def run(self,*event):
#I tried for WAY to long to get this to work with the high-level interface
#but it usually resulted in the output having a very long label that
#contained all the data in the vector we passed to it, still not sure
#how to resolve it
args = rdict()
args['mu'] = self.parameters['Mu'].get()
args['conf.level'] = float( self.parameters['Confidence Level'].get() )
#args['paired'] = self.parameters['Paired'].get()
args['var.equal'] = self.parameters['Equal Variances'].get()
#R takes the first character as an argument
args['alternative'] = self.parameters['Alternative Hypothesis'].get().lower()[0]
args = rsession.arguments_to_string(args)
print args
s1lab,s1data = self.custom_variables['samples'][0]
if len(self.custom_variables['samples'])>1:
s2lab,s2data = self.custom_variables['samples'][1]
x = self.active_r_object[s1lab]
y = self.active_r_object[s2lab]
anova_output = rsession.r('t.test(x=%s,y=%s,data=%s,%s)' % (x, y, self.active_r_object.label, args))
else:
x = self.active_r_object[s1lab]
anova_output = rsession.r('t.test(x=%s,data=%s,%s)' % (x, self.active_r_object.label, args))
name = 'T-Test ' + str(self.active_r_object.label)
self.parent.render_description(anova_output,label=name)
def run(self,*event):
#I tried for WAY to long to get this to work with the high-level interface
#but it usually resulted in the output having a very long label that
#contained all the data in the vector we passed to it, still not sure
#how to resolve it
args = rdict()
args['correct'] = self.parameters['Apply continuity correction'].get()
args = rsession.arguments_to_string(args)
s1lab,s1data = self.custom_variables['samples'][0]
if len(self.custom_variables['samples'])==2:
s2lab,s2data = self.custom_variables['samples'][1]
x = self.active_r_object[s1lab]
y = self.active_r_object[s2lab]
anova_output = rsession.r('mcnemar.test(x=%s,y=%s%s)' % (x, y, args))
else:
info('Please select two samples.')
name = 'T-Test ' + str(self.active_r_object.label)
self.parent.render_description(anova_output,label=name)
def plot_fit(self,event,image=None):
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
#These return the parent$column strings
x = self.active_r_object[ilab]
y = self.active_r_object[dlab]
fit = rsession.r('lm(%s~poly(%s,%i))' % (y,x,self.degree.get_value()))
self.result = fit
data={'x':idata,'y':ddata}
args={'xlab':ilab,'ylab':dlab}
plotwindow = rsession.RPlotThread(data=data,
args=args,
par_mode=(1,1),
export=image,
type='scatter')
plotwindow.start()
seq=rsession.r['seq']
min=rsession.r['min']
max=rsession.r['max']
object = rsession.env[self.active_r_object.label]
range = seq(min(object), max(object), length_out = 100)
predicted = rsession.r['predict'](fit,data_frame=range)
plotwindow.add_cmd(rsession.ro.r.lines,idata,predicted,col='red')
self.active_threads.append(plotwindow)
def plot_fit(self,event,image=None):
if not self.variables_are_set():
error('Required variables are not set.')
return
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
x = self.active_r_object[ilab]
y = self.active_r_object[dlab]
fit = rsession.r('lm(%s ~ %s,data=%s)' % (y,x,self.active_r_object.label))
self.result = fit
data={'x':idata,'y':ddata}
args={'xlab':ilab,'ylab':dlab}
plotwindow = rsession.RPlotThread(data=data,
args=args,
type='scatter',
export=image,
par_mode=(1,1))
plotwindow.add_cmd(rsession.r['abline'],fit,col='red')
plotwindow.start()
self.active_threads.append(plotwindow)
def freqtable(self,event):
condition = variable_prompt.VariablePrompt(['Condition(s):']).get_inputs()
#This is dirty but ro.RFormula will not parse
try:
filter = rsession.r('with(%s,table(%s))' % (self.active_robject.label,condition[0]))
except:
error('Formula is not valid. \n Enter conditions seperated by commas. Ex: V1>3,V4<1')
return
self.output(filter,title='Frequency Table: ' + condition[0])
self.active_output = rsession.description(filter,tabelize=False)
def export_latex(self,event):
try:
rsession.r('library(xtable)')
rprint = rsession.r['print']
xtable = rsession.r['xtable']
table = rsession.r['table']
activetab = self.builder.get_object('main_tabview').get_current_page()
#Data Tab
if activetab == 0:
if not self.active_robject:
error('No active dataframe.')
