implement calling community_leiden from igraph v1.2.7

TomKellyGenetics · TomKellyGenetics · commit 855e801ce681 · 2022-04-29T06:52:09.000+09:00
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -21,6 +21,7 @@
 ^\ .DS_Store$
 revdep
 images
+lastMiKTeXException
 
 ## Files to ignore when building package
 
@@ -46,3 +47,4 @@ images
 ^vignettes/bash_py_time$
 ^revdep$
 ^aarhus_mplex.rda$
+^CRAN-SUBMISSION$
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,8 +1,8 @@
 Package: leiden
 Type: Package
 Title: R Implementation of Leiden Clustering Algorithm
-Version: 0.4.1.9001
-Date: 2021-07-27
+Version: 0.4.2
+Date: 2022-05-01
 Authors@R: c(person("S. Thomas", "Kelly", email = "tomkellygenetics@gmail.com", role = c("aut", "cre", "trl")),
            person("Vincent A.", "Traag", email = "v.a.traag@cwts.leidenuniv.nl", role = c("com")))
 Description: Implements the 'Python leidenalg' module to be called in R.
@@ -16,21 +16,20 @@ Imports:
     methods,
     reticulate,
     Matrix,
-    igraph (> 1.2.6)
+    igraph (>= 1.2.7)
 Encoding: UTF-8
 LazyData: false
-RoxygenNote: 7.1.1
+RoxygenNote: 7.1.2
 Suggests: 
     bipartite,
     covr,
     data.table,
     devtools,
     graphsim,
     knitr,
-    multiplex,
-    multinet,
     markdown,
     multiplex,
+    multinet,
     network,
     RColorBrewer,
     remotes,
diff --git a/NAMESPACE b/NAMESPACE
@@ -22,7 +22,7 @@ importFrom(igraph,graph_from_adjacency_matrix)
 importFrom(igraph,is.igraph)
 importFrom(igraph,is_bipartite)
 importFrom(igraph,is_directed)
-importFrom(igraph,is.named)
+importFrom(igraph,is_named)
 importFrom(igraph,is_weighted)
 importFrom(igraph,laplacian_matrix)
 importFrom(igraph,membership)
diff --git a/NEWS.md b/NEWS.md
@@ -1,3 +1,11 @@
+# leiden 0.4.2
+
+* migrates bug fixes to conda environment and limit on total cells to refactored package
+
+* allows calling igraph::community_leiden for supported parameters (requires igraph v1.2.7 or later)
+
+* automatically calls native R version of leiden rather than Python to improve performance
+
 # leiden 0.4.1
 
 * migrates changes to retain on CRAN to leiden 0.4.0 (alpha)
@@ -6,6 +14,23 @@
 
 * migrate to calling community_leiden in igraph
 
+# leiden 0.3.10
+
+* removes limitation on number of cells (disables scientific notation within function call): resolves #12
+
+* resolves conflict between base and r-reticulate conda environments on loading: resolves #20
+
+* updates conda environment in interactive sessions only for compliance to CRAN checks
+
+* resolves formatting error in Rmarkdown vignettes (https://github.com/yihui/knitr/issues/2057)
+
+* update testing for bipartite graphs for compatibility with newer version
+
+# leiden 0.3.9
+
+Updates maintainer contact details.
+>>>>>>> master
+
 # leiden 0.3.8
 
 * bug fixes for vignettes to retain on CRAN
diff --git a/R/find_partition.R b/R/find_partition.R
@@ -22,6 +22,11 @@ run_bipartite_partitioning <- function(py_graph,
                                        seed = NULL,
                                        n_iterations = 2L,
                                        max_comm_size = 0L){
+  # disable printing numerals in scientific notation
+  oo <- options(scipen = 100000000000)
+  # restore options when function terminates
+  on.exit(options(oo))
+
   #import python modules with reticulate
   numpy <- import("numpy", delay_load = TRUE)
   leidenalg <- import("leidenalg", delay_load = TRUE)
@@ -85,8 +90,12 @@ n_iterations = 2L,
 max_comm_size = 0L,
 degree_as_node_size = TRUE,
 legacy = FALSE
-degree_as_node_size = TRUE
 ) {
+  # disable printing numerals in scientific notation
+  oo <- options(scipen = 100000000000)
+  # restore options when function terminates
+  on.exit(options(oo))
+
   #import python modules with reticulate
   numpy <- import("numpy", delay_load = TRUE)
   leidenalg <- import("leidenalg", delay_load = TRUE)
@@ -289,6 +298,11 @@ max_comm_size = 0L,
 degree_as_node_size = TRUE,
 legacy = FALSE
 ) {
+  # disable printing numerals in scientific notation
+  oo <- options(scipen = 100000000000)
+  # restore options when function terminates
+  on.exit(options(oo))
+
   #import python modules with reticulate
   numpy <- import("numpy", delay_load = TRUE)
   leidenalg <- import("leidenalg", delay_load = TRUE)
diff --git a/R/leiden.R b/R/leiden.R
diff --git a/README.md b/README.md
@@ -1,6 +1,6 @@
 # Leiden Algorithm
 
