Distinguish directedness of author networks and edge-construction algorithm #137
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -15,6 +15,7 @@ | |
| ## Copyright 2017 by Christian Hechtl <[email protected]> | ||
| ## Copyright 2017 by Felix Prasse <[email protected]> | ||
| ## Copyright 2018 by Barbara Eckl <[email protected]> | ||
| ## Copyright 2018 by Thomas Bock <[email protected]> | ||
| ## All Rights Reserved. | ||
|
|
||
|
|
||
|
|
@@ -238,6 +239,47 @@ test_that("Network construction of the undirected author-cochange network", { | |
| expect_true(igraph::identical_graphs(network.built, network.expected)) | ||
| }) | ||
|
|
||
| test_that("Network construction of the undirected but temorally ordered author-cochange network", { | ||
|
|
||
| ## configurations | ||
| proj.conf = ProjectConf$new(CF.DATA, CF.SELECTION.PROCESS, CASESTUDY, ARTIFACT) | ||
| proj.conf$update.value("artifact.filter.base", FALSE) | ||
| net.conf = NetworkConf$new() | ||
| net.conf$update.values(updated.values = list(author.relation = "cochange", respect.temporal.order = TRUE)) | ||
|
There was a problem hiding this comment. Can you please additionally set |
||
|
|
||
| ## construct objects | ||
| proj.data = ProjectData$new(project.conf = proj.conf) | ||
| network.builder = NetworkBuilder$new(project.data = proj.data, network.conf = net.conf) | ||
|
|
||
| ## build network | ||
| network.built = network.builder$get.author.network() | ||
|
|
||
| ## vertex attributes | ||
| authors = data.frame(name = c("Björn", "Olaf", "Karl", "Thomas"), | ||
| kind = TYPE.AUTHOR, | ||
| type = TYPE.AUTHOR) | ||
|
|
||
| ## edge attributes | ||
| data = data.frame(comb.1. = c("Olaf", "Karl", "Thomas", "Thomas"), | ||
| comb.2. = c("Björn", "Olaf", "Olaf", "Karl"), | ||
| date = get.date.from.string(c("2016-07-12 16:00:45", "2016-07-12 16:06:10", | ||
| "2016-07-12 16:06:32", "2016-07-12 16:06:32")), | ||
| artifact.type = "Feature", | ||
| hash = c("5a5ec9675e98187e1e92561e1888aa6f04faa338", "1143db502761379c2bfcecc2007fc34282e7ee61", | ||
| "0a1a5c523d835459c42f33e863623138555e2526", "0a1a5c523d835459c42f33e863623138555e2526"), | ||
| file = c("test.c", "test3.c", "test2.c", "test2.c"), | ||
| artifact = c("A", "Base_Feature", "Base_Feature", "Base_Feature"), | ||
| weight = 1, | ||
| type = TYPE.EDGES.INTRA, | ||
| relation = "cochange" | ||
| ) | ||
|
|
||
| ## build expected network | ||
| network.expected = igraph::graph.data.frame(data, directed = FALSE, vertices = authors) | ||
|
|
||
| expect_true(igraph::identical_graphs(network.built, network.expected)) | ||
| }) | ||
|
|
||
| test_that("Network construction of the directed author-cochange network", { | ||
|
|
||
| ## configurations | ||
|
|
@@ -279,6 +321,52 @@ test_that("Network construction of the directed author-cochange network", { | |
| expect_true(igraph::identical_graphs(network.built, network.expected)) | ||
| }) | ||
|
|
||
| test_that("Network construction of the directed author-cochange network without respecting temporal order", { | ||
|
|
||
| ## configurations | ||
| proj.conf = ProjectConf$new(CF.DATA, CF.SELECTION.PROCESS, CASESTUDY, ARTIFACT) | ||
| proj.conf$update.value("artifact.filter.base", FALSE) | ||
| net.conf = NetworkConf$new() | ||
| net.conf$update.values(updated.values = list(author.relation = "cochange", author.directed = TRUE, | ||
| respect.temporal.order = FALSE)) | ||
|
There was a problem hiding this comment. The indentation is wrong here. |
||
|
|
||
| ## construct objects | ||
| proj.data = ProjectData$new(project.conf = proj.conf) | ||
| network.builder = NetworkBuilder$new(project.data = proj.data, network.conf = net.conf) | ||
|
|
||
| ## build network | ||
| network.built = network.builder$get.author.network() | ||
|
|
||
| ## vertex attributes | ||
| authors = data.frame(name = c("Björn", "Olaf", "Karl", "Thomas"), | ||
| kind = TYPE.AUTHOR, | ||
| type = TYPE.AUTHOR) | ||
|
|
||
| ## edge attributes | ||
