# This R environment comes with all of CRAN preinstalled, as well as many other helpful packages
# The environment is defined by the kaggle/rstats docker image: https://github.com/kaggle/docker-rstats
# For example, here's several helpful packages to load in 

library(ggplot2) # Data visualization
library(readr) # CSV file I/O, e.g. the read_csv function

# Input data files are available in the "../input/" directory.
# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the input directory

system("ls ../input")

# Any results you write to the current directory are saved as output.

#Load data.table for fast reads and aggreagtions
library(data.table)

# Input data files are available in the "../input/" directory.

# Read in only required columns and force to numeric to ensure that subsequent 
# aggregation when calculating medians works
train <- fread('../input/train.csv', 
               select = c('Agencia_ID', 'Cliente_ID', 'Producto_ID', 'Demanda_uni_equil'),
               colClasses=c(Agencia_ID="numeric", Cliente_ID="numeric",Producto_ID="numeric",Demanda_uni_equil="numeric"))


#Print first 6 rows and show that conversion from integer has been successful
head(train)
sapply(train, class)


# In case the pair of product/agent pair has a median available, we use that as the predicted value.
# If not, then the product median is checked. If a value is not found, then the global median is used.

# set a table key to enable fast aggregations
setkey(train, Producto_ID, Agencia_ID, Cliente_ID)

#calculate the overall median
median <- train[, median(Demanda_uni_equil)]
#median = 3
print("M1")

#Calculate the appropriate mean; call it m2
app_Mean <- train[, abs(exp(mean(log(Demanda_uni_equil+1))))-1, by = Producto_ID]
setnames(app_Mean,"V1","m2")
print("m2")

#Calculate the appropriate mean; call it m2
app_Agent_Mean <- train[, abs(exp(mean(log(Demanda_uni_equil+1))))-1, by = .(Producto_ID,Agencia_ID,Cliente_ID)]
setnames(app_Agent_Mean,"V1","m3")
print("m3")

###################################################################################################
#  
# That's the 'modeling' done now need to apply scoring to test set
# 
###################################################################################################

# Read in Test data 
# Read in only required columns and force to numeric

test <- fread('../input/test.csv', 
               select = c('id','Agencia_ID', 'Producto_ID', 'Cliente_ID'),
               colClasses=c(Agencia_ID="numeric",Producto_ID="numeric", Cliente_ID="numeric"))


#Print first 6 rows and show that conversion from integer has been successful
head(test)
sapply(test, class)

# set a table key to enable fast joins to predictions
setkey(test, Producto_ID, Agencia_ID, Cliente_ID)

# Create table called submit that joins medians (in field M3) by Product and Agent to test data set
submit <- merge(test, app_Agent_Mean, all.x = TRUE)

# add column M2 that contains median by Product
submit$m2 <- merge(test, app_Mean, by = "Producto_ID", all.x = TRUE)$m2

# Now create Predictions column; intially set to be M3 which contains median by product and agent
submit$Pred <- submit$m3

# where median by product and agent is null use median by product (M2)
submit[is.na(m3)]$Pred <- round(submit[is.na(m3)]$m2)

# where median by product is null use overall median
submit[is.na(Pred)]$Pred <- median

# now relabel columns ready for creating submission
setnames(submit,"Pred","Demanda_uni_equil")

# check all looks OK
head(submit)

# Write out submission file.
# Any results you write to the current directory are saved as output.
write.csv(submit[,.(id,Demanda_uni_equil)],"submit_med_by_Agency_Client.csv", row.names = FALSE)