.), h, v, aspect=False)
# the width and height of the rectangle is ignored.
ax = LocatableAxes(fig1, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig1.add_axes(ax)
ax.plot([1, 2, 3])
def fix_sz_fig(width, height, width_ext=2, padding=.5):
# These code is to make fixed size article
# Makesure matplotlib can draw Chinese characters
plt.rcParams['font.sans-serif'] = ['SimHei']
# Init figure
fig = plt.figure(figsize=(width + 2 * padding + width_ext,
height + 2 * padding))
# The first items are for padding and the second items are for the axes.
# Article width and height
w = [Size.Fixed(padding), Size.Fixed(width)]
h = [Size.Fixed(padding), Size.Fixed(height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), w, h, aspect=False)
# the width and height of the rectangle is ignored.
ax = Axes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
return fig, ax
def __init__(self,
ax_size,
wspace,
dpi,
cbar,
cbar_width,
left,
bottom,
right,
top,
fig_func=None):
if not cbar:
wspace = 0
cbar_width = 0
# convert arguments from pixels to inches
ax_size = np.asarray(ax_size) / dpi
wspace, cbar_width, left, bottom, right, top = np.array(
(wspace, cbar_width, left, bottom, right, top)) / dpi
horiz = [left, ax_size[0], wspace, cbar_width, right]
vert = [bottom, ax_size[1], top]
figsize = (sum(horiz), sum(vert))
fig_func = plt.figure if fig_func is None else fig_func
fig = fig_func(figsize=figsize, dpi=dpi)
horiz, vert = list(map(Size.Fixed, horiz)), list(map(Size.Fixed, vert))
divider = Divider(fig, (0.0, 0.0, 1., 1.), horiz, vert, aspect=False)
ax = LocatableAxes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
if cbar:
cax = LocatableAxes(fig, divider.get_position())
cax.set_axes_locator(divider.new_locator(nx=3, ny=1))
fig.add_axes(cax)
self.fig = fig
self.ax = ax
self.cax = cax if cbar else None
def create_plot(two_sided=False,
colors=[
'#6F4C9B', '#5568B8', '#4D8AC6', '#60AB9E', '#77B77D',
'#A6BE54', '#D1B541', '#E49C39', '#DF4828', '#990D38'
],
markers=['o', 'v', '^', 's', 'D', '*'],
figsize=(5, 3.4)):
"""
Crée un environnement de plot
Parameters
----------
twosided : bool
allows to change the size of the figure accordingly.
colors : list of strings
a default list exists but this allows to change it if u want
markers : list of strings
a default list of markers exists, but u can change it if needed
Returns
-------
fig, ax : matplotlib objects to be used as normal
"""
color = cycle(colors)
marker = cycle(markers)
if two_sided:
fig = plt.figure(figsize=(3.4, 3.4))
else:
fig = plt.figure(figsize=figsize)
# The first & third items are for padding and the second items are for the
# axes. Sizes are in inches.
h = [Size.Fixed(1.0), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
# the width and height of the rectangle is ignored.
ax = Axes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
return fig, ax, color, marker
def get_divider(self, fig):
self.border = 0.1
while self.actual_size[0] < (
self.size_on_screen[0] - 2 * self.border
) and self.actual_size[1] < (self.size_on_screen[1] - 2 * self.border):
self.border += 0.1
x_factor = self.actual_size[0] / (self.size_on_screen[0] -
2 * self.border)
y_factor = self.actual_size[1] / (self.size_on_screen[1] -
2 * self.border)
self.scale = max(x_factor, y_factor)
draw_size = (self.actual_size[0] / self.scale,
self.actual_size[1] / self.scale)
h = [Size.Fixed(self.border), Size.Fixed(draw_size[0])]
v = [Size.Fixed(self.border), Size.Fixed(draw_size[1])]
return Divider(
fig,
(self.location_on_screen[0], self.location_on_screen[1], 1, 1),
h,
v,
aspect=False)
def set_size3(ax, w, h):
# https://newbedev.com/axes-class-set-explicitly-size-width-height-of-axes-in-given-units
from mpl_toolkits.axes_grid1 import Divider, Size
axew = w / 2.54
axeh = h / 2.54
# lets use the tight layout function to get a good padding size for our axes labels.
# fig = plt.gcf()
# ax = plt.gca()
fig = ax.get_figure()
fig.tight_layout()
# obtain the current ratio values for padding and fix size
oldw, oldh = fig.get_size_inches()
l = ax.figure.subplotpars.left
r = ax.figure.subplotpars.right
t = ax.figure.subplotpars.top
b = ax.figure.subplotpars.bottom
# work out what the new ratio values for padding are, and the new fig size.
# ps: adding a bit to neww and newh gives more padding
# the axis size is set from axew and axeh
neww = axew + oldw * (1 - r + l) + 0.4
newh = axeh + oldh * (1 - t + b) + 0.4
newr = r * oldw / neww - 0.4
newl = l * oldw / neww + 0.4
newt = t * oldh / newh - 0.4
newb = b * oldh / newh + 0.4
# right(top) padding, fixed axes size, left(bottom) pading
hori = [Size.Scaled(newr), Size.Fixed(axew), Size.Scaled(newl)]
vert = [Size.Scaled(newt), Size.Fixed(axeh), Size.Scaled(newb)]
divider = Divider(fig, (0.0, 0.0, 1.0, 1.0), hori, vert, aspect=False)
# the width and height of the rectangle is ignored.
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
# we need to resize the figure now, as we have may have made our axes bigger than in.
fig.set_size_inches(neww, newh)
def subplot_array(self, hsize, vsize=(1.0, ), figsize=(10, 10)):
""" Use the axes_divider module to make a single row of plots
hsize : list of floats
horizontal spacing: alternates Scaled for plot, Fixed for between plots
vsize : list of floats
vertical spacing
ref: http://matplotlib.org/mpl_toolkits/axes_grid/users/axes_divider.html
"""
nx = (len(hsize) + 1) / 2
ny = (len(vsize) + 1) / 2
fig, axx = plt.subplots(
ny, nx, squeeze=False,
figsize=figsize) # just to make the axes, will move them
sizer = lambda x, i: axes_size.Scaled(
x) if i % 2 == 0 else axes_size.Fixed(x)
horiz = [sizer(h, i) for i, h in enumerate(hsize)]
vert = [sizer(v, i) for i, v in enumerate(vsize)]
divider = Divider(fig, (0.1, 0.1, 0.8, 0.8), horiz, vert, aspect=False)
for i, ax in enumerate(axx.flatten()):
iy = i // nx
ix = i % nx
ax.set_axes_locator(divider.new_locator(nx=2 * ix, ny=2 * iy))
return fig, axx
def cpu_svg(self,
host=None,
title="CPU Usage",
ylabel="Percent",
b64encode=True):
fig = pyplot.figure(figsize=self.FIGSIZE)
h = [Size.Fixed(1.2), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
plt = Axes(fig, divider.get_position())
plt.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(plt)
result = self.result
runinfo = result.runinfo
# ----
# The X limits are -rampupMins, runMins, Y limits are 0..100
# ----
plt.set_xlim(-int(runinfo['rampupMins']), int(runinfo['runMins']))
plt.set_ylim(0, 100)
plt.axvspan(-int(runinfo['rampupMins']), 0, facecolor='0.2', alpha=0.1)
# ----
# Gather all the relevant data
# ----
x, y_user = self._get_metric_xy({
'host': host,
'metric': 'cpu.percent-user',
})
_, y_sys = self._get_metric_xy(
{
'host': host,
'metric': 'cpu.percent-system',
}, x)
_, y_wait = self._get_metric_xy(
{
'host': host,
'metric': 'cpu.percent-wait',
}, x)
_, y_intr = self._get_metric_xy(
{
'host': host,
'metric': 'cpu.percent-interrupt',
}, x)
_, y_softirq = self._get_metric_xy(
{
'host': host,
'metric': 'cpu.percent-softirq',
}, x)
_, y_idle = self._get_metric_xy(
{
'host': host,
'metric': 'cpu.percent-idle',
}, x)
