Add Commit Networks

NEWS.md

@@ -12,17 +12,21 @@
 - Add line-based code coverage reports into CI pipeline. Coverage reports are generated by `coverage.R` (PR #262, 10cac49d005e87c3964cc61711e7f5acef749626, b3b9f4ac7a9911bd00293c68fac88e0f9033bdfb, c815d18dc6266d620a7a145493417b87ac08679e, e8093525fdaf46e54f2f7fcc6358ca7892e795e5, 32d04823e2007c63d2a43ce59bea3057327c19a7)
 - Add the possibility to split data time-based by multiple data sources (PR #261, 1088395f46b84028c8d7c463ca86b5dc38500c26, e1f79fc9e40cd6f41c946be42db364b2101cfe10, 0bb187fec0fd801d7634bf8d5180525770f6ab0b, 371a97ac6ebf3de4fe9360dea79d62e2ed3ef585)
 - Add tests for uncovered functionality in `util-misc.R` and `util-networks.R` (PR #264, ff30f3238b1bf2539280d0d055a5d925c197c271, af80551d0615a49b86e45ff596bd75941ee88f91)
+- Add commit network as a new type of network. It uses commits as vertices and connects them either via cochange or commit interactions. This includes adding new config parameters and a function for adding vertex attributes to a commit network(PR #263, ab73271781e8e9a0715f784936df4b371d64c338, ab73271781e8e9a0715f784936df4b371d64c338, cd9a930fcb54ff465c2a5a7c43cfe82ac15c134d)
 
 ### Changed/Improved
 
 - Change the default value for the `issues.from.source` configuration parameter. Instead of reading JIRA and GitHub issues together, which was the previous default, the new default value causes only GitHub issue data to be read. To restore the previous default behavior and read data from both issue sources, this now needs to be manually configured when needed. (PR #264, 5ff83c364f6bfc1e6ff95e9c5f1087e031c48a5d, 8c8080cb9caf115f19d9f145ad6e6c108b131a67, 8bcbc81db521877908d2e5c2989082ed672f2a3b)
 - Replace deprecated `igraph` functions by their preferred alternatives (PR #264, 0df9d5bf6bafbb5d440f4c47db4ec901cf11f037)
 - Deprecate support for R version 3.6 (PR #264, c8e6f45111e487fadbe7f0a13c7595eb23f3af6e, fb3f5474259d4a88f4ff545691cca9d1ccde90e3)
 - Explicitly add R version 4.4 to the CI test pipeline (c8e6f45111e487fadbe7f0a13c7595eb23f3af6e)
+- Refactor function `construct.edge.list.from.key.value.list` to be more readable(PR #263, 05c3bc09cb1d396fd59c34a88030cdca58fd04dd)
 
 ### Fixed
 
 - Fix the creation of edgelists for issue-based artifact-networks by correctly iterating over the issue data (PR #264, 321d85043112971c04998249c14a0677a32c9004)
+- Fix networks based upon commit interaction data to also have the attribute `artifact.type`(PR #263, 849123a8b7d898fbb1343745ecffc1f6000c9367)
+- Fix endless recursion that could occur when commit interaction data was configured and commit data is empty (PR #263, 3fb7437b68950303916b62984fa449732c70353e)
 
 ## 4.4
 

README.md

@@ -630,7 +630,7 @@ Updates to the parameters can be done by calling `NetworkConf$update.variables(.
 - `author.relation`
     * The relation(s) among authors, encoded as edges in an author network
     * **Note**: The  author--artifact relation in bipartite and multi networks is configured by `artifact.relation`!
-    * possible values: [*`"mail"`*, `"cochange"`, `"issue"`]
+    * possible values: [*`"mail"`*, `"cochange"`, `"issue"`, `commit.interaction`]
 - `author.directed`
     * The directedness of edges in an author network
     * [`TRUE`, *`FALSE`*]
@@ -649,7 +649,7 @@ Updates to the parameters can be done by calling `NetworkConf$update.variables(.
 - `artifact.relation`
     * The relation(s) among artifacts, encoded as edges in an artifact network
     * **Note**: Additionally, this relation configures also the author--artifact relation in bipartite and multi networks!
-    * possible values: [*`"cochange"`*, `"callgraph"`, `"mail"`, `"issue"`]
+    * possible values: [*`"cochange"`*, `"callgraph"`, `"mail"`, `"issue"`, `commit.interaction`]
 - `artifact.directed`
     * The directedness of edges in an artifact network
     * **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.

