updates, inherits changes

CHANGELOG.md

@@ -1,5 +1,10 @@
 ## Upcoming
 
+## [1.4.7] - 2022-04-Sept
+- Fixes for Matrix, updated versions 1.4.2, 1.5.0
+- Use `inherits()` for conditionals with class()
+
+
 ## [1.4.6] - 2022-30-March
 
 - Fixed bug with `snn` argument in `buildGraph()` due to the parameter `snn.k.self`

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: conos
 Title: Clustering on Network of Samples
-Version: 1.4.6
+Version: 1.4.7
 Authors@R: c(person("Viktor","Petukhov", email="[email protected]", role="aut"), person("Nikolas","Barkas", email="[email protected]", role="aut"), person("Peter", "Kharchenko", email = "[email protected]", role = "aut"), person("Weiliang", "Qiu", email = "[email protected]", role = c("ctb")), person("Evan", "Biederstedt", email="[email protected]", role=c("aut", "cre")))
 Description: Wires together large collections of single-cell RNA-seq datasets, which allows for both the identification of recurrent cell clusters and the propagation of information between datasets in multi-sample or atlas-scale collections. 'Conos' focuses on the uniform mapping of homologous cell types across heterogeneous sample collections. For instance, users could investigate a collection of dozens of peripheral blood samples from cancer patients combined with dozens of controls, which perhaps includes samples of a related tissue such as lymph nodes. This package interacts with data available through the 'conosPanel' package, which is available in a 'drat' repository. To access this data package, see the instructions at <https://github.com/kharchenkolab/conos>. The size of the 'conosPanel' package is approximately 12 MB.
 License: GPL-3

R/de_functions.R

@@ -34,7 +34,7 @@ validatePerCellTypeParams <- function(con.obj, groups, sample.groups, ref.level,
   if (is.null(ref.level)) stop('reference level is not defined')
   ## todo: check samplegrousp are named
   if(is.null(names(sample.groups))) stop('sample.groups must be named')
-  if(class(groups) != 'factor') stop('groups must be a factor')
+  if(!inherits(groups,'factor')) stop('groups must be a factor')
   if(any(grepl(cluster.sep.chr, names(con.obj$samples),fixed=TRUE)))
     stop('cluster.sep.chr must not be part of any sample name')
   if(any(grepl(cluster.sep.chr,levels(groups),fixed=TRUE)))
@@ -47,10 +47,10 @@ validateBetweenCellTypeParams <- function(con.obj, groups, sample.groups, refgro
     stop("You have to install DESeq2 package to use differential expression")
   }
 
-  if (class(con.obj) != 'Conos') stop('con.obj must be a conos object')
+  if (!inherits(con.obj,'Conos')) stop('con.obj must be a conos object')
   if (is.null(groups) ) stop('groups must be specified');
   if (is.null(sample.groups) ) stop('sample.groups must be specified')
-  if (class(sample.groups) != 'list' ) stop('sample.groups must be a list');
+  if (!inherits(sample.groups,'list')) stop('sample.groups must be a list');
   #if ( length(sample.groups) != 2 ) stop('sample.groups must be of length 2');
   if (!all(unlist(lapply(sample.groups, function(x) class(x) == 'character'))) )
     stop('sample.groups must be a list of character vectors');
@@ -62,7 +62,7 @@ validateBetweenCellTypeParams <- function(con.obj, groups, sample.groups, refgro
   if (is.null(altgroup)) stop('altgroup is not defined')
   ## todo: check samplegrousp are named
   if(is.null(names(sample.groups))) stop('sample.groups must be named')
-  if(class(groups) != 'factor') stop('groups must be a factor')
+  if(!inherits(groups,'factor')) stop('groups must be a factor')
   if(any(grepl(cluster.sep.chr, names(con.obj$samples),fixed=TRUE)))
     stop('cluster.sep.chr must not be part of any sample name')
   if(any(grepl(cluster.sep.chr,levels(groups),fixed=TRUE)))

README.md

@@ -238,5 +238,5 @@ The R package can be cited as:
 ```
 Viktor Petukhov, Nikolas Barkas, Peter Kharchenko, and Evan
 Biederstedt (2021). conos: Clustering on Network of Samples. R
-package version 1.4.6.
+package version 1.4.7.
 ```