# ----
# It is possible that the mqtt based metric collector produces
# CSV files with different numbers of lines. We cut all of them
# to a combined minimum length.
# ----
min_len = min(len(x), len(y_user), len(y_sys), len(y_wait),
len(y_intr), len(y_softirq), len(y_idle))
x = x[:min_len]
y_user = y_user[:min_len]
y_sys = y_sys[:min_len]
y_wait = y_wait[:min_len]
y_intr = y_intr[:min_len]
y_softirq = y_softirq[:min_len]
y_idle = y_idle[:min_len]
# ----
# Plot the CPU usage
# ----
plt.stackplot(x,
y_user,
y_sys,
y_wait,
y_intr,
y_softirq,
y_idle,
colors=[
'g',
'b',
'r',
'c',
'm',
'y',
],
labels=[
'User',
'System',
'Wait',
'Interrupt',
'SoftIrq',
'Idle',
])
# ----
# Title and labels
# ----
plt.set_title(title)
plt.set_xlabel("Elapsed Minutes")
plt.set_ylabel(ylabel)
plt.legend(loc='upper left')
plt.grid()
# ----
# Now turn this into an in-memory SVG and return it as requested
# (raw or b64-encoded)
# ----
buf = io.StringIO()
pyplot.savefig(buf, format='svg')
if not b64encode:
return buf.getvalue()
return base64.b64encode(buf.getvalue().encode('utf-8')).decode('utf-8')
import mpl_toolkits.axes_grid1.axes_size as Size
from mpl_toolkits.axes_grid1 import Divider
import matplotlib.pyplot as plt
fig = plt.figure(figsize=(5.5, 4))
rect = (0.1, 0.1, 0.8, 0.8)
ax = [fig.add_axes(rect, label="%d" % i) for i in range(4)]
horiz = [Size.AxesX(ax[0]), Size.Fixed(.5), Size.AxesX(ax[1])]
vert = [Size.AxesY(ax[0]), Size.Fixed(.5), Size.AxesY(ax[2])]
divider = Divider(fig, rect, horiz, vert, aspect=False)
ax[0].set_axes_locator(divider.new_locator(nx=0, ny=0))
ax[1].set_axes_locator(divider.new_locator(nx=2, ny=0))
ax[2].set_axes_locator(divider.new_locator(nx=0, ny=2))
ax[3].set_axes_locator(divider.new_locator(nx=2, ny=2))
ax[0].set_xlim(0, 2)
ax[1].set_xlim(0, 1)
ax[0].set_ylim(0, 1)
ax[2].set_ylim(0, 2)
divider.set_aspect(1.)
for ax1 in ax:
ax1.tick_params(labelbottom=False, labelleft=False)
plt.show()
def segtrends(x, segments=2, charts=True):
"""
Turn minitrends to iterative process more easily adaptable to
implementation in simple trading systems; allows backtesting functionality.
:param x: One-dimensional data set
:param window: How long the trendlines should be. If window < 1, then it
will be taken as a percentage of the size of the data
:param charts: Boolean value saying whether to print chart to screen
"""
import numpy as np
y = np.array(x['Close'])
# for i in range(len(y),1200):
# y = np.append(y, y[len(y)-1])
# Implement trendlines
segments = int(segments)
maxima = np.ones(segments)
minima = np.ones(segments)
segsize = int(len(y) / segments)
for i in range(1, segments + 1):
ind2 = i * segsize
ind1 = ind2 - segsize
maxima[i - 1] = max(y[ind1:ind2])
minima[i - 1] = min(y[ind1:ind2])
# Find the indexes of these maxima in the data
x_maxima = np.ones(segments)
x_minima = np.ones(segments)
for i in range(0, segments):
x_maxima[i] = np.where(y == maxima[i])[0][0]
x_minima[i] = np.where(y == minima[i])[0][0]
if charts:
import matplotlib.pyplot as plt
plt.rc('font', size=6)
fig = plt.figure(figsize=(8, 6))
h = [Size.Fixed(0.5), Size.Fixed(7.)]
v = [Size.Fixed(0.7), Size.Fixed(5.)]
divider = Divider(fig, (0.0, 0.0, 0., 0.), h, v, aspect=False)
ax = Axes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
plt.plot(y, linewidth=1)
plt.grid(True)
for i in range(0, segments - 1):
maxslope = (maxima[i + 1] - maxima[i]) / (x_maxima[i + 1] -
x_maxima[i])
a_max = maxima[i] - (maxslope * x_maxima[i])
b_max = maxima[i] + (maxslope * (len(y) + 300 - x_maxima[i]))
maxline = np.linspace(a_max, b_max, len(y) + 300)
minslope = (minima[i + 1] - minima[i]) / (x_minima[i + 1] -
x_minima[i])
a_min = minima[i] - (minslope * x_minima[i])
b_min = minima[i] + (minslope * (len(y) + 300 - x_minima[i]))
minline = np.linspace(a_min, b_min, len(y) + 300)
if charts:
plt.plot(maxline, 'g', linewidth=0.15)
plt.plot(minline, 'r', linewidth=.15)
if charts:
#plt.show()
plt.ylim(min(y[-120:] * .9), max(y[-120:] * 1.1))
plt.rc('xtick', labelsize=6)
plt.xticks(range(0, 1200, 4), x.index[range(0, 910, 4)], rotation=90)
plt.rc('xtick', labelsize=6)
plt.xlim(800, 1000)
# plt.xlim(datetime.datetime(2016,1,1),datetime.datetime(2022,1,1))
plt.savefig('C:/git/TRADE/chart.png', dpi=750)
plt.close()
# OUTPUT
return x_maxima, maxima, x_minima, minima
def plot_stratigraphy(section,
style,
ncols=1,
linewidth=1,
col_spacings=0.5,
col_widths=1):
"""
Initialize a figure and subplots, with the stratigraphic section
already plotted on the first axis.
