v3.3 #139
Merged
v3.3 #139
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The function takes project data and returns a dataframe containing the commit count per person counting commits, where the person is author and committer. Works with SQL, analogous to other methods in the same file. It is used for the adding the commit count as vertex attribute. Signed-off-by: Klara Schlueter <[email protected]>
… attribute The function adds the commit count per person (counting commits, where the person is the author AND the committer) to the vertices of the networks given as a list, analogous to already existing functions in the same class. Signed-off-by: Klara Schlueter <[email protected]>
Signed-off-by: Klara Schlueter <[email protected]>
The function takes project data and returns a dataframe containing the commit count per person (counting commits, where the person is author OR committer), working wie SQL, analogous to already existing methods in the same file.
The function adds the commit count per person (counting the commits, where the person is author OR commiter) as attributes to the vertices of the given networks, working analogous to already existing methods in the same file. Signed-off-by: Klara Schlueter <[email protected]>
The commit data in the test data is complemented by committer information in addition to author information. Signed-off-by: Klara Schlueter <[email protected]>
Signed-off-by: Klara Schlueter <[email protected]>
The functions for adding a commit count as vertex attribute set as default value an Integer now. Signed-off-by: Klara Schlueter <[email protected]>
As committer data was added to test commit data earlier, the tests for reading and cutting the test project data expected incomplete data. To resolve this, tests in test-data-cut.R, test-network-cut.R and test-read.R are adapted to the new test data. Signed-off-by: Klara Schlueter <[email protected]>
Signed-off-by: Klara Schlueter <[email protected]>
Signed-off-by: Klara Schlueter <[email protected]>
Signed-off-by: Klara Schlueter <[email protected]>
…andor Complement commit count vertex attribute possibilities Reviewed-by: Claus Hunsen <[email protected]> Reviewed-by: Christian Hechtl <[email protected]>
To hopefully get predictable results under Windows and get rid of an error message issued by the current code, we now set the locale to "english-us" when Windows is detected. For all other operating systems, the previous setting is preserved. Signed-off-by: Claus Hunsen <[email protected]>
For clearer guidelines for the user/developer regarding issues and pull requests, the corresponding templates are updated. We now distinguish bug reports and feature requests for issues and represent a checklist with prerequisites for pull requests. The template files are now located in the folder '.github/' in the project's root folder. The contribution guide is updated accordingly. Signed-off-by: Claus Hunsen <[email protected]>
In TravisCI, using a matrix build, we can run the test suite for several versions of R. In this patch, we enable the matrix build for the R versions 3.3, 3.4, and 3.5. This way, we may find incompatibility issues early in the process. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
See travis-ci/travis-ci#9654 for details. Signed-off-by: Claus Hunsen <[email protected]>
When using the editor Remarkable for Markdown editing, the second-level lists are not rendered when using only two spaces of further indentation. To fix this, all second-level lists are further indented. Signed-off-by: Claus Hunsen <[email protected]>
For better comprehensibility, we add a section on the library's functionality to the README file. This includes a short list of data sources (grouped by type), very basic information on network construction (igraph, network attributes, types of networks), and the list of available files/modules. The paragraph on the various types of networks is not formulated yet, so this is a work-in-progress patch. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
For a better overview and to give more information to the interested reader/user, the README section on mandatory vertex and edge attributes is further extended. For each mandatory attribute, we now give information on the attribute's name, the reasoning and information stored in the attribute, and, if applicable, possible values. This fixes #121. Signed-off-by: Claus Hunsen <[email protected]>
After removing 'grayscale' as a parameter to the plotting functions, the
corresponding showcases need to be removed. Additionally, an additional
edge for the sample network now inherits all necessary edge attributes
('relation' and 'artifact.type').
Signed-off-by: Claus Hunsen <[email protected]>
Until now, all network attributes set on an igraph object have been removed when calling 'simplify.network'. Now, these attributes are preserved by saving the before simplification and re-setting the attributes afterwards. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
As the legends for the network plots can get really long, a complete line-oriented layout for them is impractical, esepcially when there are multiple relations present in the network to be plotted. To mitigate the problem, the legend is now column-oriented by default. This means, individually, the multiple sublegends/guides are presented in a column- oriented way, while the sublegends are positioned next to each other in a horizontal manner. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
As pointed out in PR #135, the instructions on adding vertex attributes to constructed networks get more and more confusing due to the increasing number of helper functions in the file `util-networks-covariates.R`. To be more precise, the functions to add vertex attributes are now described with the pattern "add.vertex.attribute.*" in the README file. Additionally, some typos are fixed. Props to @bockthom for pointing this out. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
In this patch, we heavily update the README file regarding the feedback of @SCPhantom: - Re-order sections, - add information on relations, - add information on PaStA and synchronicity data, - clarify on range/revision difference, - clarify on (non-)dependency to packrat, and - update how-to code to use sample data. Signed-off-by: Claus Hunsen <[email protected]> Signed-off-by: Barbara Eckl <[email protected]> Signed-off-by: Thomas Bock <[email protected]>
As the bipartite edges in the exemplary plot has been directed from the artifacts to the authors (which is wrong!), the plot is updated with all edges reversed. Additionally, the code to generate the plot and dump it to disk is added to the showcase file. Thanks to @ecklbarb and @bockthom for spotting this. Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
