import cPickle
import rpy2.robjects as R

IN = open("data.pickle")
data = cPickle.load(IN)
IN.close()

#Create an R data frame.
dataframe = {}
dataframe['username'] = R.StrVector((data['username']))
dataframe['join_date'] = R.StrVector((data['join_date']))
dataframe['posts'] = R.IntVector((data['posts']))
dataframe['last_visit'] = R.StrVector((data['last_visit']))
dataframe['birthday'] = R.StrVector((data['birthday']))
dataframe['age'] = R.IntVector((data['age']))
MyRDataframe = R.DataFrame(dataframe)

print MyRDataframe.colnames

#Plot ages
hist = R.r.hist
R.r.png('~/Desktop/hist.png',width=300,height=300)
hist(MyRDataframe[2], main="", xlab="", br=20)
R.r['dev.off']()   #No built in function

#Plot and fit a model for to posts/activity.
#First we need to compute how many days the user has been active.

activity = R.r(r'''
    function(x, y) {
        if (is.na(x) | is.na(y)) NA
        else {
            date.1 <- strptime(x, "%m-%d-%Y")
            date.2 <- strptime(y, "%m-%d-%Y")
            difftime(date.1, date.2, units='days') 
        }
    }
''')

as_numeric = R.r['as.numeric']
days_active = activity(dataframe['join_date'], dataframe['last_visit'])
days_active = R.r.abs(days_active)
days_active = as_numeric(days_active)

plot = R.r.plot
R.r.png('~/Desktop/plot.png', width=300, height=300)
plot(days_active, MyRDataframe[3], pch='.')
R.r['dev.off']()

#Fit a linear model.
fm1a = R.Formula('y ~ x')
env = fm1a.environment
env['y'] = days_active
env['x'] = MyRDataframe[3]
#stats = importr('lm')
fit = R.r.lm(fm1a)
summary = R.r.summary(fit)
#Now we can program with the results of fit in Python.
print summary[3]

intercept = summary[3][0]
slope = summary[3][1]

print "We are in Python now."
print "A linear model is     POSTS = %f + %f * DAYS_ACTIVE" % (intercept, slope)