This function is intended to act similar to `plt.subplots()` in that
it will initialize the figure and axis handles. However, given that
controlling the exact width and height of the axes is critical, this
function is necessary to initialize the handles correctly.
Note that setting the figsize itself is not sufficient to control
the width and height of the axes exactly, since the figsize includes
the size of the padding around the axes.
Parameters
----------
section : Section
A Section object.
style : Style
A Style object.
ncols : int
The number of axes that will be in the figure (including the
axis with the stratigraphic section.)
linewidth : float
The linewidth when drawing the stratigraphic section.
col_spacings : float or array_like
The spacing between the axes in inches. If a float, this value
will be interpreted as uniform spacing between the axes. If an
array, the length of the array must be ncols - 1, and the values
will be the spacing between the individual axes.
col_widths : float or array_like
The width of the axes as a ratio relative to the width of the
stratigraphic column. If a float, this value will be interpreted
as a uniform width of the axes, excluding the stratigraphic
column, for which the width is explicitly set in the Style. If
an array, the length of the array must be ncols - 1, and the
values will be the widths of the individual axes excluding the
stratigraphic column.
Returns
-------
fig : matplotlib Figure
Figure handle.
ax : matplotlib Axes
Axis handle.
"""
# get the attributes - implicitly checks if the attributes exist
color_attribute = getattr(section, style.color_attribute)
width_attribute = getattr(section, style.width_attribute)
# initialize
fig, ax = plt.subplots(nrows=1, ncols=ncols, sharey=True)
# get the first axis
if ncols == 1:
ax0 = ax
else:
ax0 = ax[0]
# determine the axis height and limits first
ax_height = style.height_scaling_factor * section.total_thickness
ax0.set_ylim(0, section.total_thickness)
# initiate counting of the stratigraphic height
strat_height = 0.0
# loop over elements of the data
for i in range(section.n_units):
# pull out the thickness
this_thickness = section.thicknesses[i]
# loop over the elements in Style to get the color and width
for j in range(style.n_color_labels):
if color_attribute[i] == style.color_labels[j]:
this_color = style.color_values[j]
for j in range(style.n_width_labels):
if width_attribute[i] == style.width_labels[j]:
this_width = style.width_values[j]
# create the rectangle
ax0.add_patch(
Rectangle((0.0, strat_height),
this_width,
this_thickness,
facecolor=this_color,
edgecolor='k',
linewidth=linewidth))
# count the stratigraphic height
strat_height = strat_height + this_thickness
# set the axis dimensions (values below are all in inches)
ax0_lower_left_x = 0.5
ax0_lower_left_y = 0.5
ax0_width = style.width_inches
h = [Size.Fixed(ax0_lower_left_x), Size.Fixed(ax0_width)]
v = [Size.Fixed(ax0_lower_left_y), Size.Fixed(ax_height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax0.set_axes_locator(divider.new_locator(nx=1, ny=1))
# prettify
ax0.set_xlim(0, 1)
ax0.set_xticks([0, 0.2, 0.4, 0.6, 0.8, 1])
for label in ax0.get_xticklabels():
label.set_rotation(270)
label.set_ha('center')
label.set_va('top')
ax0.set_axisbelow(True)
ax0.xaxis.grid(ls='--')
ax0.spines['top'].set_visible(False)
ax0.spines['right'].set_visible(False)
ax0.set_ylabel('stratigraphic height [m]')
# turning the spines off creates some clipping mask issues
# so just turn the clipping masks off
for obj in ax0.findobj():
obj.set_clip_on(False)
# check if the col_spacing and col_widths is of the correct format
if type(col_spacings) == list or type(col_spacings) == np.ndarray:
if len(col_spacings) != ncols - 1:
raise Exception('col_spacings must be either a float or '
'array_like with length ncols-1.')
if type(col_widths) == list or type(col_widths) == np.ndarray:
if len(col_widths) != ncols - 1:
raise Exception('col_widths must be either a float or '
'array_like with length ncols-1.')
# set up the other axes
if ncols != 1:
# iterate through the axes
for i in range(1, ncols):
# get the spacing and width values
if type(col_spacings) == list or type(col_spacings) == np.ndarray:
col_spacing = col_spacings[i - 1]
else:
col_spacing = col_spacings
if type(col_widths) == list or type(col_widths) == np.ndarray:
col_width = col_widths[i - 1] * ax0_width
else:
col_width = col_widths * ax0_width
# adjust the axis
if i == 1:
axn_lower_left_x = ax0_lower_left_x + ax0_width + col_spacing
else:
axn_lower_left_x = axn_lower_left_x + axn_width + col_spacing
axn_lower_left_y = ax0_lower_left_y
axn_width = col_width
h = [Size.Fixed(axn_lower_left_x), Size.Fixed(axn_width)]
v = [Size.Fixed(axn_lower_left_y), Size.Fixed(ax_height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax[i].set_axes_locator(divider.new_locator(nx=1, ny=1))
# set the figsize here too to account for the axis size manipulation
if ncols != 1:
fig.set_size_inches(
ax0_lower_left_x * 2 + np.sum(col_spacings) + np.sum(col_widths) +
ax0_width, ax0_lower_left_y * 2 + ax_height)
else:
fig.set_size_inches(ax0_lower_left_x * 2 + ax0_width,
ax0_lower_left_y * 2 + ax_height)
return fig, ax
def __init__(self, images, shape):
fig = plt.figure(figsize=(10, 6))
# subplot positions
h_nav = [Size.Fixed(0.5), Size.Fixed(2.5)]
v_nav = [Size.Fixed(0.5), Size.Scaled(1.0), Size.Fixed(0.5)]
h_im = [Size.Fixed(0.5), Size.Scaled(1.0), Size.Fixed(0.5)]
v_im = [Size.Fixed(0.5), Size.Scaled(1.0), Size.Fixed(0.5)]
nav = Divider(fig, (0.0, 0.0, 0.2, 1.), h_nav, v_nav, aspect=False)
image = Divider(fig, (0.2, 0.0, 0.8, 1.), h_im, v_im, aspect=True)
image.set_anchor('C')
# Toolbar menu box
ax1 = LocatableAxes(fig, nav.get_position())
ax1.set_axes_locator(nav.new_locator(nx=1, ny=1))
