Return distance between individuals with edge_dist

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: spatsoc
 Title: Group Animal Relocation Data by Spatial and Temporal Relationship
-Version: 0.1.9
+Version: 0.1.10
 Authors@R: 
     c(person(given = "Alec L.",
              family = "Robitaille",

NEWS.md

@@ -1,4 +1,7 @@
-# v 0.1.9 
+# v 0.1.10
+* added optional return of distance between individuals with `edge_dist` ([PR 19](https://github.com/ropensci/spatsoc/pull/19))
+
+# v 0.1.9 (2019-05-14)
 * fixed bug for randomizations type 'step' and 'daily' ([PR 13](https://github.com/ropensci/spatsoc/pull/13)). 
 * clarified `SIMPLIFY=FALSE` in SNA vignette. 
 

R/edge_dist.R

@@ -15,10 +15,10 @@
 #' The \code{splitBy} argument offers further control over grouping. If within your \code{DT}, you have multiple populations, subgroups or other distinct parts, you can provide the name of the column which identifies them to \code{splitBy}. \code{edge_dist} will only consider rows within each \code{splitBy} subgroup.
 #'
 #' @inheritParams group_pts
+#' @param returnDist boolean indicating if the distance between individuals should be returned. If FALSE (default), only ID1, ID2 columns (and timegroup, splitBy columns if provided) are returned. If TRUE, another column "distance" is returned indicating the distance between ID1 and ID2.
 #' @param fillNA boolean indicating if NAs should be returned for individuals that were not within the threshold distance of any other. If TRUE, NAs are returned. If FALSE, only edges between individuals within the threshold distance are returned.
 #'
-#'
-#' @return \code{edge_dist} returns a \code{data.table}  with three columns: timegroup, ID1 and ID2.
+#' @return \code{edge_dist} returns a \code{data.table} with columns ID1, ID2, timegroup (if supplied) and any columns provided in splitBy. If 'returnDist' is TRUE, column 'distance' is returned indicating the distance between ID1 and ID2.
 #'
 #' The ID1 and ID2 columns represent the edges defined by the spatial (and temporal with \code{group_times}) thresholds.
 #'
@@ -41,13 +41,14 @@
 #'
 #' # Edge list generation
 #' edge_dist(DT, threshold = 100, id = 'ID',
-#'           coords = c('X', 'Y'), timegroup = 'timegroup', fillNA = TRUE)
+#'           coords = c('X', 'Y'), timegroup = 'timegroup', returnDist = TRUE, fillNA = TRUE)
 edge_dist <- function(DT = NULL,
                       threshold = NULL,
                       id = NULL,
                       coords = NULL,
                       timegroup = NULL,
                       splitBy = NULL,
+                      returnDist = FALSE,
                       fillNA = TRUE) {
   # due to NSE notes in R CMD check
   N <- Var1 <- Var2 <- value <- . <- NULL
@@ -104,7 +105,7 @@ edge_dist <- function(DT = NULL,
         )
         )
     }
-    }
+  }
 
   if (is.null(timegroup) && is.null(splitBy)) {
     splitBy <- NULL
@@ -129,8 +130,15 @@ edge_dist <- function(DT = NULL,
       as.matrix(stats::dist(.SD[, 2:3], method = 'euclidean'))
     diag(distMatrix) <- NA
     w <- which(distMatrix < threshold, arr.ind = TRUE)
-    list(ID1 = .SD[[1]][w[, 1]],
-         ID2 = .SD[[1]][w[, 2]])
+
+    if (returnDist) {
+      list(ID1 = .SD[[1]][w[, 1]],
+           ID2 = .SD[[1]][w[, 2]],
+           distance = distMatrix[w])
+    } else {
+      list(ID1 = .SD[[1]][w[, 1]],
+           ID2 = .SD[[1]][w[, 2]])
+    }
   },
   by = splitBy, .SDcols = c(id, coords)]
 
@@ -143,6 +151,5 @@ edge_dist <- function(DT = NULL,
   } else {
     return(edges)
   }
-
 }
 

README.Rmd

@@ -20,9 +20,12 @@ knitr::opts_chunk$set(message = FALSE,
 <!-- badges: end -->
 
 # spatsoc 
-spatsoc is an R package for detecting spatial and temporal groups in GPS relocations. It can be used to convert GPS relocations to gambit-of-the-group format to build proximity-based social networks with grouping and edge-list generating functions. In addition, the `randomizations` function provides data-stream randomization methods suitable for GPS data and the `get_gbi` function generates group by individual matrices useful for building networks with `asnipe::get_network`. 
 
-See below for installation and basic usage. 
+### [News](#news) | [Installation](#installation) | [Usage](#usage) | [Contributing](#contributing)
+
+`spatsoc` is an R package for detecting spatial and temporal groups in GPS relocations. It can be used to convert GPS relocations to gambit-of-the-group format to build proximity-based social networks with grouping and edge-list generating functions. In addition, the `randomizations` function provides data-stream randomization methods suitable for GPS data and the `get_gbi` function generates group by individual matrices useful for building networks with `asnipe::get_network`. 
+
+See below for [installation](#installation) and basic [usage](#usage). 
 
