Commit | Line | Data |
---|---|---|
ad642dc6 BA |
1 | ## File : 03_compute-sum-of-classes_2009.r |
2 | ## Description : Calculer les synchrones pour chaque groupe obtenu par le | |
3 | ## clustering. | |
4 | ||
5 | rm(list = ls()) | |
6 | ||
7 | setwd("~/ownCloud/projects/2014_EDF-Orsay-Lyon2/codes/") | |
8 | ||
9 | ## 1. Read auxiliar data files #### | |
10 | ||
11 | identifiants <- read.table("identifs.txt")[ ,1] | |
12 | dates0 <- read.table("datesall.txt")[, 1] | |
13 | dates <- dates0[grep("2009", dates0)] | |
14 | rm(dates0) | |
15 | ||
16 | n <- length(identifiants) | |
17 | p <- length(dates) | |
18 | ||
19 | blocks <- c(rep(6500, 3), 5511) | |
20 | ||
21 | # Fit of the clustering : clfit | |
22 | #load('~/Documents/projects/2014_EDF-Orsay-Lyon2/res/clfit500.Rdata') | |
23 | #load('~/Documents/projects/2014_EDF-Orsay-Lyon2/res/clfit200.Rdata') | |
24 | load('../res/clfit200RC.Rdata') | |
25 | # table(clfit$clustering) | |
26 | ||
27 | names(clfit$clustering) <- sample(names(clfit$clustering))# for random classes | |
28 | dfclust <- data.frame(cluster = clfit$clustering) | |
29 | # write.table(dfclust, file = "../res/clfit200-random.txt") | |
30 | ||
31 | K <- nrow(clfit$clusinfo) | |
32 | #dfclust <- head(dfclust, 50) # just for testing purpouses | |
33 | synchros <- matrix(0, ncol = p, nrow = K) | |
34 | rm(clfit) | |
35 | ||
36 | ## 2. Process the large file #### | |
37 | ||
38 | close(con) | |
39 | con <- file("~/tmp/2009_full.txt") # Establish a connection to the file | |
40 | open(con, "r") # Open the connection | |
41 | ||
42 | for(b in seq_along(blocks)){ # Reading loop | |
43 | nb <- blocks[b] | |
44 | actual <- readLines(con = con, n = nb ) | |
45 | auxmat <- matrix(unlist(strsplit(actual, " ")), ncol = p + 1, byrow = TRUE) | |
46 | rm(actual) | |
47 | ||
48 | datamat <- t(apply(auxmat[, -1], 1, as.numeric)) | |
49 | rownames(datamat) <- substr(auxmat[, 1], 2, 7) | |
50 | rm(auxmat) | |
51 | ||
52 | # obtain for each line of datamat the cluster membership (if any) | |
53 | clustfactor <- dfclust$cluster[match(rownames(datamat), rownames(dfclust))] | |
54 | ||
55 | # if(length(which(rownames(datamat) == "218095"))>0) print(b); | |
56 | ||
572d139a BA |
57 | |
58 | #TODO: en C : lourd | |
ad642dc6 BA |
59 | for(k in 1:K){ |
60 | clustk <- which(clustfactor == k) | |
61 | if(length(clustk) > 0) { | |
62 | ||
63 | if(length(clustk) > 1) { | |
64 | synchrosk <- colSums(datamat[which(clustfactor == k), ]) | |
65 | } else { | |
66 | synchrosk <- datamat[which(clustfactor == k), ] | |
67 | } | |
68 | synchros[k, ] <- synchros[k, ] + synchrosk | |
69 | rm(synchrosk) | |
70 | } | |
71 | } | |
72 | } | |
73 | ||
74 | close(con) # close connection to the file | |
75 | ||
76 | synchros <- data.frame(t(synchros), total = colSums(synchros)) | |
77 | ||
78 | # write.table(synchros, file = "~/tmp/2009_synchros200RC.txt") | |
79 | # write.table(synchros, file = "~/tmp/2009_synchros200-random.txt") | |
80 |