ax1.get_xaxis().set_visible(False)
ax1.get_yaxis().set_visible(False)
fig.add_axes(ax1, label='toolbar')
ax1.text(0.05, 0.45, "Filter", weight='heavy',
transform=ax1.transAxes)
# Image space
ax2 = LocatableAxes(fig, image.get_position())
ax2.set_axes_locator(image.new_locator(nx=1, ny=1))
fig.add_axes(ax2, label='image_space')
self.callback = ImageIndex(images, shape, fig)
# Navigation
## Go to
ax_text_index = plt.axes([0.59, 0.05, 0.1, 0.075])
ip = InsetPosition(ax1, [0.2, 0.84, 0.3, 0.05])
ax_text_index.set_axes_locator(ip)
entry_index = TextBox(ax_text_index, 'Go to', initial="0")
entry_index.on_submit(self.callback.submit_index)
## Previous
ax_prev = plt.axes([0.7, 0.05, 0.075, 0.075])
ip = InsetPosition(ax1, [0.55, 0.84, 0.15, 0.05])
ax_prev.set_axes_locator(ip)
bprev = Button(ax_prev, '<<')
bprev.on_clicked(self.callback.prev)
## Next
ax_next = plt.axes([0.81, 0.05, 0.075, 0.075])
ip = InsetPosition(ax1, [0.75, 0.84, 0.15, 0.05])
ax_next.set_axes_locator(ip)
bnext = Button(ax_next, '>>')
bnext.on_clicked(self.callback.next)
# Bounding Boxes
ax_chec = plt.axes([0.1, 0.05, 0.35, 0.075])
ip = InsetPosition(ax1, [0.05, 0.5, 0.9, 0.3])
ax_chec.set_axes_locator(ip)
ax_chec.text(0.05, 0.85, "Bounding Boxes", transform=ax_chec.transAxes)
check = CheckButtons(ax_chec,
('characters', 'lines'),
(False, False))
check.on_clicked(self.callback.update_bboxes)
# Filtering
## Image
ax_text_image = plt.axes([0.1, 0.1, 0.1, 0.075])
ip = InsetPosition(ax1, [0.26, 0.38, 0.64, 0.05])
ax_text_image.set_axes_locator(ip)
entry_image = TextBox(ax_text_image, 'images',
initial="image_id,image_id")
entry_image.on_submit(self.callback.submit_images)
## Characters
ax_text_char = plt.axes([0.1, 0.2, 0.1, 0.075])
ip = InsetPosition(ax1, [0.21, 0.3, 0.69, 0.05])
ax_text_char.set_axes_locator(ip)
entry_char = TextBox(ax_text_char, 'chars',
initial="U+3055,U+3056")
entry_char.on_submit(self.callback.submit_chars)
## Reset
ax_reset = plt.axes([0., 0., 1., 1.])
ip = InsetPosition(ax1, [0.05, 0.2, 0.2, 0.05])
ax_reset.set_axes_locator(ip)
breset = Button(ax_reset, 'Reset')
breset.on_clicked(self.callback.reset)
plt.show()
ax.tick_params(bottom=False, labelbottom=False,
left=False, labelleft=False)
fig = plt.figure(figsize=(12, 6))
sfs = fig.subfigures(1, 2)
sfs[0].suptitle("Fixed axes sizes, fixed paddings")
# Sizes are in inches.
horiz = [Size.Fixed(1.), Size.Fixed(.5), Size.Fixed(1.5), Size.Fixed(.5)]
vert = [Size.Fixed(1.5), Size.Fixed(.5), Size.Fixed(1.)]
rect = (0.1, 0.1, 0.8, 0.8)
# Divide the axes rectangle into a grid with sizes specified by horiz * vert.
div = Divider(sfs[0], rect, horiz, vert, aspect=False)
# The rect parameter will actually be ignored and overridden by axes_locator.
ax1 = sfs[0].add_axes(rect, axes_locator=div.new_locator(nx=0, ny=0))
label_axes(ax1, "nx=0, ny=0")
ax2 = sfs[0].add_axes(rect, axes_locator=div.new_locator(nx=0, ny=2))
label_axes(ax2, "nx=0, ny=2")
ax3 = sfs[0].add_axes(rect, axes_locator=div.new_locator(nx=2, ny=2))
label_axes(ax3, "nx=2, ny=2")
ax4 = sfs[0].add_axes(rect, axes_locator=div.new_locator(nx=2, nx1=4, ny=0))
label_axes(ax4, "nx=2, nx1=4, ny=0")
sfs[1].suptitle("Scalable axes sizes, fixed paddings")
# Fixed sizes are in inches, scaled sizes are relative.
horiz = [Size.Scaled(1.5), Size.Fixed(.5), Size.Scaled(1.), Size.Scaled(.5)]
def plot_legend(self, legend_unit_height=0.25):
"""
Plot a legend for this Style object.
If the color and width labels are the same, a single legend will
be created. Otherwise, two legends will be created - one with
the color labels, and the other with the width labels.
Parameters
----------
legend_unit_height : float
A scaling factor to modify the height of each unit in the
legend only.
Returns
-------
fig : matplotlib Figure
Figure handle.
ax : matplotlib Axes
Axis handle.
"""
# print some plotting values
print('stratigraphic height scaling : 1 distance unit = 1 inch * {}'.
format(self.height_scaling_factor))
print('width value of 1 will be : {} inches'.format(
self.width_inches))
# extract attributes
color_labels = self.color_labels
width_labels = self.width_labels
color_values = self.color_values
width_values = self.width_values
# if the color and width labels are different
if np.any(~(color_labels == width_labels)):
# sort the widths
width_sort_inds = np.argsort(width_values)
width_labels = width_labels[width_sort_inds]
width_values = width_values[width_sort_inds]
# initialize the figure
fig, ax = plt.subplots(nrows=1, ncols=2, sharey=True)
# determine the axis height and limits first
if self.n_color_labels > self.n_width_labels:
ax_height = legend_unit_height * self.n_color_labels
ax[0].set_ylim(0, self.n_color_labels)
ax[1].set_ylim(0, self.n_color_labels)
else:
ax_height = legend_unit_height * self.n_width_labels
ax[0].set_ylim(0, self.n_width_labels)
ax[1].set_ylim(0, self.n_width_labels)
# plot the colors
strat_height_colors = 0
for i in range(len(color_labels)):
# create the rectangle - with thickness of 1
ax[0].add_patch(
Rectangle((0.0, strat_height_colors),
1,
1,
facecolor=color_values[i],
edgecolor='k'))
# label the unit
ax[0].text(1.2,
strat_height_colors + 0.5,
color_labels[i],
horizontalalignment='left',
verticalalignment='center')
# count the height
strat_height_colors = strat_height_colors + 1
# set the axis dimensions (values below are all in inches)
ax0_lower_left_x = 0.5
ax0_lower_left_y = 0.5
ax0_width = 0.5
h = [Size.Fixed(ax0_lower_left_x), Size.Fixed(ax0_width)]
v = [Size.Fixed(ax0_lower_left_y), Size.Fixed(ax_height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax[0].set_axes_locator(divider.new_locator(nx=1, ny=1))
# set the limits
ax[0].set_xlim(0, 1)
# prettify
ax[0].set_xticks([])
ax[0].set_yticklabels([])
ax[0].set_yticks([])
# plot the widths
strat_height_widths = 0
for i in range(len(width_labels)):