| data = data.frame(comb.1. = c("Björn", "Björn", "Olaf", "Olaf", "Olaf", "Olaf", "Karl", "Karl"), | ||
| comb.2. = c("Olaf", "Olaf", "Karl", "Karl", "Thomas", "Thomas", "Thomas", "Thomas"), | ||
|
There was a problem hiding this comment. Unfortunately, the easy fix for issue #6 will not be as easy as expected... 😞 The network constructed here is – sadly – wrong! The idea was that, in a network ignoring the temporal order of events (i.e.,
Okay, how do we solve this? In the last hours, I debugged this and prepared a patch in the end. Unfortunately, this temporarily breaks some network tests with issue data because the order of edges is messed up. Please add the patch to your branch and fix the tests. There was a problem hiding this comment. Oh, now I know what I did forget to do. Before opening this PR, I wanted to check something but did not remember any more what I had wanted to check. When I have read your comment, I remembered again, that I wanted to check for the edge directions again. I am sorry for that. Moreover, it was not intended that it is your job to spend hours on debugging that. I would just have been fine with a review comment like "the edge directions are wrong here, please debug that". I will have a look at your patch tomorrow and also will fix the other tests. That's the least I can do. There was a problem hiding this comment.
I guess, everybody knows that feeling. 😁
Sorry for the misleading formulation. For the patch, I needed only some short time, I just spent some further time on debugging what was happening inside
You are welcome. Keep me posted. |
||
| date = get.date.from.string(c("2016-07-12 15:58:59", "2016-07-12 16:00:45", "2016-07-12 16:05:41", | ||
| "2016-07-12 16:06:10", "2016-07-12 16:05:41", "2016-07-12 16:06:32", | ||
| "2016-07-12 16:06:10", "2016-07-12 16:06:32")), | ||
| artifact.type = "Feature", | ||
| hash = c("72c8dd25d3dd6d18f46e2b26a5f5b1e2e8dc28d0", "5a5ec9675e98187e1e92561e1888aa6f04faa338", | ||
| "3a0ed78458b3976243db6829f63eba3eead26774", "1143db502761379c2bfcecc2007fc34282e7ee61", | ||
| "3a0ed78458b3976243db6829f63eba3eead26774", "0a1a5c523d835459c42f33e863623138555e2526", | ||
| "1143db502761379c2bfcecc2007fc34282e7ee61", "0a1a5c523d835459c42f33e863623138555e2526"), | ||
| file = c("test.c", "test.c", "test2.c", "test3.c", "test2.c", "test2.c", "test3.c", "test2.c"), | ||
| artifact = c("A", "A", "Base_Feature", "Base_Feature", "Base_Feature", "Base_Feature", | ||
| "Base_Feature", "Base_Feature"), | ||
| weight = 1, | ||
| type = TYPE.EDGES.INTRA, | ||
| relation = "cochange" | ||
| ) | ||
|
|
||
| ## build expected network | ||
| network.expected = igraph::graph.data.frame(data, directed = TRUE, vertices = authors) | ||
|
|
||
| expect_true(igraph::identical_graphs(network.built, network.expected)) | ||
| }) | ||
|
|
||
| test_that("Network construction of the undirected simplified author-cochange network", { | ||
|
|
||
| ## configurations | ||
|
|
||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I like this description, but it seems a bit long for this list. What do you think about moving most of this description to the subsection
Relationsin the SectionNetwork construction(or a new subsection)? There, we would have enough space to also insert the table from issue #6 (or an equivalent illustration).Alternatively, we could introduce an
**Example**item as forskip.threshold. 😉 Although, I think that a new section as suggested above would be more appropriate as the matter is quite important.Additionally:
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Adding all the information regarding network construction would overload the
Relationssection. Therefore, I would add a new section after theRelationssection, namedDifferent edge-construction algorithms for author networks. There we can add also the table from issue #6.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
👍 I like the heading for the new subsection, although, I would remove the "Different".