-## leiden development version 0.4.1
+## leiden version 0.4.2
 
 [![CRAN_Status_Badge](http://www.r-pkg.org/badges/version/leiden)](https://cran.r-project.org/package=leiden)
 [![Travis Build Status](https://travis-ci.com/TomKellyGenetics/leiden.svg?branch=master)](https://travis-ci.com/TomKellyGenetics/leiden)
@@ -283,16 +283,16 @@ Please cite this implementation R in if you use it:
 ```
 To cite the leiden package in publications use:
 
-  S. Thomas Kelly (2021). leiden: R implementation of the Leiden algorithm. R
-  package version 0.4.0 https://github.com/TomKellyGenetics/leiden
+  S. Thomas Kelly (2022). leiden: R implementation of the Leiden algorithm. R
+  package version 0.4.2 https://github.com/TomKellyGenetics/leiden
 
 A BibTeX entry for LaTeX users is
 
   @Manual{,
     title = {leiden: R implementation of the Leiden algorithm},
     author = {S. Thomas Kelly},
-    year = {2021},
-    note = {R package version 0.4.0},
+    year = {2022},
+    note = {R package version 0.4.2},
     url = {https://github.com/TomKellyGenetics/leiden},
   }
  ```
diff --git a/cran-comments.md b/cran-comments.md
@@ -5,33 +5,24 @@
 * rhub (release) Ubuntu Linux 16.04 LTS, R-release, GCC
 * Fedora (devel) Linux, R-devel, clang, gfortran
 * MacOS 10.14.6 R 3.6.1 
-* MacOS 10.15.7 R 4.1.0
+* MacOS 10.15.7 R 4.2.0
 
 ## R CMD check results
 
 0 errors ✔ | 0 warnings ✔ | 0 notes ✔
 
 ## Updates
 
-Changes to vignettes to address the following issue:
+Updates calling native R version in igraph v1.2.7 when applicable.
+Significant performance improvements are expected without python
+dependencies for some parameters.
 
-Specifically, please see the problems for the Debian-based checks for
-r-devel.
-
-These are from
-
-      \item \code{matrix(x, n, m)} now warns in more cases where
-      \code{length(x)} differs from than \code{n * m}; suggested by Abby
-      Spurdle and Wolfgang Huber in Feb 2021 on the R-devel mailing
-      list.
-
-      This warning can be turned into an error by setting environment
-      variable \env{_R_CHECK_MATRIX_DATA_} to \samp{TRUE}: \command{R
-        CMD check --as-cran} does so unless it is already set.
-
-in current r-devel, where the Debian-based checks are now performed with
-_R_CHECK_MATRIX_DATA_=TRUE.
+Breaking changes are possible with the new implementation.
+A legacy mode with setting the random seed is supported 
+to reproduce previous results.
 
 ## Python integration
 
+Python is a soft dependency which is still required for some functions but is not essential for core functionality any longer. It is retained for backwards compatibility.
+
 Vignettes are disabled when python is not available. This works on Linux test environments (with python available) and windows test environments (without python or pip packages).
diff --git a/inst/CITATION b/inst/CITATION
@@ -3,21 +3,15 @@ citHeader("To cite the leiden package in publications use:")
 citEntry(entry = "Manual",
          title        = "leiden: R implementation of the Leiden algorithm",
          author       = personList(as.person("S. Thomas Kelly")),
-<<<<<<< HEAD
-         year         = "2020",
-         note         = "R package version 0.4.0",
+         year         = "2022",
+         note         = "R package version 0.4.2",
          url          = "https://github.com/TomKellyGenetics/leiden",
 
          textVersion  =
-           paste("S. Thomas Kelly (2020). leiden: R implementation of the Leiden algorithm. R package version 0.4.0",
-=======
-         year         = "2021",
-         note         = "R package version 0.3.8",
+           paste("S. Thomas Kelly (2020). leiden: R implementation of the Leiden algorithm. R package version 0.4.2",
+         year         = "2022",
+         note         = "R package version 0.4.2",
          url          = "https://github.com/TomKellyGenetics/leiden",
-
-         textVersion  =
-           paste("S. Thomas Kelly (2021). leiden: R implementation of the Leiden algorithm. R package version 0.3.8",
->>>>>>> master
                  "https://github.com/TomKellyGenetics/leiden")
 )
 
diff --git a/tests/testthat/test_bipartite.R b/tests/testthat/test_bipartite.R
@@ -63,7 +63,7 @@ skip_if_no_python <- function() {
     testthat::skip("leidenalg not available for testing")
 }
 
-
+set.seed(9000)
 test_that("run with CPMVertexPartition.Bipartite", {
   skip_if_no_python()
   partition <- leiden(bipartite_graph,
@@ -78,6 +78,7 @@ test_that("run with CPMVertexPartition.Bipartite", {
   bipartite_graph
 })
 