# create the rectangle
ax[1].add_patch(
Rectangle((0.0, strat_height_widths),
width_values[i],
1,
facecolor='grey',
edgecolor='k'))
# the label
ax[1].text(1.1,
strat_height_widths + 0.5,
width_labels[i],
horizontalalignment='left',
verticalalignment='center')
# count the height
strat_height_widths = strat_height_widths + 1
# set the axis dimensions (values below are all in inches)
ax1_lower_left_x = ax0_lower_left_x + ax0_width + 2
ax1_lower_left_y = ax0_lower_left_y
ax1_width = self.width_inches
h = [Size.Fixed(ax1_lower_left_x), Size.Fixed(ax1_width)]
v = [Size.Fixed(ax1_lower_left_y), Size.Fixed(ax_height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax[1].set_axes_locator(divider.new_locator(nx=1, ny=1))
# prettify
ax[1].set_xlim(0, 1)
ax[1].set_xticks([0, 0.2, 0.4, 0.6, 0.8, 1])
for label in ax[1].get_xticklabels():
label.set_rotation(270)
label.set_ha('center')
label.set_va('top')
ax[1].set_axisbelow(True)
ax[1].xaxis.grid(ls='--')
ax[1].set_yticklabels([])
ax[1].set_yticks([])
ax[1].spines['top'].set_visible(False)
ax[1].spines['right'].set_visible(False)
# turning the spines off creates some clipping mask issues
# so just turn the clipping masks off
for obj in fig.findobj():
obj.set_clip_on(False)
# not really necessary since the axis sizes are already
# forced, but we may as well set the figsize here too
fig.set_size_inches(
ax1_lower_left_x + ax1_width + ax0_lower_left_x,
ax1_lower_left_y * 2 + ax_height)
# if the color and width labels are the same
else:
# sort by width
width_sort_inds = np.argsort(width_values)
color_labels = color_labels[width_sort_inds]
width_labels = width_labels[width_sort_inds]
color_values = color_values[width_sort_inds]
width_values = width_values[width_sort_inds]
# initiate fig and ax
fig, ax = plt.subplots()
# determine the axis height and limits first
ax_height = legend_unit_height * self.n_color_labels
ax.set_ylim(0, self.n_color_labels)
# initiate counting of the stratigraphic height
strat_height = 0
# loop over each item
for i in range(len(color_labels)):
# create the rectangle - with thickness of 1
ax.add_patch(
Rectangle((0.0, strat_height),
width_values[i],
1,
facecolor=color_values[i],
edgecolor='k'))
# label the unit
ax.text(1.1,
strat_height + 0.5,
color_labels[i],
horizontalalignment='left',
verticalalignment='center')
# count the stratigraphic height
strat_height = strat_height + 1
# set the axis dimensions (values below are all in inches)
ax_lower_left_x = 0.5
ax_lower_left_y = 0.5
ax_width = self.width_inches
h = [Size.Fixed(ax_lower_left_x), Size.Fixed(ax_width)]
v = [Size.Fixed(ax_lower_left_y), Size.Fixed(ax_height)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
# prettify
ax.set_xlim(0, 1)
ax.set_xticks([0, 0.2, 0.4, 0.6, 0.8, 1])
for label in ax.get_xticklabels():
label.set_rotation(270)
label.set_ha('center')
label.set_va('top')
ax.set_axisbelow(True)
ax.xaxis.grid(ls='--')
ax.set_yticklabels([])
ax.set_yticks([])
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
# turning the spines off creates some clipping mask issues
# so just turn the clipping masks off
for obj in fig.findobj():
obj.set_clip_on(False)
# not really necessary since the axis sizes are already
# forced, but we may as well set the figsize here too
fig.set_size_inches(ax_lower_left_x * 2 + ax_width,
ax_lower_left_y * 2 + ax_height)
return fig, ax
def plot_variance_multi(savefile,
data,
labels,
colors,
ylabel,
ylim,
ylim_unit=1,
ylim_scale=5,
colors2=None,
hline=None,
vline=None,
locator=-1,
draw_dist=True,
subsample=1):
plt.clf()
plt.rcParams["font.size"] = 14
colors2 = colors if colors2 is None else colors2
m = len(data)
#fig, ax = plt.subplots(1, 1)
fig = plt.figure(1, figsize=(4, 4))
# The first items are for padding and the second items are for the axes.
# sizes are in inch.
h = [Size.Fixed(0.65), Size.Fixed(5.5)]
v = [Size.Fixed(0.7), Size.Fixed(4.)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
# the width and height of the rectangle is ignored.
ax = LocatableAxes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
if not hasattr(draw_dist, '__getitem__'):
v = draw_dist
draw_dist = [v for i in range(m)]
print(draw_dist)
for k in range(m)[::-1]:
plot_m = numpy.median(data[k], 0)
sorted = numpy.sort(data[k], 0)
n = data[k].shape[0]
T = data[k].shape[1]
t = numpy.arange(0, T)
fin = plot_m[-1]
best = sorted[0][-1]
alpha = 1.0 / (n // 2) * 0.4
if subsample > 1:
t = subsample_data(t, subsample)
plot_m = subsample_data(plot_m, subsample)
sorted = subsample_data(sorted, subsample)
if k == 0:
if ylim is None:
ymax = min([fin * ylim_scale, sorted[-1, 0], 90])
ymin = numpy.floor(numpy.max(best / ylim_unit - 0.5,
0)) * ylim_unit
ymin = 0
ylim = [ymin, numpy.ceil(ymax / ylim_unit) * ylim_unit]
else:
ylim = ylim
# 3: 0
if draw_dist[k]:
print('k', k)
for i in range(n // 2):
ax.fill_between(t,
sorted[-1 - i],
sorted[i],
facecolor=colors2[k],
alpha=alpha)
for k in range(m)[::-1]:
plot_m = numpy.median(data[k], 0)
t = numpy.arange(0, T)
if subsample > 1:
t = subsample_data(t, subsample)
plot_m = subsample_data(plot_m, subsample)
ax.plot(t, plot_m, lw=1, color=colors[k], ls='-', label=labels[k])
if hline is not None:
ax.plot((0, T), (hline, hline), color='gray', ls=':')
if vline is not None:
ax.plot((vline, vline), (ylim[0], ylim[1]), color='gray', ls=':')
ax.legend(loc='upper right')
ax.set_xlabel('Iterations')
ax.set_ylabel(ylabel)
ax.set_xlim([0, T])
ax.set_ylim(ylim)
if locator >= 0:
ax.yaxis.set_major_locator(MultipleLocator(locator))
ax.grid()
plt.savefig(savefile + '.png')
plt.savefig(savefile + '.pdf')
def __init__(self,
parent,
file_dialog_service,
h_margin=(0.8, 0.1),
v_margin=(0.5, 0.15),
h_axes=[Size.Scaled(1.0)],
v_axes=[Size.Scaled(1.0)],
nx_default=1,
ny_default=1):
QWidget.__init__(self, parent)
