Merge pull request #250 from bockthom/thomas-updates

Enable simplification of multi-relation edges & make multi-network construction work with igraph version 2.0.1.1

Reviewed-by: Christian Hechtl <[email protected]>

.github/workflows/pull_request.yml

@@ -12,6 +12,7 @@
 ## 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 ##
 ## Copyright 2023 by Maximilian Löffler <[email protected]>
+## Copyright 2024 by Thomas Bock <[email protected]>
 ## All Rights Reserved.
 
 name: Build Status
@@ -29,17 +30,17 @@ permissions:
 jobs:
   build:
     name: Build
-    
+
     # change to 'runs-on: self-hosted' to run on self-hosted runners (https://docs.github.com/en/actions/using-jobs/choosing-the-runner-for-a-job)
     runs-on: ubuntu-latest
-    
+
     strategy:
       fail-fast: false
       matrix:
-        r-version: ['3.6', '4.0', '4.1', '4.2', 'latest']
+        r-version: ['3.6', '4.0', '4.1', '4.2', '4.3', 'latest']
 
     steps:
-      - name: Checkout Repo 
+      - name: Checkout Repo
         uses: actions/checkout@v3
 
       - name: Update system
@@ -56,7 +57,7 @@ jobs:
         uses: r-lib/actions/setup-r@v2
         with:
           r-version: ${{ matrix.r-version }}
-      
+
       - name: Install dependencies
         run: Rscript install.R
 

NEWS.md

@@ -6,10 +6,11 @@
 
 ### Added
 
-- Add issue-based artifact-networks (PR #244, 98a93ee721a293410623aafe46890cfba9d81e72, 771bcc8d961d419b53a1e891e9dc536371f1143b)
+- Add issue-based artifact-networks (PR #244, 98a93ee721a293410623aafe46890cfba9d81e72, 771bcc8d961d419b53a1e891e9dc536371f1143b, 368e79264adf5a5358c04518c94ad2e1c13e212b)
 - Add a new `split.data.by.bins` function (not to be confused with a previously existing function that had the same name and was renamed in this context), which splits data based on given activity-based bins (PR #244, ece569ceaf557bb38cd0cfad437b69b30fe8a698, ed5feb214a123b605c9513262f187cfd72b9e1f4)
 - Add new `assert.sparse.matrices.equal` function to compare two sparse matrices for equality for testing purposes (PR #248, 9784cdf12d1497ee122e2ae73b768b8c334210d4, d9f1a8d90e00a634d7caeb5e7f8f262776496838)
 - Add tests for file `util-networks.misc.R` for issue #242 (PR #248, f3202a6f96723d11c170346556d036cf087521c8, 030574b9d0f3435db4032d0e195a3d407fb7244b, 380b02234275127297fcd508772c69db21c216de, 8b803c50d60fc593e4e527a08fd4c2068d801a48, 7335c3dd4d0302b024a66d18701d9800ed3fe806, 6b600df04bec1fe70c272604f274ec5309840e65)
+- Add the possibility to simplify edges of multiple-relation networks into a single edge at all instead of a single edge per relation (PR #250, 2105ea89b5227e7c9fa78fea9de1977f2d9e8faa)
 
 ### Changed/Improved
 
@@ -20,6 +21,7 @@
 - Change `util-tensor.R` to correctly use the new output format of `get.author.names.from.network` (PR #248, 72b663ebf7169c0da5c687fe215529f3be0c08c5)
 - Throw an error in `convert.adjacency.matrix.list.to.array` if the function is called with wrong parameters (PR #248, ece2d38b4972745af3a83e06f32317a06465a345, 1a3e510df15f5fa4e920e9fce3e0e162c27cd6d1)
 - Rename `compare.networks` to `assert.networks.equal` to better match the purpose of the function (PR #248, d9f1a8d90e00a634d7caeb5e7f8f262776496838)
+- Explicitly add R version 4.3 to the CI test pipeline (9f346d5bc3cfc553f01e5e80f0bbe51e1dc2b53e)
 
