# File: descriptive-32K.r

rm(list = ls())
#library(dplyr)

## Following lines allow to construct the data file
## totconso-par-indiv.txt. 
## They are very time&memory-consumming 
#library(data.table)  
#d10 <- fread("~/tmp/data/2010.csv")
#tot <- d10[, list(tot= sum(CPP_PUISSANCE_BRUTE), 
#                  min= min(CPP_PUISSANCE_BRUTE), 
#                  max= max(CPP_PUISSANCE_BRUTE), 
#                  sd = sd( CPP_PUISSANCE_BRUTE)), 
#                  by = FK_CCU_ID]
# write.table(file = "~/tmp/data/totconso-par-indiv.txt", tot)

tot     <- read.table("~/tmp/data/totconso-par-indiv.txt")
synchro <- read.table("~/tmp/data/syncrone2010.txt")
mat_synchro <- matrix(c(synchro$V1[1:2], synchro$V1) / 1e6, 
                      ncol = 48,
                      byrow = TRUE)

xHour <- seq(1/2, 24, length.out = 48)
xDate <- seq.Date(from = as.Date("2010/1/1"),
                  to   = as.Date("2011/1/1"),
                  length.out = 365 * 48 - 2)  

# Filter infra daily
plot(xDate, filter(synchro$V1 / 1e6, rep(1 / 48, 48)), 
     type = 'l', 
     main = 'Synchrone 2010',
     ylab = 'Charge (en Mwh)',
     xlab = 'Time')

# Daily
matplot(xHour, scale(t(mat_synchro)), type = 'l', lty = 1, 
        col  = rgb(0, 0, 0, .25),
        main = 'Synchrone jounalière (norm.)',
        xlab = 'Heures', ylab = 'Charge (en Mwh)')


tot_order <- tot[order(tot$tot), ]
max_order <- tot[order(tot$max), ]

## Mesures aberrantes ?
tail(cbind(max_order[,1], max_order[,-1] / 1e3))


## I want to recuperate ids: 186733, 191819, 193174
## which corresponds to places 1, 2481, 3478
# identiants <- read.table('~/tmp/data/identifs.txt')
# match(c(186733, 191819, 193174), identiants$V1)

#c186733 <- d10[, FK_CCU_ID == 186733]
#c191819 <- d10[, FK_CCU_ID == 191819]
#c193174 <- d10[, FK_CCU_ID == 193174]

c186733 <- read.csv("~/tmp/data/c186733.csv")[, 4] / 1e3

mat_186733 <- matrix(c(c186733[1:2], c186733), 
                      ncol = 48,
                      byrow = TRUE)

# Filter infra daily
plot(xDate, filter(c186733, rep(1 / 48, 48)), 
     type = 'l', 
     main = "Consommation individuelle (2010)",
     ylab = 'Charge (en Mwh)',
     xlab = 'Temps')

# Daily (raw)
matplot(xHour, t(mat_186733), type = 'l', lty = 1, 
        col  = rgb(0, 0, 0, .25),
        #main = "Consommation individuelle journalière "
        xlab = 'Heures', ylab = 'Charge (en Mwh)')

# Daily (standarized)
matplot(xHour, scale(t(mat_186733)), type = 'l', lty = 1, 
        col  = rgb(0, 0, 0, .25),
        xlab = 'Heures', ylab = 'Charge (en Mwh)')