self._file_dialog_service = file_dialog_service
self._figure = Figure()
self._canvas = FigureCanvas(self._figure)
h = [Size.Fixed(h_margin[0]), *h_axes, Size.Fixed(h_margin[1])]
v = [Size.Fixed(v_margin[0]), *v_axes, Size.Fixed(v_margin[1])]
self._divider = Divider(self._figure, (0.0, 0.0, 1.0, 1.0),
h,
v,
aspect=False)
self._axes = LocatableAxes(self._figure, self._divider.get_position())
self._axes.set_axes_locator(
self._divider.new_locator(nx=nx_default, ny=ny_default))
self._axes.set_zorder(2)
self._axes.patch.set_visible(False)
for spine in ['top', 'right']:
self._axes.spines[spine].set_visible(False)
self._figure.add_axes(self._axes)
self._canvas.setParent(self)
self._layout = QVBoxLayout(self)
self._layout.setContentsMargins(0, 0, 0, 0)
self._layout.addWidget(self._canvas)
self.setLayout(self._layout)
self._figure.canvas.mpl_connect('scroll_event', self._on_scroll)
self._xy_extents = None
self._background_cache = None
self._decoration_artists = []
self._is_panning = False
self._zoom_selector = _RectangleSelector(self._axes,
self._zoom_selected)
self._zoom_selector.set_active(False)
self._x_extent_padding = 0.01
self._y_extent_padding = 0.01
self._axes.ticklabel_format(style='sci', axis='x', scilimits=(-4, 4))
self._axes.ticklabel_format(style='sci', axis='y', scilimits=(-4, 4))
self._active_tools = {}
self._span = _SpanSeletor(self._axes,
self._handle_span_select,
'horizontal',
rectprops=dict(alpha=0.2,
facecolor='red',
edgecolor='k'),
span_stays=True)
self._span.set_on_select_none(self._handle_span_select_none)
self.span = self._previous_span = None
self._span_center_mouse_event = None
self._span_left_mouse_event = None
self._span_right_mouse_event = None
self._figure.canvas.mpl_connect('button_press_event',
self._handle_press)
self._figure.canvas.mpl_connect('motion_notify_event',
self._handle_move)
self._figure.canvas.mpl_connect('button_release_event',
self._handle_release)
self._figure.canvas.mpl_connect('resize_event', self._handle_resize)
self.activateTool(ToolType.span, self.isActiveDefault(ToolType.span))
self._pan_event = None
self._pending_draw = None
self._pending_artists_draw = None
self._other_draw_events = []
self._draw_timer = QTimer(self)
self._draw_timer.timeout.connect(self._do_draw_events)
self._draw_timer.start(20)
self._zoom_skew = None
self._menu = QMenu(self)
self._copy_image_action = QAction(self.tr('Copy To Clipboard'), self)
self._copy_image_action.triggered.connect(self.copyToClipboard)
self._copy_image_action.setShortcuts(QKeySequence.Copy)
self._save_image_action = QAction(self.tr('Save As Image'), self)
self._save_image_action.triggered.connect(self.saveAsImage)
self._show_table_action = QAction(self.tr('Show Table'), self)
self._show_table_action.triggered.connect(self.showTable)
self._menu.addAction(self._copy_image_action)
self._menu.addAction(self._save_image_action)
self._menu.addAction(self._show_table_action)
self.addAction(self._copy_image_action)
self._table_view = None
self._single_axis_zoom_enabled = True
self._cached_label_width_height = None
if hasattr(type(self), 'dataChanged'):
self.dataChanged.connect(self._on_data_changed)
self._options_view = None
self._secondary_axes = self._secondary_y_extent = self._secondary_x_extent = None
self._legend = None
self._draggable_legend = None
self._setting_axis_limits = False
self.hasHiddenSeries = False
def _setup_image_grid(self, figure, dims):
"""
An example of how grid, parameter and view numbers change
for dims = 4
The numbers in the grid are:
parameter x, parameter y
grid x, grid y
-----------------------------------------------------
| | | | |
| 0, 0 | 1, 0 | 2, 0 | 3, 0 |
| -, - | -, - | -, - | -, - |
| | | | |
==============------------|------------|------------|
I I | | |
I 0, 1 I 1, 1 | 2, 1 | 3, 1 |
I 0, 0 I 1, 0 | 2, 0 | -, - |
I I | | |
I------------==============------------|------------|
I | I | |
I 0, 2 | 1, 2 I 2, 2 | 3, 2 |
I 0, 1 | 1, 1 I 2, 1 | -, - |
I | I | |
I------------|------------==============------------|
I | | I |
I 0, 3 | 1, 3 | 2, 3 I 3, 3 |
I 0, 2 | 1, 2 | 2, 2 I -, - |
I | | I |
I======================================I------------|
From the above it should be clear that:
parameter x = grid x
parameter y = grid y + 1
grid nr = grid y + grid x * (dims - 1)
view nr = grid nr - (grid nr / dims) * ((grid nr / dims) +1) / 2
"""
image_grid = ImageGrid(
figure,
111,
nrows_ncols=(dims - 1, dims - 1),
cbar_location="right",
cbar_mode="single",
# add_all=False,
aspect=False,
axes_pad=0.1,
direction="column")
rect = (0.1, 0.1, 0.8, 0.8)
horiz = [Size.Fixed(.1)
] + [Size.Scaled(1.), Size.Fixed(.1)] * max(dims - 1, 1) + [
Size.Fixed(0.15)
]
vert = [Size.Fixed(.1)
] + [Size.Scaled(1.), Size.Fixed(.1)] * max(dims - 1, 1)
# divide the axes rectangle into grid whose size is specified by horiz * vert
divider = Divider(figure, rect, horiz, vert) # , aspect=False)
# Helper to get the axis for par x and y:
def get_grid(parx, pary):
gridx = parx
gridy = pary - 1
return image_grid[gridy + gridx * (dims - 1)]
# Helper to get the grid locator for par x and par y
def get_locator(parx, pary):
gridx = parx
gridy = pary - 1
nx = 1 + gridx * 2
ny = 1 + (dims - gridy - 2) * 2
return divider.new_locator(nx=nx, ny=ny)
# Hide the unused plots & setup the used ones:
for parx, pary in itertools.product(list(range(self.num_params - 1)),
list(range(1, self.num_params))):
# Calculate the grid position:
ax = get_grid(parx, pary)
# Setup axes:
if pary <= parx:
ax.set_visible(False)
else:
ax.set_axes_locator(get_locator(parx, pary))
ax.set_visible(True)
return image_grid, divider, get_grid
ax3.plot(t, x3)
# <markdowncell>
# make_axes_locatable() gives one way to handle this, by giving the append_axes() method. Another way is by creating a Divider object and giving it the layout up front.