 ### Fixed
 
@@ -30,6 +32,8 @@
 - Fix `get.expanded.adjacency` to work if the provided author list does not contain all authors from network and add a warning when that happens since it causes some authors from the network to be lost in the resulting matrix (PR #248, ff59017e114b10812dcfb1704a19e01fc1586a13)
 - Fix `get.expanded.adjacency.matrices` to have correct names for the columns and rows (PR #248, e72eff864a1cb1a4aecd430e450d4a6a5044fdf2)
 - Fix `get.expanded.adjacency.cumulated` so that it works if `weighted` parameter is set to `FALSE` (PR #248, 2fb9a5d446653f6aee808cbfc87c2dafeb9a749a)
+- Fix broken error loggingin `metrics.smallworldness` (03e06881f06abf30d44b69d7988873f20b95232d)
+- Fix multi-network construction to work with igraph version 2.0.1.1, which does not allow to add an empty list of vertices (PR #250, 5547896faa279f6adaae4b2b77c7ab9623ddf256)
 
 
 ## 4.3

README.md

@@ -632,7 +632,7 @@ Updates to the parameters can be done by calling `NetworkConf$update.variables(.
     * possible values: [*`"cochange"`*, `"callgraph"`, `"mail"`, `"issue"`]
 - `artifact.directed`
     * The directedness of edges in an artifact network
-    * **Note**: This parameter does not take effect for now, as the `cochange` relation is always undirected, while the `callgraph` relation is always directed. For the other relations (`mail` and `issue`), we currently do not have data available to exhibit edge information.
+    * **Note**: This parameter does only affect the `issue` relation, as the `cochange` relation is always undirected, while the `callgraph` relation is always directed. For the `mail`, we currently do not have data available to exhibit edge information.
   * [`TRUE`, *`FALSE`*]
 - `edge.attributes`
     * The list of edge-attribute names and information
@@ -654,6 +654,10 @@ Updates to the parameters can be done by calling `NetworkConf$update.variables(.
 - `simplify`
     * Perform edge contraction to retrieve a simplified network
     * [`TRUE`, *`FALSE`*]
+- `simplify.multiple.relations`
+    * Whether the simplified network should contract edges of multiple relations into a single edge or not (if not, there will be one edge for each relation, resulting in possibly more than one edge between a pair of vertices)
+    * **Note** This parameter does not take effect if ``simplify = FALSE``!
+    * [`TRUE`, *`FALSE`*]
 - `skip.threshold`
     * The upper bound for total amount of edges to build for a subset of the data, i.e., not building any edges for the subset exceeding the limit
     * any positive integer

util-conf.R

@@ -18,7 +18,7 @@
 ## Copyright 2020-2021 by Christian Hechtl <[email protected]>
 ## Copyright 2017 by Felix Prasse <[email protected]>
 ## Copyright 2017-2019 by Thomas Bock <[email protected]>
-## Copyright 2021, 2023 by Thomas Bock <[email protected]>
+## Copyright 2021, 2023-2024 by Thomas Bock <[email protected]>
 ## Copyright 2018 by Barbara Eckl <[email protected]>
 ## Copyright 2018-2019 by Jakob Kronawitter <[email protected]>
 ## Copyright 2019 by Anselm Fehnker <[email protected]>
@@ -863,6 +863,12 @@ NetworkConf = R6::R6Class("NetworkConf", inherit = Conf,
                 allowed = c(TRUE, FALSE),
                 allowed.number = 1
             ),
+            simplify.multiple.relations = list(
+               default = FALSE,
+               type = "logical",
+               allowed = c(TRUE, FALSE),
+               allowed.number = 1
+           ),
             skip.threshold = list(
                 default = Inf,
                 type = "numeric",

util-networks-metrics.R

@@ -12,7 +12,7 @@
 ## 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 ##
 ## Copyright 2015, 2019 by Thomas Bock <[email protected]>
-## Copyright 2021, 2023 by Thomas Bock <[email protected]>
+## Copyright 2021, 2023-2024 by Thomas Bock <[email protected]>
 ## Copyright 2017 by Raphael Nömmer <[email protected]>
 ## Copyright 2017-2019 by Claus Hunsen <[email protected]>
 ## Copyright 2017-2018 by Christian Hechtl <[email protected]>
@@ -167,7 +167,7 @@ metrics.smallworldness = function(network) {
     if (!is.simple(network)) {
         ## if this is not the case, raise an error and stop the execution
         error.message = "The input network has too many edges. Try again with a simplified network."
-        logging::error(error.message)
+        logging::logerror(error.message)
         stop(error.message)
     }
 

