def _dataframe_changed(self, old, new):
''' Handles how updates occur when dataframe changes. Evaluates if columns or columnlabels
have been changed. Provides entry condition as well.
Note: New automatically sets self.dataframe, so when I refer to new, I am actually
referring to self.dataframe. Using "new" instead of self.dataframe is just for readability'''
### Initialize plot first time dataframe is passed into the class. Boolean listeners
### for dataframe behave oddly, so uses self.plotdata for entry condition.
if not self.plotdata:
self.plotdata = PandasPlotData(df=new)
self.originaldata = new
### Try to infer plot title from dataframe name ###
try:
self.plot_title = new.name
except AttributeError:
pass
### Draw barebones of plot
self.draw_plot()
### Draw lines
self._draw_lines()
return
### Decide to update columns or completely redraw dataframe.
else:
labelold = self._getlabelarray(old)
labelnew = self._getlabelarray(new)
### Have columns been added or removed?
if len(labelold) != len(labelnew):
self._overwrite_plotdata
### Has index along primaryaxis changed any?
### Pandas index comparison is a bit tricky so just list conver
elif list(labelold) != list(labelnew):
self._overwrite_plotdata
else:
print 'updating frame'
self.plotdata.set_df(new)
def _dataframe_changed(self, old, new):
''' Handles how updates occur when dataframe changes. Evaluates if columns or columnlabels
have been changed. Provides entry condition as well.
Note: New automatically sets self.dataframe, so when I refer to new, I am actually
referring to self.dataframe. Using "new" instead of self.dataframe is just for readability'''
### Initialize plot first time dataframe is passed into the class. Boolean listeners
### for dataframe behave oddly, so uses self.plotdata for entry condition.
if not self.plotdata:
self.plotdata=PandasPlotData(df=new)
self.originaldata=new
### Try to infer plot title from dataframe name ###
try:
self.plot_title=new.name
except AttributeError:
pass
### Draw barebones of plot
self.draw_plot()
### Draw lines
self._draw_lines()
return
### Decide to update columns or completely redraw dataframe.
else:
labelold=self._getlabelarray(old)
labelnew=self._getlabelarray(new)
### Have columns been added or removed?
if len(labelold) != len(labelnew):
self._overwrite_plotdata
### Has index along primaryaxis changed any?
### Pandas index comparison is a bit tricky so just list conver
elif list(labelold) != list(labelnew):
self._overwrite_plotdata
else:
print 'updating frame'
self.plotdata.set_df(new)
class PandasPlot(HasTraits):
'''Plot to interface a pandas dataframe'''
### Actually
originaldata = Instance(DataFrame)
dataframe = Instance(DataFrame)
primaryaxis = Enum(0, 1, 'index',
'columns') #Integer or string, defaults to 0
_primaryaxis = Property(Enum(0, 1), depends_on='primaryaxis') #0,1
indexname = Str('Index')
columnname = Str('Column')
### Traits of the actual plot and plotdatasource handler ###
plot = Instance(Plot)
plotdata = Instance(
PandasPlotData) #Data stored at any given time in the plot
### Aesthetic traits
plot_title = Str('') #Defaults from dataframe name
linecolor = Str('black')
pointcolor = Str('red')
linestyle = Enum('line', 'scatter', 'both')
markersize = Range(low=1, high=5)
### Traits for testing
df_new = Button
df_change = Button
re_bin = Range(1, 100)
def _dataframe_changed(self, old, new):
''' Handles how updates occur when dataframe changes. Evaluates if columns or columnlabels
have been changed. Provides entry condition as well.
Note: New automatically sets self.dataframe, so when I refer to new, I am actually
referring to self.dataframe. Using "new" instead of self.dataframe is just for readability'''
### Initialize plot first time dataframe is passed into the class. Boolean listeners
### for dataframe behave oddly, so uses self.plotdata for entry condition.
if not self.plotdata:
self.plotdata = PandasPlotData(df=new)
self.originaldata = new
### Try to infer plot title from dataframe name ###
try:
self.plot_title = new.name
except AttributeError:
pass
### Draw barebones of plot
self.draw_plot()
### Draw lines
self._draw_lines()
return
### Decide to update columns or completely redraw dataframe.
else:
labelold = self._getlabelarray(old)
labelnew = self._getlabelarray(new)
### Have columns been added or removed?
if len(labelold) != len(labelnew):
self._overwrite_plotdata
### Has index along primaryaxis changed any?
### Pandas index comparison is a bit tricky so just list conver
elif list(labelold) != list(labelnew):
self._overwrite_plotdata
else:
print 'updating frame'
self.plotdata.set_df(new)
### Properties ###
def _get__primaryaxis(self):
''' Converts 'columns' or 'index' to 0,1. Mask basically'''
if self.primaryaxis == 'columns' or self.primaryaxis == 0:
return 0
else:
return 1
def _getlabelarray(self, dataframe):
''' Conveience method to get label along working dimension of a dataframe.'''