# <codecell>
from mpl_toolkits.axes_grid1 import Size, Divider
fig = plt.figure()
ax = [fig.add_axes([0.1, 0.1, 0.8, 0.8], label='%d' % i)
for i in range(3)] # Create 3 Axes to be sized later
horiz = [Size.Scaled(2), Size.Fixed(0.5), Size.Scaled(3), Size.Fixed(0.01)]
vert = [Size.Scaled(1), Size.Fixed(0.5), Size.Scaled(1)]
div = Divider(fig, (0.1, 0.1, 0.8, 0.8), horiz, vert, aspect=False)
ax[0].set_axes_locator(div.new_locator(nx=0, ny=0))
ax[1].set_axes_locator(div.new_locator(nx=2, ny=0))
ax[2].set_axes_locator(div.new_locator(nx=0, nx1=2, ny=2))
ax[0].plot(t, x1)
ax[1].plot(t, x2)
ax[2].plot(t, x3)
# <rawcell>
# One obvious application of this is for doing grid of plots with colorbars. For this use case we can go one step further and make use of the ImageGrid class.
# <codecell>
sumFinalAnalysis = sum(lstFinalAnalysis)
sumFinalAnalysisTE = np.matmul(-np.transpose(npTE), lstFinalAnalysis)
A = np.array([[ndados, sumTE], [sumTE, prodTE]])
B = np.array([[sumFinalAnalysis], [sumFinalAnalysisTE]])
invA = np.linalg.inv(A)
P = np.matmul(invA, B)
T2[indexValue] = (P[1][0])
for i, t in enumerate(T2):
T2[i] = math.inf if t[0] == 0 else 1000 / t[0]
T2Star = np.reshape(T2, (400, 400))
fig = plt.figure(figsize=(6, 6))
h = [Size.Fixed(1.0), Size.Fixed(5.)]
v = [Size.Fixed(1.0), Size.Fixed(5.)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
ax = Axes(fig, divider.get_position())
ax.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(ax)
plt.pcolor(np.flip(T2Star, 0), cmap=jet_combined, vmin=0, vmax=180)
toc = time.perf_counter()
print(f"Executed in {toc - tic:0.4f} seconds")
# toggle_selector.RS = RectangleSelector(ax, line_select_callback,
# drawtype='box', useblit=True,
# button=[1, 3], # don't use middle button
# minspanx=5, minspany=5,
# spancoords='pixels',
# interactive=True)
# plt.connect('key_press_event', toggle_selector)
def setup_axes(df, plates, attrs):
from mpl_toolkits.axes_grid1 import Divider
from mpl_toolkits.axes_grid1.axes_size import Fixed
# These assumptions let us simplify some code, and should always be true.
assert len(plates) > 0
assert len(attrs) > 0
# Determine how much data will be shown in the figure:
num_plates = len(plates)
num_attrs = len(attrs)
dims = Dimensions(df)
bar_label_width = guess_attr_label_width(df, attrs)
# Define the grid on which the axes will live:
h_divs = [
LEFT_MARGIN,
]
for _ in plates:
h_divs += [
CELL_SIZE * dims.num_cols,
PAD_WIDTH,
]
h_divs[-1:] = [
BAR_PAD_WIDTH,
BAR_WIDTH,
RIGHT_MARGIN + bar_label_width,
]
v_divs = [
TOP_MARGIN,
]
for attr in attrs:
v_divs += [
max(
CELL_SIZE * dims.num_rows,
BAR_WIDTH * dims.num_values[attr],
),
PAD_HEIGHT,
]
v_divs[-1:] = [
BOTTOM_MARGIN,
]
# Add up all the divisions to get the width and height of the figure:
figsize = sum(h_divs), sum(v_divs)
# Make the figure:
fig, axes = plt.subplots(
num_attrs,
num_plates + 1, # +1 for the colorbar axes.
figsize=figsize,
squeeze=False,
)
# Position the axes:
rect = 0.0, 0.0, 1, 1
h_divs = [Fixed(x) for x in h_divs]
v_divs = [Fixed(x) for x in reversed(v_divs)]
divider = Divider(fig, rect, h_divs, v_divs, aspect=False)
for i in range(num_attrs):
for j in range(num_plates + 1):
loc = divider.new_locator(nx=2 * j + 1, ny=2 * (num_attrs - i) - 1)
axes[i, j].set_axes_locator(loc)
return fig, axes, dims
def tpmc_svg(self, b64encode=True):
fig = pyplot.figure(figsize=self.FIGSIZE)
h = [Size.Fixed(1.2), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
plt = Axes(fig, divider.get_position())
plt.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(plt)
result = self.result
runinfo = result.runinfo
# ----
# The X limits are -rampupMins, runMins
# ----
plt.set_xlim(-int(runinfo['rampupMins']), int(runinfo['runMins']))
plt.axvspan(-int(runinfo['rampupMins']), 0, facecolor='0.2', alpha=0.1)
# ----
# offset the timestamps by -rampupMins so that the graph
# starts with negative minutes elapsed and switches to
# positive when the measurement begins.
# ----
offset = (int(runinfo['rampupMins'])) * 60.0
# ----
# NEW_ORDER transactions per minute. First get the timestamp
# and number of transactions from the result data.
# The X vector then is the sorted unique timestamps rounded
# to an interval.
# ----
interval = 10
data = numpy.array([[(int(tup[0] / interval) * interval - offset) / 60,
tup[1] * (60 / interval)]
for tup in result.result_ttype['NEW_ORDER']])
x = sorted(numpy.unique(data[:, 0]))
# ----
# The Y vector is the sums of transactions grouped by X
# ----
y = []
for ts in x:
tmp = data[numpy.where(data[:, 0] == ts)]
y.append(numpy.sum(tmp[:, 1]))
# ----
# Plot the NOPM and add all the decorations
# ----
plt.plot(x, y, 'b')
plt.set_title("NEW_ORDER Transactions per Minute (tpmC)")
plt.set_xlabel("Elapsed Minutes")
plt.set_ylabel("tpmC")
plt.grid()
buf = io.StringIO()
pyplot.savefig(buf, format='svg')
if not b64encode:
return buf.getvalue()
return base64.b64encode(buf.getvalue().encode('utf-8')).decode('utf-8')
# # if use_alpha:
# rgba = colors.hsv_to_rgb(hsv)
# rgba = np.dstack([rgba, alpha])
# # else:
# hsv[:, :, 2] = alpha.clip(min=0.2, max=value)
# rgba_l = colors.hsv_to_rgb(hsv)
# alpha_l = np.abs(1 - alpha).clip(min=value, max=1)
# hsv[:, :, 2] = alpha_l
# rgba_lr = colors.hsv_to_rgb(hsv)
cmap = colors.ListedColormap(cmap, name='test1')
fig = plt.figure(1, figsize=(12, 8))
h = [Size.Fixed(0.), Size.Fixed(6.5)]
v = [Size.Fixed(0.5), Size.Fixed(3.25)]
win = Divider(fig, (0.1, 0.1, 0.8, 0.8), h, v, aspect=False)
ax = Axes(fig, win.get_position())
ax.set_axes_locator(win.new_locator(nx=1, ny=1))
fig.add_axes(ax)
# fig = plt.gcf()
for ii in range(1, 2):
if ii == 0:
to_plot = rgb
title = 'Model'
elif ii == 1:
to_plot = rgba
title = 'Model + Resolution (Transparency)'
elif ii == 2:
to_plot = rgba_l
title = 'Model + Resolution (Lightness)'
else:
def memory_svg(self,
host=None,
title="Memory Usage",
unit="Bytes",
factor=1.0,
b64encode=True):
fig = pyplot.figure(figsize=self.FIGSIZE)
h = [Size.Fixed(1.2), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
plt = Axes(fig, divider.get_position())
plt.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(plt)
result = self.result
runinfo = result.runinfo
# ----
# The X limits are -rampupMins, runMins
# ----
plt.set_xlim(-int(runinfo['rampupMins']), int(runinfo['runMins']))
plt.axvspan(-int(runinfo['rampupMins']), 0, facecolor='0.2', alpha=0.1)
x, y_used = self._get_metric_xy({
'host': host,
'metric': 'memory.memory-used',
'factor': factor,
})
_, y_cached = self._get_metric_xy(
{
'host': host,
'metric': 'memory.memory-cached',
'factor': factor,
}, x)
_, y_buffered = self._get_metric_xy(
{
'host': host,
'metric': 'memory.memory-buffered',
'factor': factor,
}, x)
_, y_free = self._get_metric_xy(
{
'host': host,
'metric': 'memory.memory-free',
'factor': factor,
}, x)