Fix two typos with missing spaces. Props to @bockthom for being that precise. Signed-off-by: Claus Hunsen <[email protected]>
Various fixes to README and contribution guide Reviewed-by: Thomas Bock <[email protected]>
…orithm From now on, the 'author.directed' parameter in the network configuration only affects the directedness of an author network, and the newly introduced parameter 'respect.temporal.order' is responsible for choosing the edge-construction algorithm. Now, it is possible to build author networks which are directed but don't respect the temporal order. That is, for two subsequent events, there will be two edges between the corresponding nodes, one in each direction. In addition, it is also possible to build author networks which are undirected but respect the temporal order. That is, only one edge (annotated with the temporally later date) will be present between the corresponding nodes of two subsequent events. So, let there be an event of author A to common artifact at time 1 and an event of B to the same artifact at time 2. Then, we are able to construct the following networks (only the list of edges is presented here): author.directed=TRUE, respect.temporal.order=TRUE: A<--2--B author.directed=TRUE, respect.temporal.order=FALSE: A--1-->B, A<--2--B author.directed=FALSE, respect.temporal.order=TRUE: A--2-- B author.directed=FALSE, respect.temporal.order=FALSE: A--1--B, A--2--B For more information, please see #6. To keep the former behavior, the default value of the newly introduced network configuration parameter 'respect.temporal.order' is NA. If the default value this parameter is used, the edge-construction function uses the 'directed' parameter for 'respect.temporal.order'. This fixes #6. Signed-off-by: Thomas Bock <[email protected]>
Add a test for constructing undirected networks which respect the temporal order and add another test for constructing directed networks without respecting the temporal order. The tests test the functionality introduced in commit 4fc59a0. Signed-off-by: Thomas Bock <[email protected]>
When working on the functions 'construct.edge.list.from.key.value.list' and 'construct.network.from.edge.list', I recognized some inconsistent uses of '#' in comments. Comments which begin at the beginning of a line should start with '##'. This commit fixes that in the above mentioned functions. Signed-off-by: Thomas Bock <[email protected]>
Update README.md accordingly after introducing the new network configuration parameter 'respect.temporal.order'. This updates the README.md regarding the changes in commit 4fc59a0. Signed-off-by: Thomas Bock <[email protected]>
Signed-off-by: Thomas Bock <[email protected]>
In the course of PR #137, it occurred that the temporal order of edges was broken in directed networks. The cause of this error is the technical implementation to obtain all combinations of items in a key-value set: The call 'combn(nodes, 2)' in the function 'construct.edge.list.from.key.value.list' does not care for the order of the items, although we need to care. To fix this problem, we need to extract the edge list for each item in an item combination separately and use the second one as the receiver of the edge. Additionally, the function 'construct.edge.list.from.key.value.list' is adapted to match the coding conventions: - access the vertex column via its name 'data.vertices', - use 'seq_len' instead of the ':' operator, and - add curly braces for a single-line if-statement. Note: This temporarily breaks some tests due to igraph's normalization on undirected edge lists which reorders some edges (apparently) randomly. Signed-off-by: Claus Hunsen <[email protected]>
Also extend documentation in the source code for the edge-construction algorithm for better comprehensibility. Signed-off-by: Thomas Bock <[email protected]>
After a change of the edge-construction algorithm in 70b3c82, for some author networks in the test suite, the internal representation of the corresponding igraph object has changed: The igraph object now stores the edges in another order (but the network itself has not changed). This commit fixes those tests. Signed-off-by: Thomas Bock <[email protected]>
-Move the detailed information on how the parameter works to an individual section. -Add an example (taken from #6) and extend it by adding loop edges. -Fix minor issues on README from review in PR #137. Signed-off-by: Thomas Bock <[email protected]>
Signed-off-by: Thomas Bock <[email protected]>
Improve the section regarding edge construction for author networks in the README.md file. Signed-off-by: Thomas Bock <[email protected]> Signed-off-by: Barbara Eckl <[email protected]> Signed-off-by: Claus Hunsen <[email protected]>
To clarify, that the network configuration parameter 'respect.temporal.order' only applies for the construction of author networks, the parameter is renamed to 'author.respect.temporal.order'. Signed-off-by: Thomas Bock <[email protected]>
Escape arrow brackets. Signed-off-by: Thomas Bock <[email protected]>
After getting feedback from @SCPhantom, improve the README section on edge-construction algorithms for author networks again. Signed-off-by: Thomas Bock <[email protected]> Signed-off-by: Barbara Eckl <[email protected]> Signed-off-by: Claus Hunsen <[email protected]>
Signed-off-by: Thomas Bock <[email protected]>
Distinguish directedness of author networks and edge-construction algorithm Reviewed-by: Claus Hunsen <[email protected]> Reviewed-by: Barbara Eckl <[email protected]> Reviewed-by: Jakob Kronawitter <[email protected]>
Signed-off-by: Claus Hunsen <[email protected]>
|
As we reviewed everything before, I will merge rightaway when the build checks finish successfully. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
3.3
Changes in detail
v3.2...d4e2f26
Added
metrics.is.scale.free()to decide whether a network is scale-free or not (80f4751, 97161b1)clear.edge.attributesto clear the edge attributes list of the network configuration (15f7587)author.respect.temporal.orderfor determining the edge-construction algorithm (Distinguish directedness of networks and edge-construction algorithm #6, 4fc59a0, fd0b07d)Changed/Improved
typeandkindattributes of vertices in bipartite networks for testing reasons (351311a)Fixed
simplify.network(inigraphlanguage, graph attributes) (424b2bc)split.networks.time.basedmethod by now splitting the networks from the earliest timestamp to the latest (1f65db3)sudocommands (baca08e)