if self._primaryaxis == 0:
return np.asarray(dataframe.columns)
else:
return np.asarray(dataframe.index)
def _df_change_fired(self):
''' set data iteratively column by column'''
self.dataframe = DataFrame((np.random.randn(100, 100)))
def _df_new_fired(self):
df2 = DataFrame((np.random.randn(100, 100)),
columns=[i * 5 for i in range(10)],
index=[i * 50 for i in range(10)])
self.dataframe = df2
self._draw_lines()
def _re_bin_changed(self):
## Fix, squashing indicies
#self.dataframe=rebin(self.dataframe, self.re_bin, axis=0, avg_fcn='mean')
self.dataframe = rolling_mean(self.originaldata, self.re_bin)
def _draw_lines(self):
''' Recasts plotdata and ties it to the plot object.'''
for idx, name in enumerate(self.plotdata.list_data(as_strings=True)):
# if idx % 50==0:
self.plot.plot((self.indexname, name))
def draw_plot(self, toolbar=True, **pltkws):
''' Draw bare plot, including main plotting area, toolbar, etc...
either at initialization or global redo'''
if toolbar:
plot = ToolbarPlot(self.plotdata, **pltkws)
else:
plot = Plot(self.plotdata, **pltkws)
# plot.active_tool = RangeSelection(plot)
# plot.overlays.append(RangeSelectionOverlay(component=plot))
plot.title = self.plot_title
plot.padding = 50
plot.legend.visible = False
plot.tools.append(PanTool(plot))
zoom = BetterSelectingZoom(component=plot,
tool_mode="box",
always_on=False)
plot.overlays.append(zoom)
#indexlabels=[str(round(i,1)) for i in self.dataframe.index]
#### If I do plot.index_axis, it actually removes the default values.
#index_axis=LabelAxis(plot, orientation='top',
## positions=range(int(float(indexlabels[0])),
## int(float(indexlabels[-1]))),
#positions=range(0,1000,100),
#labels=indexlabels,#, resizable='hv',
#title=self.indexname)
#plot.underlays.append(index_axis)
#plot.value_axis= LabelAxis(plot, orientation='left', positions=range(self.t_axis_samples),
#mainlabels=['t1', 't2', 't3', 't4','t5', 't6'], resizable='hv',
#title=self.t_axis_title)
self.plot = plot
return
def _overwrite_plotdata(self):
'''When a new instance of PandasPlotData is created, this overwrites the
data source and updates the axis values.'''
self.plotdata = PandasPlotData(dataframe=self.dataframe)
#self.plot.index_axis=LabelAxis(self.plot, orientation='bottom', positions=[0,4,42],
#labels=self.plotdata._rowmasks.keys(), resizable='hv',
#title=self.indexname)
#plot.value_axis= LabelAxis(plot, orientation='left', positions=range(self.t_axis_samples),
#mainlabels=['t1', 't2', 't3', 't4','t5', 't6'], resizable='hv',
#title=self.t_axis_title)
#axis_traits_group=VGroup(
#HGroup(
#Item('x_axis_title'), Item('t_axis_title'), Item('plot_title'),
#),
#HGroup(
#Item('markersize'),Item('linestyle'),
#),
#)
#sample_group=Group(
#HGroup(Item('spacing'), Item('samp_style') )
#)
main_group = Group(
Item('plot', editor=ComponentEditor(), show_label=False),
# Item('df_new'), Item('df_change'),
Item('re_bin', style='simple'),
#Include('sample_group'),
#Include('axis_traits_group')
)
traits_view = View(Include('main_group'))
def _overwrite_plotdata(self):
'''When a new instance of PandasPlotData is created, this overwrites the
data source and updates the axis values.'''
self.plotdata = PandasPlotData(dataframe=self.dataframe)
class PandasPlot(HasTraits):
'''Plot to interface a pandas dataframe'''
### Actually
originaldata=Instance(DataFrame)
dataframe=Instance(DataFrame)
primaryaxis=Enum(0,1,'index','columns') #Integer or string, defaults to 0
_primaryaxis=Property(Enum(0,1), depends_on='primaryaxis') #0,1
indexname=Str('Index')
columnname=Str('Column')
### Traits of the actual plot and plotdatasource handler ###
plot=Instance(Plot)
plotdata=Instance(PandasPlotData) #Data stored at any given time in the plot
### Aesthetic traits
plot_title=Str('') #Defaults from dataframe name
linecolor=Str('black')
pointcolor=Str('red')
linestyle=Enum('line', 'scatter', 'both')
markersize=Range(low=1, high=5)
### Traits for testing
df_new=Button
df_change=Button
re_bin=Range(1,100)
def _dataframe_changed(self, old, new):
''' Handles how updates occur when dataframe changes. Evaluates if columns or columnlabels
have been changed. Provides entry condition as well.