util-conf.R

@@ -63,6 +63,8 @@ ARTIFACT.CODEFACE = list(
     "file"     = "File"
 )
 
+ARTIFACT.COMMIT.INTERACTION = "CommitInteraction"
+
 
 ## / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / / /
 ## Conf --------------------------------------------------------------------

util-data.R

@@ -415,11 +415,13 @@ ProjectData = R6::R6Class("ProjectData",
         #'
         #' This method should be called whenever the field \code{commit.interactions} is changed.
         update.commit.interactions = function() {
-            stacktrace = get.stacktrace(sys.calls())
-            caller = get.second.last.element(stacktrace)
-            if (self$is.data.source.cached("commit.interactions") &&
-                (is.na(caller)|| paste(caller, collapse = " ") != "self$set.commits(commit.data)")) {
-                if (!self$is.data.source.cached("commits.unfiltered")) {
+            if (self$is.data.source.cached("commit.interactions")) {
+                ## check if caller was 'set.commits'. If so, or if commits are already filtered,
+                ## do not get the commits again.
+                stacktrace = get.stacktrace(sys.calls())
+                caller = get.second.last.element(stacktrace)
+                if (!self$is.data.source.cached("commits.unfiltered") &&
+                    (is.na(caller) || paste(caller, collapse = " ") != "self$set.commits(commit.data)")) {
                     self$get.commits()
                 }
 

util-networks.R

@@ -249,7 +249,7 @@
             colnames(edges)[2] = "from"
             colnames(edges)[4] = "hash"
             if (nrow(edges) > 0) {
-                edges[["artifact.type"]] = "CommitInteraction"
+                edges[["artifact.type"]] = ARTIFACT.COMMIT.INTERACTION
             }
             author.net.data = list(vertices = vertices, edges = edges)
             ## construct the network
@@ -700,8 +700,8 @@
 
             ## do not compute anything more than once
             if (!is.null(private$commits.network.commit.interaction)) {
                 logging::logdebug("get.commit.network.commit.interaction: finished. (already existing)")
                 return(private$commits.network.commit.interaction)
             }
 
             ## get the hashes that appear in the commit-interaction data as the vertices of the network
@@ -715,7 +715,7 @@
             edges = edges[, c("base.hash", "commit.hash", "func", "interacting.author",
                               "file", "base.author", "base.func", "base.file")]
             if (nrow(edges) > 0) {
-                edges[["artifact.type"]] = "CommitInteraction"
+                edges[["artifact.type"]] = ARTIFACT.COMMIT.INTERACTION
             }
             colnames(edges)[1] = "to"
             colnames(edges)[2] = "from"
@@ -1443,9 +1443,9 @@
 
     ## Skip artifacts with many, many edges
     if (number.edges > network.conf$get.value("skip.threshold")) {
         logging::logwarn("Skipping edges for %s '%s' due to amount (> %s).",
                          attr(set, "group.type"), attr(set, "group.name"), network.conf$get.value("skip.threshold"))
         return(NULL)
     }
 
     if (network.type == "commit") {
@@ -1523,9 +1523,9 @@
 
     ## Skip artifacts with many, many edges
     if (number.edges > network.conf$get.value("skip.threshold")) {
         logging::logwarn("Skipping edges for %s '%s' due to amount (> %s).",
                          attr(set, "group.type"), attr(set, "group.name"), network.conf$get.value("skip.threshold"))
         return(NULL)
     }
 
     ## get vertex data
@@ -1533,7 +1533,7 @@
 
     ## break if there is no author
     if (length(vertices) < 1) {
         return(NULL)
     }
 
     ## if there is only one author, just create the vertex, but no edges