+set.seed(9000)
 test_that("run with CPMVertexPartition.Bipartite and degree as node size", {
   skip_if_no_python()
   partition <- leiden(bipartite_graph,
@@ -86,22 +87,23 @@ test_that("run with CPMVertexPartition.Bipartite and degree as node size", {
                       degree_as_node_size = TRUE,
                       seed = 9001)
   expect_length(partition, length(V(bipartite_graph)))
-  expect_equal(sort(unique(partition)), c(1, 2))
+  expect_equal(sort(unique(partition)), c(1, 2, 3, 4))
   expect_equal(partition,
-               c(1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1,
-                 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2))
+               c(1, 1, 1, 1, 1, 1, 3, 4, 3, 3, 2, 2, 2, 2, 2, 2, 4, 4, 1, 1,
+                 1, 1, 1, 1, 3, 3, 4, 2, 4, 2, 2, 2))
 })
 
+set.seed(9000)
 test_that("run with ModularityVertexPartition.Bipartite", {
   skip_if_no_python()
   partition <- leiden(bipartite_graph,
                       partition_type = "ModularityVertexPartition.Bipartite",
                       resolution_parameter = 0.01,
                       seed = 9001)
   expect_length(partition, length(V(bipartite_graph)))
-  expect_equal(sort(unique(partition)), c(1, 2))
+  expect_equal(sort(unique(partition)), c(1, 2, 3, 4))
   expect_equal(partition,
-               c(1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1,
-                 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2))
+               c(1, 1, 1, 1, 1, 1, 3, 4, 3, 3, 2, 2, 2, 2, 2, 2, 4, 4, 1, 1,
+                 1, 1, 1, 1, 3, 3, 4, 2, 4, 2, 2, 2))
 })
 
diff --git a/tests/testthat/test_igraph.R b/tests/testthat/test_igraph.R
@@ -80,7 +80,6 @@ test_that("run with named adjacency matrix", {
   partition <- leiden(snn_graph)
   expect_length(partition, 100)
 })
-<<<<<<< HEAD
 
 mat1 <- matrix(round(runif(10000, 0, 1)), 100, 100)
 mat2 <- matrix(round(rbinom(10000, 1, 0.1)), 100, 100)
@@ -204,5 +203,3 @@ test_that("run consistent results bewteen igraph and reticulate legacy mode with
                        resolution_parameter = 0.5)
   expect_true(all(table(partition, partition2) %in% c(0, 100)))
 })
-=======
->>>>>>> master
diff --git a/tests/testthat/test_multiplex.R b/tests/testthat/test_multiplex.R
@@ -1,10 +1,7 @@
 library("reticulate")
 library("igraph")
 library("leiden")
-<<<<<<< HEAD
 library("multiplex")
-=======
->>>>>>> master
 set.seed(9000)
 
 context("running Leiden on multiplex igraph objects")
diff --git a/tests/testthat/test_weighted.R b/tests/testthat/test_weighted.R
@@ -162,7 +162,3 @@ test_that("same output with different input class", {
 # sum(diag(table(part_mat2, part_graph2))) == length(part_mat)
 # sum(diag(table(part_mat2, part_graph2))) == nrow(mat)
 # sum(diag(table(part_mat2, part_graph2))) == length(V(graph))
-<<<<<<< HEAD
-=======
-
->>>>>>> master
diff --git a/vignettes/benchmarking.Rmd b/vignettes/benchmarking.Rmd
@@ -109,7 +109,7 @@ First we load the packages:
 import igraph as ig
 print("igraph", ig.__version__)
 import leidenalg as la
-print("leidenalg", la.__version__)
+print("leidenalg", la.version)
 ```
 
 Then we load the Zachary karate club example data from igraph.
@@ -148,7 +148,7 @@ plot(graph_object, vertex.color = node.cols, layout=layout_with_kk)
 
 We can reproduce passing arguments in this manner as well.
 
-````{python, eval=module}
+```{python, eval=module}
 partition = la.find_partition(G, la.CPMVertexPartition, resolution_parameter = 0.05)
 print(partition)
 partition
@@ -173,7 +173,7 @@ plot(graph_object, vertex.color = node.cols, layout=layout_with_kk)
 
 We can run the RBC vertex method which generalises the modularity vertex partition.
 
-````{python, eval=module}
+```{python, eval=module}
 partition = la.find_partition(G, la.RBConfigurationVertexPartition, resolution_parameter = 1.5)
 print(partition)
 partition
@@ -350,7 +350,7 @@ plot(G, vertex.color = node.cols, layout=layout_with_kk)
 
 We can run the RBC vertex method which generalises the modularity vertex partition.
 
-````{r, eval=module}
+```{r, eval=module}
 partition <- leiden(G, "RBConfigurationVertexPartition", resolution_parameter = 1.5)
 partition
 ```