Note: New automatically sets self.dataframe, so when I refer to new, I am actually
referring to self.dataframe. Using "new" instead of self.dataframe is just for readability'''
### Initialize plot first time dataframe is passed into the class. Boolean listeners
### for dataframe behave oddly, so uses self.plotdata for entry condition.
if not self.plotdata:
self.plotdata=PandasPlotData(dataframe=new, primaryaxis=self.primaryaxis, primaryname=self.columnname,\
secondaryname=self.indexname, add_labels_to_extras=True)
self.originaldata=new
### Try to infer plot title from dataframe name ###
try:
self.plot_title=new.name
except AttributeError:
pass
### Draw barebones of plot
self.draw_plot()
### Draw lines
self._draw_lines()
return
### Decide to update columns or completely redraw dataframe.
else:
labelold=self._getlabelarray(old)
labelnew=self._getlabelarray(new)
### Have columns been added or removed?
if len(labelold) != len(labelnew):
self._overwrite_plotdata
### Has index along primaryaxis changed any?
### Pandas index comparison is a bit tricky so just list conver
elif list(labelold) != list(labelnew):
self._overwrite_plotdata
else:
print 'updating frame'
self.plotdata.update_dataframe(new)
### Properties ###
def _get__primaryaxis(self):
''' Converts 'columns' or 'index' to 0,1. Mask basically'''
if self.primaryaxis=='columns' or self.primaryaxis==0:
return 0
else:
return 1
def _getlabelarray(self, dataframe):
''' Conveience method to get label along working dimension of a dataframe.'''
if self._primaryaxis==0:
return np.asarray(dataframe.columns)
else:
return np.asarray(dataframe.index)
def _df_change_fired(self):
''' set data iteratively column by column'''
self.dataframe=DataFrame((np.random.randn(100,100)))
def _df_new_fired(self):
df2=DataFrame((np.random.randn(100,100)), columns=[i*5 for i in range(10)], index=[i*50 for i in range(10)])
self.dataframe=df2
self._draw_lines()
def _re_bin_changed(self):
## Fix, squashing indicies
#self.dataframe=rebin(self.dataframe, self.re_bin, axis=0, avg_fcn='mean')
self.dataframe=rolling_mean(self.originaldata, self.re_bin)
def _draw_lines(self):
''' Recasts plotdata and ties it to the plot object.'''
for idx, name in enumerate(self.plotdata.list_data(as_strings=True)):
# if idx % 50==0:
self.plot.plot((self.indexname, name))
def draw_plot(self, toolbar=True, **pltkws):
''' Draw bare plot, including main plotting area, toolbar, etc...
either at initialization or global redo'''
if toolbar:
plot=ToolbarPlot(self.plotdata, **pltkws)
else:
plot=Plot(self.plotdata, **pltkws)
plot.title = self.plot_title
plot.padding = 50
plot.legend.visible=False
plot.tools.append(PanTool(plot))
zoom=BetterSelectingZoom(component=plot, tool_mode="box", always_on=False)
plot.overlays.append(zoom)
#indexlabels=[str(round(i,1)) for i in self.dataframe.index]
#### If I do plot.index_axis, it actually removes the default values.
#index_axis=LabelAxis(plot, orientation='top',
## positions=range(int(float(indexlabels[0])),
## int(float(indexlabels[-1]))),
#positions=range(0,1000,100),
#labels=indexlabels,#, resizable='hv',
#title=self.indexname)
#plot.underlays.append(index_axis)
#plot.value_axis= LabelAxis(plot, orientation='left', positions=range(self.t_axis_samples),
#mainlabels=['t1', 't2', 't3', 't4','t5', 't6'], resizable='hv',
#title=self.t_axis_title)
self.plot=plot
return
def _overwrite_plotdata(self):
'''When a new instance of PandasPlotData is created, this overwrites the
data source and updates the axis values.'''
self.plotdata=PandasPlotData(dataframe=self.dataframe)
#self.plot.index_axis=LabelAxis(self.plot, orientation='bottom', positions=[0,4,42],
#labels=self.plotdata._rowmasks.keys(), resizable='hv',
#title=self.indexname)
#plot.value_axis= LabelAxis(plot, orientation='left', positions=range(self.t_axis_samples),
#mainlabels=['t1', 't2', 't3', 't4','t5', 't6'], resizable='hv',
#title=self.t_axis_title)
#axis_traits_group=VGroup(
#HGroup(
#Item('x_axis_title'), Item('t_axis_title'), Item('plot_title'),
#),
#HGroup(
#Item('markersize'),Item('linestyle'),
#),
#)
#sample_group=Group(
#HGroup(Item('spacing'), Item('samp_style') )
#)
main_group=Group(
Item('plot', editor=ComponentEditor(), show_label=False),
# Item('df_new'), Item('df_change'),
Item('re_bin', style='simple'),
#Include('sample_group'),
#Include('axis_traits_group')
)
traits_view=View( Include('main_group') )
def _overwrite_plotdata(self):
'''When a new instance of PandasPlotData is created, this overwrites the
data source and updates the axis values.'''
self.plotdata=PandasPlotData(dataframe=self.dataframe)