 For more details, see the [blog post](https://ropensci.org/blog/2018/12/04/spatsoc/) and vignettes:
 
@@ -32,7 +35,9 @@ For more details, see the [blog post](https://ropensci.org/blog/2018/12/04/spats
 
 
 
-## News
+
+
+## [News](http://spatsoc.robitalec.ca/news/index.html)
 New edge-list generating functions added (feedback welcome as always!):
 
   * `edge_nn`

README.md

@@ -15,16 +15,18 @@ checks](https://cranchecks.info/badges/summary/spatsoc)](https://cran.r-project.
 
 # spatsoc
 
-spatsoc is an R package for detecting spatial and temporal groups in GPS
-relocations. It can be used to convert GPS relocations to
+### [News](#news) | [Installation](#installation) | [Usage](#usage) | [Contributing](#contributing)
+
+`spatsoc` is an R package for detecting spatial and temporal groups in
+GPS relocations. It can be used to convert GPS relocations to
 gambit-of-the-group format to build proximity-based social networks with
 grouping and edge-list generating functions. In addition, the
 `randomizations` function provides data-stream randomization methods
 suitable for GPS data and the `get_gbi` function generates group by
 individual matrices useful for building networks with
 `asnipe::get_network`.
 
-See below for installation and basic usage.
+See below for [installation](#installation) and basic [usage](#usage).
 
 For more details, see the [blog
 post](https://ropensci.org/blog/2018/12/04/spatsoc/) and vignettes:
@@ -36,7 +38,7 @@ post](https://ropensci.org/blog/2018/12/04/spatsoc/) and vignettes:
   - [Using spatsoc in social network
     analysis](http://spatsoc.robitalec.ca/articles/using-in-sna.html)
 
-## News
+## [News](http://spatsoc.robitalec.ca/news/index.html)
 
 New edge-list generating functions added (feedback welcome as always\!):
 

codemeta.json

@@ -17,7 +17,7 @@
   ],
   "issueTracker": "https://github.com/ropensci/spatsoc/issues",
   "license": "https://spdx.org/licenses/GPL-3.0",
-  "version": "0.1.9",
+  "version": "0.1.10",
   "programmingLanguage": {
     "@type": "ComputerLanguage",
     "name": "R",
@@ -224,7 +224,7 @@
   ],
   "releaseNotes": "https://github.com/ropensci/spatsoc//blob/master/NEWS.md",
   "readme": "https://github.com/ropensci/spatsoc/blob/master/README.md",
-  "fileSize": "366.364KB",
+  "fileSize": "366.853KB",
   "contIntegration": "https://codecov.io/gl/robit.a/spatsoc",
   "developmentStatus": "http://www.repostatus.org/#active",
   "review": {

tests/testthat/test-edge-dist.R

@@ -222,3 +222,63 @@ test_that('returned IDs make sense', {
   expect_true(eDT[ID1 == ID2, .N] == 0)
 
 })
+
+
+
+test_that('returnDist works', {
+  copyDT <- copy(DT)[, datetime := as.POSIXct(datetime)]
+  group_times(copyDT, datetime = 'datetime', threshold = '10 minutes')
+
+  thresh <- 50
+  withDist <- edge_dist(
+    copyDT,
+    threshold = thresh,
+    id = 'ID',
+    coords = c('X', 'Y'),
+    timegroup = 'timegroup',
+    returnDist = TRUE,
+    fillNA = TRUE
+  )
+
+  woDist <- edge_dist(
+    copyDT,
+    threshold = thresh,
+    id = 'ID',
+    coords = c('X', 'Y'),
+    timegroup = 'timegroup',
+    returnDist = FALSE,
+    fillNA = TRUE
+  )
+
+  expect_equal(withDist[, .(ID1, ID2, timegroup)],
+               woDist[, .(ID1, ID2, timegroup)])
+
+  expect_equal(nrow(withDist),
+               nrow(woDist))
+
+  expect_equal(withDist[is.na(ID2)],
+               withDist[is.na(distance)])
+
+  expect_equal(withDist[!is.na(ID2)],
+               withDist[!is.na(distance)])
+
+  expect_lt(withDist[, max(distance, na.rm = TRUE)],
+            thresh)
+
+
+  withDistNoNA <- edge_dist(
+    copyDT,
+    threshold = thresh,
+    id = 'ID',
+    coords = c('X', 'Y'),
+    timegroup = 'timegroup',
+    returnDist = TRUE,
+    fillNA = FALSE
+  )
+
+  expect_true(withDistNoNA[is.na(distance), .N] == 0)
+  expect_true(withDistNoNA[is.na(ID2), .N] == 0)
+  expect_lt(withDistNoNA[, max(distance, na.rm = TRUE)],
+            thresh)
+
+})