Commit | Line | Data |
---|---|---|
3ddf1c12 | 1 | context("computeFilaments") |
1e20780e | 2 | |
af3b84f4 BA |
3 | #shorthand: map 1->1, 2->2, 3->3, 4->1, ..., 149->2, 150->3 |
4 | I = function(i) | |
5 | (i-1) %% 3 + 1 | |
6 | ||
44a9990b | 7 | #MOCK data; NOTE: could be in inst/testdata as well |
af3b84f4 | 8 | getDataTest = function(n) |
a65907cc | 9 | { |
af3b84f4 | 10 | data = Data$new() |
8f84543c BA |
11 | x = seq(0,9.5,0.1) |
12 | L = length(x) #96 1/4h | |
a65907cc BA |
13 | s1 = cos(x) |
14 | s2 = sin(x) | |
15 | s3 = c( s1[1:(L%/%2)] , s2[(L%/%2+1):L] ) | |
8f84543c BA |
16 | #sum((s1-s2)^2) == 96 |
17 | #sum((s1-s3)^2) == 58 | |
18 | #sum((s2-s3)^2) == 38 | |
19 | s = list(s1, s2, s3) | |
a65907cc BA |
20 | series = list() |
21 | for (i in seq_len(n)) | |
1e20780e | 22 | { |
af3b84f4 BA |
23 | serie = s[[I(i)]] + rnorm(L,sd=0.01) |
24 | level = mean(serie) | |
25 | serie = serie - level | |
a65907cc | 26 | # 10 series with NAs for index 2 |
af3b84f4 | 27 | if (I(i) == 2 && i >= 60 && i<= 90) |
a65907cc | 28 | serie[sample(seq_len(L),1)] = NA |
8f84543c BA |
29 | time = as.POSIXct(i*15*60, origin="2007-01-01", tz="GMT") |
30 | exo = runif(4) | |
31 | exo_hat = runif(4) | |
32 | data$append(time, serie, level, exo, exo_hat) | |
6d97bfec | 33 | } |
af3b84f4 | 34 | data |
6d97bfec BA |
35 | } |
36 | ||
af3b84f4 | 37 | test_that("output is as expected on simulated series", |
6d97bfec | 38 | { |
af3b84f4 | 39 | data = getDataTest(150) |
a65907cc | 40 | |
44a9990b | 41 | # index 143 : serie type 2 |
8f84543c BA |
42 | pred = computeForecast(data, 143, "Neighbors", "Zero", |
43 | horizon=length(data$getSerie(1)), simtype="endo", h_window=1) | |
44 | f = computeFilaments(data, pred, 1, limit=60, plot=FALSE) | |
98e958ca BA |
45 | |
46 | # Expected output: 50-3-10 series of type 2, then 23 series of type 3 (closest next) | |
47 | expect_identical(length(f$neighb_indices), as.integer(60)) | |
48 | expect_identical(length(f$colors), as.integer(60)) | |
49 | expect_equal(f$index, 143) | |
50 | expect_true(all(I(f$neighb_indices) >= 2)) | |
51 | for (i in 1:37) | |
6d97bfec | 52 | { |
98e958ca | 53 | expect_equal(I(f$neighb_indices[i]), 2) |
44a9990b | 54 | expect_match(f$colors[i], f$colors[1]) |
6d97bfec | 55 | } |
98e958ca | 56 | for (i in 38:60) |
6d97bfec | 57 | { |
98e958ca BA |
58 | expect_equal(I(f$neighb_indices[i]), 3) |
59 | expect_match(f$colors[i], f$colors[38]) | |
6d97bfec | 60 | } |
98e958ca BA |
61 | expect_match(f$colors[1], "#1*") |
62 | expect_match(f$colors[38], "#E*") | |
6d97bfec | 63 | |
af3b84f4 | 64 | # index 142 : serie type 1 |
8f84543c BA |
65 | pred = computeForecast(data, 142, "Neighbors", "Zero", |
66 | horizon=length(data$getSerie(1)), simtype="endo", h_window=1) | |
67 | f = computeFilaments(data, pred, 1, limit=50, plot=FALSE) | |
98e958ca BA |
68 | |
69 | # Expected output: 50-10-3 series of type 1, then 13 series of type 3 (closest next) | |
44a9990b | 70 | # NOTE: -10 because only past days with no-NAs tomorrow => exclude type 1 in [60,90[ |
98e958ca BA |
71 | expect_identical(length(f$neighb_indices), as.integer(50)) |
72 | expect_identical(length(f$colors), as.integer(50)) | |
73 | expect_equal(f$index, 142) | |
74 | expect_true(all(I(f$neighb_indices) != 2)) | |
75 | for (i in 1:37) | |
6d97bfec | 76 | { |
98e958ca | 77 | expect_equal(I(f$neighb_indices[i]), 1) |
6d97bfec BA |
78 | expect_match(f$colors[i], f$colors[1]) |
79 | } | |
98e958ca | 80 | for (i in 38:50) |
6d97bfec | 81 | { |
98e958ca BA |
82 | expect_equal(I(f$neighb_indices[i]), 3) |
83 | expect_match(f$colors[i], f$colors[38]) | |
6d97bfec | 84 | } |
98e958ca BA |
85 | expect_match(f$colors[1], "#1*") |
86 | expect_match(f$colors[38], "#E*") | |
1e20780e | 87 | }) |