util-networks.R

@@ -15,7 +15,7 @@
 ## Copyright 2017 by Raphael Nömmer <[email protected]>
 ## Copyright 2017-2018 by Christian Hechtl <[email protected]>
 ## Copyright 2017-2019 by Thomas Bock <[email protected]>
-## Copyright 2021, 2023 by Thomas Bock <[email protected]>
+## Copyright 2021, 2023-2024 by Thomas Bock <[email protected]>
 ## Copyright 2018 by Barbara Eckl <[email protected]>
 ## Copyright 2018-2019 by Jakob Kronawitter <[email protected]>
 ## Copyright 2020 by Anselm Fehnker <[email protected]>
@@ -958,8 +958,10 @@ NetworkBuilder = R6::R6Class("NetworkBuilder",
             ## to be consistent with bipartite networks
             artifacts.to.add.kind[artifacts.to.add.kind == "IssueEvent"] = "Issue"
 
-            artifacts.net = artifacts.net + igraph::vertices(artifacts.to.add, type = TYPE.ARTIFACT,
-                                                             kind = artifacts.to.add.kind)
+            if (length(artifacts.to.add) > 0) {
+                artifacts.net = artifacts.net + igraph::vertices(artifacts.to.add, type = TYPE.ARTIFACT,
+                                                                 kind = artifacts.to.add.kind)
+            }
 
             ## check directedness and adapt artifact network if needed
             if (igraph::is.directed(authors.net) && !igraph::is.directed(artifacts.net)) {
@@ -1260,7 +1262,8 @@ construct.network.from.edge.list = function(vertices, edge.list, network.conf, d
 
     ## transform multiple edges to edge weights
     if (network.conf$get.value("simplify")) {
-        net = simplify.network(net)
+        net = simplify.network(net,
+                               simplify.multiple.relations = network.conf$get.value("simplify.multiple.relations"))
     }
 
     logging::logdebug("construct.network.from.edge.list: finished.")
@@ -1536,16 +1539,19 @@ add.attributes.to.network = function(network, type = c("vertex", "edge"), attrib
 #' @param network the given network
 #' @param remove.multiple whether to contract multiple edges between the same pair of vertices [default: TRUE]
 #' @param remove.loops whether to remove loops [default: TRUE]
+#' @param simplify.multiple.relations whether to combine edges of multiple relations into
+#'        one simplified edge [default: FALSE]
 #'
 #' @return the simplified network
-simplify.network = function(network, remove.multiple = TRUE, remove.loops = TRUE) {
+simplify.network = function(network, remove.multiple = TRUE, remove.loops = TRUE,
+                            simplify.multiple.relations = FALSE) {
     logging::logdebug("simplify.network: starting.")
     logging::loginfo("Simplifying network.")
 
     ## save network attributes, otherwise they get lost
     network.attributes = igraph::get.graph.attribute(network)
 
-    if (length(unique(igraph::get.edge.attribute(network, "relation"))) > 1) {
+    if (!simplify.multiple.relations && length(unique(igraph::get.edge.attribute(network, "relation"))) > 1) {
         ## data frame of the network
         edge.data = igraph::as_data_frame(network, what = "edges")
         vertex.data = igraph::as_data_frame(network, what = "vertices")
@@ -1587,17 +1593,20 @@ simplify.network = function(network, remove.multiple = TRUE, remove.loops = TRUE
 #' @param networks the list of networks
 #' @param remove.multiple whether to contract multiple edges between the same pair of vertices [default: TRUE]
 #' @param remove.loops whether to remove loops [default: TRUE]
+#' @param simplify.multiple.relations whether to combine edges of multiple relations into
+#'                                    one simplified edge [default: FALSE]
 #'
 #' @return the simplified networks
-simplify.networks = function(networks, remove.multiple = TRUE, remove.loops = TRUE) {
+simplify.networks = function(networks, remove.multiple = TRUE, remove.loops = TRUE,
+                             simplify.multiple.relations = FALSE) {
     logging::logdebug("simplify.networks: starting.")
     logging::loginfo(
         "Simplifying networks (names = [%s]).",
         paste(names(networks), collapse = ", ")
     )
 
     nets = parallel::mclapply(networks, simplify.network, remove.multiple = remove.multiple,
-                              remove.loops = remove.loops)
+                              remove.loops = remove.loops, simplify.multiple.relations = simplify.multiple.relations)
 
     logging::logdebug("simplify.networks: finished.")
     return(nets)