return
object = self.active_robject.object
label = self.active_robject.label
text = str(xtable(object,caption=label))
TextOutput(text)
#Output Tab
elif activetab == 1:
if not self.active_output:
error('No active output.')
return
object = self.active_output.object
label = self.active_output.label
if not label:
label = rsession.null
if self.active_output.tabelize:
text = str(xtable(table(object),caption=label))
else:
text = str(xtable(object,caption=label))
TextOutput(text)
#Console Tab
elif activetab == 2:
console = self.builder.get_object('consoleview').get_buffer()
lower,upper = console.get_bounds()
text = console.get_text(lower,upper)
TextOutput('\\begin{verbatim}\n' + text + '\n\end{verbatim}')
except:
error("Could not load the xtable library.")
def plot_fit(self,event,image=None):
if not self.variables_are_set():
error('Required variables are not set.')
return
rsession.r('par(mfrow=c(1,1))')
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
#fit = rsession.r['loess'](idata,ddata).object
data={'x':idata,'y':ddata}
args={'xlab':ilab,'ylab':dlab}
plotwindow = rsession.RPlotThread(data=data,
args=args,
type='scatter.smooth')
plotwindow.start()
self.active_threads.append(plotwindow)
def plot_summary(self,event,image=None):
if not self.variables_are_set():
error('Required variables are not set.')
return
rsession.r('par(mfrow=c(2,2))')
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
x = self.active_r_object[ilab]
y = self.active_r_object[dlab]
fit = rsession.r('lm(%s ~ %s,data=%s)' % (x,y,self.active_r_object.label))
self.result = fit
plotwindow = rsession.RPlotThread(data=fit,
type='general',
par_mode=True)
plotwindow.start()
self.active_threads.append(plotwindow)
def transform(self,event):
cols = self.active_robject.columns.keys()
transforms = zip(cols,variable_prompt.VariablePrompt(cols).get_inputs(allow_empty=True))
try:
args = rdict()
argument_string = ''
for col,transform in transforms:
if col and transform:
argument_string += ',%s=%s' % (col,transform)
print argument_string
label = self.active_robject.label
obj = rsession.env[label]
transformed = rsession.r('transform(%s %s)' % (label,argument_string))
#transformed = rsession.r['transform'](obj,args)
except:
error('Transformations were not well formed')
return
rsession.env[label] = transformed
self.sync_with_r()
def plot_summary(self,event,image=None):
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
#These return the parent$column strings
x = self.active_r_object[ilab]
y = self.active_r_object[dlab]
fit = rsession.r('lm(%s~poly(%s,%i))' % (y,x,self.degree.get_value()))
self.result = fit
plotwindow = rsession.RPlotThread(data=fit,
type='general',
par_mode=(2,2),
export=image)
plotwindow.start()
self.active_threads.append(plotwindow)
def run(self,*event):
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
x = self.active_r_object[ilab]
y = self.active_r_object[dlab]
fit = rsession.r('lm(%s ~ %s,data=%s)' % (y,x,self.active_r_object.label))
if idata == ddata:
error('Variables are identical, results may not be accurate.')
