def __init__(self,
eigen_decomp,
year_axis=False,
fig=None,
ax=None,
interactive=False,
Title=None,
figsize=(6, 4)):
"""
Initialize the plot widget
:param: eigen_decomp: An Eigen_Decomp object
:param: year_axis: set to true if X axis should correspond to the months of the year.
"""
self.eigen_decomp = eigen_decomp
self.interactive = interactive
self.fig = fig
self.ax = ax
#print('in recon_plot.__init__',ax)
self.Title = Title
self.figsize = figsize
self.i = 0
self.year_axis = year_axis
self.yearPlotter = None
if year_axis:
self.yearPlotter = YearPlotter()
if not self.interactive:
self.plot_combination(**self.eigen_decomp.coeff)
return None
def __init__(self, eigen_decomp, year_axis=False):
"""
Initialize the plot widget
:param: eigen_decomp: An Eigen_Decomp object
:param: year_axis: set to true if X axis should correspond to the months of the year.
"""
self.eigen_decomp = eigen_decomp
#self.C=eigen_decomp.C
#self.coeff=eigen_decomp.coeff
#self.mean=eigen_decomp.mean
self.year_axis = year_axis
self.yearPlotter = None
if year_axis:
self.yearPlotter = YearPlotter()
self.plot_combination(**self.eigen_decomp.coeff)
return None
class recon_plot:
"""A class for creating an interactive demonstration of approximating
a function with an orthonormal set of function"""
def __init__(self, eigen_decomp, year_axis=False):
"""
Initialize the plot widget
:param: eigen_decomp: An Eigen_Decomp object
:param: year_axis: set to true if X axis should correspond to the months of the year.
"""
self.eigen_decomp = eigen_decomp
#self.C=eigen_decomp.C
#self.coeff=eigen_decomp.coeff
#self.mean=eigen_decomp.mean
self.year_axis = year_axis
self.yearPlotter = None
if year_axis:
self.yearPlotter = YearPlotter()
self.plot_combination(**self.eigen_decomp.coeff)
return None
def get_widgets(self):
"""return the slider widget that are to be used
:returns: widget_list: the list of widgets in order
widget_dict: a dictionary of the widget to be used in `interact
:todo: make the sliders smaller: http://ipywidgets.readthedocs.io/en/latest/examples/Widget%20Styling.html
"""
coeff = self.eigen_decomp.C
widge_dict = {}
widge_list = []
for i in range(self.eigen_decomp.n):
if coeff[i] > 0:
r = [0, coeff[i] * 2]
else:
r = [coeff[i] * 2, 0]
widge_list.append(widgets.FloatSlider(min=r[0],max=r[1],step=coeff[i]/10.,\
value=coeff[i],orientation='vertical',decription='v'+str(i)))
widge_dict['c' + str(i)] = widge_list[-1]
return widge_list, widge_dict
def plot(self, y, label=''):
if self.year_axis:
self.yearPlotter.plot(y, self.fig, self.ax, label=label)
else:
self.ax.plot(self.eigen_decomp.x, y, label=label)
def plot_combination(self, **coeff):
"""the plotting function that is called by `interactive`
generates the plot according the the parameters set by the sliders
:returns: None
"""
A = self.eigen_decomp.mean
self.fig = plt.figure(figsize=(8, 6))
self.ax = self.fig.add_axes([0, 0, 1, 1])
self.plot(A, label='mean')
for i in range(self.eigen_decomp.n):
g = self.eigen_decomp.v[i] * coeff['c' + str(i)]
A = A + g
self.plot(A, label='c' + str(i))
self.plot(self.eigen_decomp.f, label='target')
self.ax.grid(figure=self.fig)
self.ax.legend()
return None
class recon_plot:
"""A class for creating an interactive demonstration of approximating
a function with an orthonormal set of function"""
def __init__(self,
eigen_decomp,
year_axis=False,
fig=None,
ax=None,
interactive=False,
Title=None,
figsize=(6, 4)):
"""
Initialize the plot widget
:param: eigen_decomp: An Eigen_Decomp object
:param: year_axis: set to true if X axis should correspond to the months of the year.
"""
self.eigen_decomp = eigen_decomp
self.interactive = interactive
self.fig = fig
self.ax = ax
#print('in recon_plot.__init__',ax)
self.Title = Title
self.figsize = figsize
self.i = 0
self.year_axis = year_axis
self.yearPlotter = None
if year_axis:
self.yearPlotter = YearPlotter()
if not self.interactive:
self.plot_combination(**self.eigen_decomp.coeff)
return None
def get_Interactive(self):
widge_list, widge_dict = self.get_widgets()
w = interactive(self.plot_combination, **widge_dict)
self.Title = 'Best reconstruction'
self.plot_combination(**self.eigen_decomp.coeff)
self.Title = 'Interactive reconstruction'
return widgets.VBox([widgets.HBox(widge_list), w.children[-1]])
def get_widgets(self):
"""return the slider widget that are to be used
:returns: widget_list: the list of widgets in order
widget_dict: a dictionary of the widget to be used in `interact
:todo: make the sliders smaller: http://ipywidgets.readthedocs.io/en/latest/examples/Widget%20Styling.html
"""
coeff = self.eigen_decomp.C
widge_dict = {}
widge_list = []
for i in range(self.eigen_decomp.n):
if coeff[i] > 0:
r = [0, coeff[i] * 2]
else:
r = [coeff[i] * 2, 0]
widge_list.append(widgets.FloatSlider(min=r[0],max=r[1],step=(r[1]-r[0])/10.,\
value=coeff[i],orientation='vertical',decription='v'+str(i)))
widge_dict['c' + str(i)] = widge_list[-1]
return widge_list, widge_dict
def plot(self, y, label=''):
if self.year_axis:
self.yearPlotter.plot(y, self.fig, self.ax, label=label)
else:
self.ax.plot(self.eigen_decomp.x, y, label=label)
def plot_combination(self, **coeff):
"""the plotting function that is called by `interactive`
generates the plot according the the parameters set by the sliders
:returns: None
"""
#print(self.i,coeff)
#self.i+=1
#return None
if self.interactive or self.fig is None:
#print('setting axis in plot_combination')
self.fig = plt.figure(figsize=self.figsize)
self.ax = self.fig.add_axes([0, 0, 1, 1])
A = self.eigen_decomp.mean
self.plot(A, label='mean')
for i in range(self.eigen_decomp.n):
#print('in plot_combinations, U.shape=',self.eigen_decomp.U.shape)
g = self.eigen_decomp.U[:, i] * coeff['c' + str(i)]
A = A + g
self.plot(A, label='c' + str(i))
self.plot(self.eigen_decomp.f, label='target')
self.ax.grid(figure=self.fig)
self.ax.legend()
self.ax.set_title(self.Title)
if self.interactive:
plt.show() # This hack was found by trial and error. I
# don't know why it works! But using the same
# "show" for both interactive and
# non-interactive creates problems.
else:
self.fig.show()
return None