# ----
# It is possible that the mqtt based metric collector produces
# CSV files with different numbers of lines. We cut all of them
# to a combined minimum length.
# ----
min_len = min(len(x), len(y_used), len(y_cached), len(y_buffered),
len(y_free))
x = x[:min_len]
y_used = y_used[:min_len]
y_cached = y_cached[:min_len]
y_buffered = y_buffered[:min_len]
y_free = y_free[:min_len]
plt.stackplot(x,
y_used,
y_cached,
y_buffered,
y_free,
colors=[
'b',
'y',
'g',
'gray',
],
labels=[
'Used',
'Cached',
'Buffered',
'Free',
])
# ----
# Title and labels
# ----
plt.set_title(title)
plt.set_xlabel("Elapsed Minutes")
plt.set_ylabel(unit)
plt.legend(loc='upper left')
plt.grid()
# ----
# Now turn this into an in-memory SVG and return it as requested
# (raw or b64-encoded)
# ----
buf = io.StringIO()
pyplot.savefig(buf, format='svg')
if not b64encode:
return buf.getvalue()
return base64.b64encode(buf.getvalue().encode('utf-8')).decode('utf-8')
data = extract_csvs(stacked_boxplot.visualization, './data/temp/csv')
level_4 = data[4]
# Filters the level 4 data to CD and UC
IBD_level_4 = level_4.loc[level_4['dx'].isin(["CD", "UC"])]
IBD_level_4 = IBD_level_4.drop(columns=metadata_columns)
IBD_level_4 = IBD_level_4.set_index('index')
# Normalise the data around 0
standardised = (IBD_level_4 - IBD_level_4.mean()) / IBD_level_4.std()
# sets the sizing so as much of the label can be shown as possible
fig = plt.figure(figsize=(11.7, 8.27))
h = [Size.Fixed(6.), Size.Scaled(.5), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(.5), Size.Fixed(.5)]
divider = Divider(fig, (0, 0, 1, 1), h, v, aspect=False)
ax = fig.add_axes(divider.get_position(),
axes_locator=divider.new_locator(nx=1, ny=1))
# Plots the relative frequency
sns.boxplot(y="variable", x="value", data=pd.melt(standardised), showfliers=False, width=.6, ax=ax)
ax.xaxis.grid(True)
ax.set(ylabel="Bacteria", xlabel="Relative Abundance",
title="Relative Abundance for IBD for taxa at level 4")
sns.despine(trim=True, left=True)
plt.savefig(out_individual_boxplot)
this_experiment = {
"_id": experiment_id,
"parent": parent_name,
"stage": CURRENT_STAGE,
def delay_avg_svg(self, ttype, b64encode=True):
fig = pyplot.figure(figsize=self.FIGSIZE)
h = [Size.Fixed(1.2), Size.Scaled(1.), Size.Fixed(.2)]
v = [Size.Fixed(0.7), Size.Scaled(1.), Size.Fixed(.5)]
divider = Divider(fig, (0.0, 0.0, 1., 1.), h, v, aspect=False)
plt = Axes(fig, divider.get_position())
plt.set_axes_locator(divider.new_locator(nx=1, ny=1))
fig.add_axes(plt)
result = self.result
runinfo = result.runinfo
# ----
# The X limits are -rampupMins, runMins
# ----
plt.set_xlim(-int(runinfo['rampupMins']), int(runinfo['runMins']))
plt.axvspan(-int(runinfo['rampupMins']), 0, facecolor='0.2', alpha=0.1)
# ----
# offset the timestamps by -rampupMins so that the graph
# starts with negative minutes elapsed and switches to
# positive when the measurement begins.
# ----
offset = (int(runinfo['rampupMins'])) * 60.0
# ----
# ttype transaction delay. First get the timestamp
# and delay numbers from the result data.
# The X vector then is the sorted unique timestamps rounded
# to an interval.
# ----
interval = 10
data = numpy.array([[(int(tup[0] / interval) * interval - offset) / 60,
tup[1], tup[5]]
for tup in result.result_ttype[ttype]])
x = sorted(numpy.unique(data[:, 0]))
# ----
# The Y vector is the sums of transactions delay divided by
# the sums of the count, grouped by X.
# ----
y = []
for ts in x:
tmp = data[numpy.where(data[:, 0] == ts)]
y.append(numpy.sum(tmp[:, 2]) / (numpy.sum(tmp[:, 1]) + 0.000001))
# ----
# Plot the ttype delay and add all the decorations
# ----
plt.plot(x, y, 'r', label='Delay')
# ----
# Now do the same aggregation for the latency
# ----
data = numpy.array([[(int(tup[0] / interval) * interval - offset) / 60,
tup[1], tup[2]]
for tup in result.result_ttype[ttype]])
# ----
# The Y vector is similar by based on latency
# ----
y = []
for ts in x:
tmp = data[numpy.where(data[:, 0] == ts)]
y.append(numpy.sum(tmp[:, 2]) / (numpy.sum(tmp[:, 1]) + 0.000001))
plt.plot(x, y, 'b', label='Latency')
plt.set_title("{} Average Latency and Delay".format(ttype))
plt.set_xlabel("Elapsed Minutes")
plt.set_ylabel("Latency/Delay in ms")
plt.legend(loc='upper left')
plt.grid()
buf = io.StringIO()
pyplot.savefig(buf, format='svg')
if not b64encode:
return buf.getvalue()
return base64.b64encode(buf.getvalue().encode('utf-8')).decode('utf-8')