anova_output = rsession.r['anova'](fit)
name = 'ANOVA ' + str(self.active_r_object.label)
#This isn't good if there are variables called x,y in the workspace
rsession.rm('x')
rsession.rm('y')
self.parent.render_description(anova_output,label=name)
def plot(self,event,image=None):
_xlab,_ylab,_title,_xmin,_ymin,_xmax,_ymax = self.fetch_labels()
data = self.data
args = rdict()
if self.parameters['Lower Panels'].get():
args['lower.panel'] = rsession.r('panel.smooth')
else:
args['lower.panel'] = rsession.null
args['gap'] = self.parameters['Plot Gap'].get()
args['pch'] = self.get_point_style()
#Advanced Parameters
plotwindow = rsession.RPlotThread(data=data,
args=args,
type='matplot',
export=image)
plotwindow.start()
self.active_threads.append(plotwindow)
def context_delcol(self,*args):
#Its much easier to resolve this on the R side, rather than using numpy
rsession.r('%s = transform(%s,%s)' % (self.active_robject.label, self.active_robject.label,self.active_column+'=NULL'))
self.sync_with_r()
def subset(self,event):
condition = variable_prompt.VariablePrompt(['Condition(s):']).get_inputs()[0]
rsession.env['subset %s' % (self.active_robject.label)] = rsession.r('subset(%s,%s)' % (self.active_robject.label,condition))
self.sync_with_r()
def shutdown_cleanly(self):
'''If the user CTRL-Cs, make sure the threads die and graphics get turned off'''
#iterate over the shared state of ThreadHandler to fetch all threads and kill them
for thread in self.active_threads.iter_shared():
thread.halt = True
rsession.r('graphics.off()')
def plot_fit(self,event,image=None):
model = self.get_model().lower()
dependent = model.partition('~')[0]
independent = model.partition('~')[2]
coefs = set()
vars = set()
#We're iterating over the variables, so that later it will be easier to do this in more than 2D
for symbol in model:
if symbol is 'x' or symbol is 'y':
vars.add(symbol)
elif symbol.isalpha():
coefs.add(symbol)
ilab,idata = self.custom_variables['independent'].get_first()
dlab,ddata = self.custom_variables['dependent'].get_first()
try:
fmla = rsession.ro.RFormula(model)
except:
error('The model you specified is not valid.')
return
env = fmla.getenvironment()
env['x'] = idata
env['y'] = ddata
#I love R, this would take like 30 hours to implement in python
#this fetches all coefficents but ignores functions like sin,cos...
coefs = list ( rsession.r['all.vars'](fmla) )
#Don't consider x and y as coefficents
coefs.remove('y')
coefs.remove('x')
data={'x':idata,'y':ddata}
args={'xlab':ilab,'ylab':dlab}
object = rsession.r(self.active_r_object.label)
startdict = {}
for i in coefs:
startdict[i]= 1
startlist=rsession.r['list'](**startdict)
seq=rsession.r['seq']
rmin=rsession.r['min']
rmax=rsession.r['max']
# There is nothing truly beautiful but that which can never be of any use whatsoever; everything
# useful is ugly; keep that in mine while reading the next 40 lines.
#Generate a range of points so our predicted curve is smooth
lenout = {'length.out':500}
df = rsession.r['data.frame']
smooth_range = rsession.r['seq']( rsession.r['min'](idata) , rsession.r['max'](idata),**lenout)
fit_found = False
predicted = None
#Try and find the fit, loop using the user provided starting values until we find a fit
while(fit_found == False):
try:
#Try to find a fit
fit = rsession.r['nls'](formula=fmla.r_repr(),data=object,start=startlist)
#Interpolate the fit over the the range of the independent variable
predicted = rsession.r['predict'](fit,df(x=smooth_range))
#If we fail to achieve convergance
except:
#Prompt the user for more information about coefficents
info('Fit could not be found, please adjust starting values for coefficents.')
startvalues = VariablePrompt(startdict.keys()).get_inputs()
if not startvalues:
return
for coef,value in zip(startdict.keys(),startvalues):
startdict[coef] = value
startlist=rsession.r['list'](**startdict)
if predicted:
#info('Fit Found')
fit_found = True
self.fit = fit
self.result = fit
plotwindow = rsession.RPlotThread(data=data,
args=args,
type='scatter',
export=image,
par_mode=(1,1))
#print 'predicted',predicted
#print 'smooth',smooth_range
plotwindow.add_cmd(rsession.ro.r.lines,smooth_range,predicted,col='red')
plotwindow.start()
self.active_threads.append(plotwindow)