Merge pull request #6 from pmcharrison/dev

Dev

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: hrep
 Title: Harmony Representations
-Version: 0.5.1
+Version: 0.6.0
 Authors@R: person("Peter", "Harrison", email = "pmc.harrison@gmail.com", role = c("aut", "cre"))
 Description: This package provides methods for representing and manipulating chord sequences.
 Depends: R (>= 3.4.0)

NAMESPACE

@@ -203,6 +203,7 @@ export(amplitude_to_dB)
 export(ascending_pc_dist)
 export(coded_vec)
 export(corpus)
+export(cosine_similarity)
 export(dB_to_amplitude)
 export(decode)
 export(encode)
@@ -232,6 +233,7 @@ export(lower)
 export(map_pc_chord_id_to_pc_set_id)
 export(metadata)
 export(midi_to_freq)
+export(milne_pc_spec_dist)
 export(milne_pc_spectrum)
 export(num_elements)
 export(num_sequences)

NEWS.md

@@ -0,0 +1,4 @@
+# hrep 0.6.0
+
+- Added a `NEWS.md` file to track changes to the package.
+- Added `milne_pc_spec_dist` for computing spectral distances between sonorities.

R/cosine-similarity.R

@@ -0,0 +1,14 @@
+#' Cosine similarity
+#'
+#' Computes the cosine similarity between two numeric vectors.
+#' @param x Numeric vector 1.
+#' @param y Numeric vector 2.
+#' @return Cosine similarity, as a numeric scalar.
+#' @export
+cosine_similarity <- function(x, y) {
+  numerator <- sum(x * y)
+  denominator <-
+    sqrt(sum(x ^ 2)) *
+    sqrt(sum(y ^ 2))
+  numerator / denominator
+}

R/milne-pc-spec-dist.R

@@ -0,0 +1,24 @@
+#' Milne's pitch-class spectral distance
+#'
+#' Computes Milne's pitch-class spectral distance measure between two sonorities.
+#'
+#' @param x First sonority, passed to \code{\link{milne_pc_spectrum}}.
+#' @param y Second sonority, passed to \code{\link{milne_pc_spectrum}}.
+#' @param ... Further arguments passed to \code{\link{milne_pc_spectrum}}.
+#'
+#' @return
+#' A numeric scalar where greater values indicate greater distance.
+#' 0 is the minimum possible distance.
+#' 1 corresponds to two uncorrelated spectra;
+#' values greater than 1 are possible if the spectra are inversely correlated.
+#'
+#' @details
+#' See \insertCite{Milne2016;textual}{hrep} for details.
+#'
+#' @export
+milne_pc_spec_dist <- function(x, y, ...) {
+  1 - cosine_similarity(
+    milne_pc_spectrum(x, ...),
+    milne_pc_spectrum(y, ...)
+  )
+}

R/pc-milne-spectrum.R → R/milne-pc-spectrum.R

@@ -1,4 +1,4 @@
-#' Constructor function for pitch-class Milne spectrum
+#' Constructor function for Milne pitch-class spectrum
 #'
 #' This function constructs a "milne_pc_spectrum" object.
 #' @param x A numeric vector of pitch-class weights,
@@ -23,11 +23,11 @@
   y
 }
 
-#' Pitch-class Milne spectrum
+#' Milne pitch-class spectrum
 #'
 #' This function represents an input object as a
-#' 'pitch-class Milne spectrum'.
-#' A pitch-class Milne spectrum defines 'perceptual weight'
+#' 'Milne pitch-class spectrum'.
+#' A Milne pitch-class spectrum defines 'perceptual weight'
 #' as a continuous function of 'pitch class'.
 #' @details
 #' This spectrum is typically constructed from musical chords

R/vec.R

@@ -71,7 +71,7 @@ is.coded.vec <- function(x) FALSE
 print.vec <- function(x, detail = FALSE, ...) {
   if (detail) {
     for (i in seq_along(x)) {
-      cat("[[", i, "]] ", capture.output(x[[i]]), "\n", sep = "")
+      cat("[[", i, "]] ", utils::capture.output(x[[i]]), "\n", sep = "")
     }
   } else {
     cat("Vector of type '", type(x),

tests/testthat/test-milne_pc_spec_dist.R

@@ -0,0 +1,95 @@
+context("test-milne_pc_spec_dist")
+
+test_that("examples", {
+  approx_equal <- function(x, y, digits = 4) {
+    all.equal(
+      round(x, digits = digits),
+      round(y, digits = digits)
+    )
+  }
+
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(0, 4, 7)
+      ),
+      0
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(0, 3, 7)
+      ),
+      0.2597
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(0, 2, 7)
+      ),
+      0.2201
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(1, 5, 8)
+      ),
+      0.8527
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(2, 7, 11)
+      ),
+      0.4155
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(0, 4, 7),
+        c(0, 5, 9)
+      ),
+      0.4155
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(2, 7, 8),
+        c(3, 4, 8, 9)
+      ),
+      0.5619
+    )
+  )
+  expect_true(
+    approx_equal(
+      milne_pc_spec_dist(
+        c(2, 6, 8, 9),
+        c(1, 4)
+      ),
+      0.5940
+    )
+  )
+})
+
+test_that("coercion", {
+  expect_equal(
+    milne_pc_spec_dist(c(0, 4, 7), c(0, 3, 7)),
+    milne_pc_spec_dist(c(60, 64, 67), c(60, 63, 67))
+  )
+
+  expect_equal(
+    milne_pc_spec_dist(c(0, 4, 7), c(0, 3, 7)),
+    milne_pc_spec_dist(pi_chord(c(60, 64, 67)), pi_chord(c(60, 63, 67)))
+  )
+})

tests/testthat/test-r-cosine-similarity.R

@@ -0,0 +1,25 @@
+context("test-r-cosine-similarity")
+
+test_that("examples", {
+  expect_equal(cosine_similarity(rep(1, times = 10),
+                                 rep(2, times = 10)),
+               1)
+  expect_equal(cosine_similarity(rep(1, times = 10),
+                                 rep(-2, times = 10)),
+               -1)
+  expect_equal(
+    cosine_similarity(
+      c(0, 3, 0, 0, 2, 0, 0, 2, 0, 5),
+      c(1, 2, 0, 0, 1, 1, 0, 1, 0, 3)
+    ),
+    0.94,
+    tolerance = 1e-2
+  )
+  expect_equal(
+    cosine_similarity(
+      c(0, 3, 0, 0, 2, 0, 0, 2, 0, 5),
+      c(1, 2, 0, 0, 1, 1, 0, 1, 0, 3)
+    ),
+    0.9356015
+  )
+})