Commit | Line | Data |
---|---|---|
56857861 | 1 | context("de_serialize") |
62deb424 | 2 | |
56857861 | 3 | test_that("serialization + getDataInFile retrieve original data / from matrix", |
62deb424 | 4 | { |
37c82bba | 5 | data_bin_file = ".epclust_test_m.bin" |
4bcfdbee BA |
6 | unlink(data_bin_file) |
7 | ||
37c82bba BA |
8 | #dataset 200 cols / 30 rows |
9 | data_ascii = matrix(runif(200*30,-10,10),nrow=30) | |
56857861 BA |
10 | nbytes = 4 #lead to a precision of 1e-7 / 1e-8 |
11 | endian = "little" | |
12 | ||
13 | #Simulate serialization in one single call | |
4bcfdbee | 14 | binarize(data_ascii, data_bin_file, 500, ",", nbytes, endian) |
56857861 BA |
15 | expect_equal(file.info(data_bin_file)$size, length(data_ascii)*nbytes+8) |
16 | for (indices in list(c(1,3,5), 3:13, c(5,20,50), c(75,130:135), 196:200)) | |
62deb424 | 17 | { |
56857861 | 18 | data_lines = getDataInFile(indices, data_bin_file, nbytes, endian) |
37c82bba | 19 | expect_equal(data_lines, data_ascii[,indices], tolerance=1e-6) |
62deb424 | 20 | } |
56857861 BA |
21 | unlink(data_bin_file) |
22 | ||
23 | #...in several calls (last call complete, next call NULL) | |
24 | for (i in 1:20) | |
4bcfdbee | 25 | binarize(data_ascii[((i-1)*10+1):(i*10),], data_bin_file, 20, ",", nbytes, endian) |
56857861 BA |
26 | expect_equal(file.info(data_bin_file)$size, length(data_ascii)*nbytes+8) |
27 | for (indices in list(c(1,3,5), 3:13, c(5,20,50), c(75,130:135), 196:200)) | |
62deb424 | 28 | { |
56857861 | 29 | data_lines = getDataInFile(indices, data_bin_file, nbytes, endian) |
37c82bba | 30 | expect_equal(data_lines, data_ascii[,indices], tolerance=1e-6) |
62deb424 | 31 | } |
56857861 BA |
32 | unlink(data_bin_file) |
33 | }) | |
34 | ||
492cd9e7 BA |
35 | test_that("serialization + transform + getDataInFile retrieve original transforms", |
36 | { | |
37c82bba | 37 | data_bin_file = ".epclust_test_t.bin" |
492cd9e7 BA |
38 | unlink(data_bin_file) |
39 | ||
37c82bba BA |
40 | #dataset 200 cols / 30 rows |
41 | data_ascii = matrix(runif(200*30,-10,10),nrow=30) | |
492cd9e7 BA |
42 | nbytes = 8 |
43 | endian = "little" | |
44 | ||
45 | binarize(data_ascii, data_bin_file, 500, ",", nbytes, endian) | |
46 | # Serialize transformation (just compute range) into a new binary file | |
37c82bba | 47 | trans_bin_file = ".epclust_test_t_trans.bin" |
492cd9e7 BA |
48 | unlink(trans_bin_file) |
49 | getSeries = function(inds) getDataInFile(inds, data_bin_file, nbytes, endian) | |
37c82bba | 50 | binarizeTransform(getSeries, function(series) apply(series, 2, range), |
492cd9e7 BA |
51 | trans_bin_file, 250, nbytes, endian) |
52 | unlink(data_bin_file) | |
53 | expect_equal(file.info(trans_bin_file)$size, 2*nrow(data_ascii)*nbytes+8) | |
54 | for (indices in list(c(1,3,5), 3:13, c(5,20,50), c(75,130:135), 196:200)) | |
55 | { | |
37c82bba BA |
56 | trans_cols = getDataInFile(indices, trans_bin_file, nbytes, endian) |
57 | expect_equal(trans_cols, apply(data_ascii[indices,],2,range), tolerance=1e-6) | |
492cd9e7 BA |
58 | } |
59 | unlink(trans_bin_file) | |
60 | }) | |
61 | ||
56857861 BA |
62 | test_that("serialization + getDataInFile retrieve original data / from connection", |
63 | { | |
37c82bba | 64 | data_bin_file = ".epclust_test_c.bin" |
4bcfdbee BA |
65 | unlink(data_bin_file) |
66 | ||
37c82bba | 67 | #dataset 300 cols / 50 rows |
56857861 BA |
68 | data_csv = system.file("testdata","de_serialize.csv",package="epclust") |
69 | nbytes = 8 | |
70 | endian = "big" | |
4bcfdbee | 71 | data_ascii = as.matrix(read.csv(data_csv, sep=";", header=FALSE)) #for ref |
56857861 BA |
72 | |
73 | #Simulate serialization in one single call | |
4bcfdbee | 74 | binarize(data_csv, data_bin_file, 350, ";", nbytes, endian) |
56857861 BA |
75 | expect_equal(file.info(data_bin_file)$size, 300*50*8+8) |
76 | for (indices in list(c(1,3,5), 3:13, c(5,20,50), c(75,130:135), 196:200)) | |
62deb424 | 77 | { |
56857861 | 78 | #HACK: as.matrix(as.data.frame( )) required to match (ref) data structure |
37c82bba BA |
79 | data_cols = as.matrix(as.data.frame( getDataInFile(indices,data_bin_file,nbytes,endian) )) |
80 | expect_equal(data_cols, data_ascii[,indices]) | |
62deb424 | 81 | } |
56857861 BA |
82 | unlink(data_bin_file) |
83 | ||
84 | #...in several calls / chunks of 29 --> 29*10 + 10, incomplete last | |
85 | data_con = file(data_csv, "r") | |
4bcfdbee | 86 | binarize(data_con, data_bin_file, 29, ";", nbytes, endian) |
56857861 BA |
87 | expect_equal(file.info(data_bin_file)$size, 300*50*8+8) |
88 | for (indices in list(c(1,3,5), 3:13, c(5,20,50), c(75,130:135), 196:200)) | |
62deb424 | 89 | { |
37c82bba BA |
90 | data_cols = as.matrix(as.data.frame( getDataInFile(indices,data_bin_file,nbytes,endian) )) |
91 | expect_equal(data_cols, data_ascii[,indices]) | |
62deb424 | 92 | } |
56857861 | 93 | unlink(data_bin_file) |
4bcfdbee | 94 | #close(data_con) --> done in binarize() |
56857861 | 95 | }) |