@@ -102,6 +102,7 @@ exportMethods(perplexity)

 exportMethods(resList)

 exportMethods(runParams)

 exportMethods(sampleLabel)

+exportMethods(selectBestModel)

 exportMethods(subsetCeldaList)

 import(Rcpp)

 import(RcppEigen)

@@ -541,50 +541,97 @@ setMethod("subsetCeldaList",

 #' @title Select best chain within each combination of parameters

 #' @description Select the chain with the best log likelihood for each

-#'  combination of tested parameters from a `celdaList` object gererated by

-#'  `celdaGridSearch()`.

-#' @param celdaList Object of class `celdaList`. An object containing celda

-#'  models returned from `celdaGridSearch()`.

-#' @param asList `TRUE` or `FALSE`. Whether to return the best model as a

-#'  `celdaList` object or not. If `FALSE`, return the best model as a

-#'  corresponding `celda_C`, `celda_G` or `celda_CG` object.

-#' @return A new `celdaList` object containing one model with the best log

-#'  likelihood for each set of parameters. If only one set of parameters is in

-#'  the `celdaList`, the best model will be returned directly instead of a

-#'  `celdaList` object.

-#' @seealso `celdaGridSearch()` can run Celda with multiple parameters and

-#'  chains in parallel. `subsetCeldaList()` can subset the `celdaList` object.

+#'  combination of tested parameters from a \code{SCE} object gererated by

+#'  \link{celdaGridSearch} or from a \code{celdaList} object.

+#' @param x Object of class \linkS4class{SingleCellExperiment} or

+#'  \code{celdaList}. An object containing celda

+#'  models returned from \link{celdaGridSearch}.

+#' @param asList \code{TRUE} or \code{FALSE}. Whether to return the

+#'  best model as a

+#'  \code{celdaList} object or not. If \code{FALSE}, return the best model as a

+#'  corresponding celda model object.

+#' @param useAssay A string specifying which \code{assay}

+#'  slot to use if \code{x} is a

+#'  \linkS4class{SingleCellExperiment} object. Default "counts".

+#' @return One of

+#' \itemize{

+#'  \item A new \linkS4class{SingleCellExperiment} object containing

+#'  one model with the best log-likelihood for each set of parameters in

+#'  \code{metadata(x)}. If there is only one set of parameters,

+#'  a new \linkS4class{SingleCellExperiment} object

+#'  with the matching model stored in the

+#'  \link[S4Vectors]{metadata}

+#'  \code{"celda_parameters"} slot will be returned. Otherwise, a new

+#'  \linkS4class{SingleCellExperiment} object with the subset models stored

+#'  in the \link[S4Vectors]{metadata}

+#'  \code{"celda_grid_search"} slot will be returned.

+#'  \item A new \code{celdaList} object containing one model with the best

+#'  log-likelihood for each set of parameters. If only one set of parameters

+#'  is in the \code{celdaList}, the best model will be returned directly

+#'  instead of a \code{celdaList} object.}

+#' @seealso \link{celdaGridSearch} \link{subsetCeldaList}

+#' @export

+setGeneric("selectBestModel", function(x, ...) {

+    standardGeneric("selectBestModel")})

+

+

+#' @rdname selectBestModel

+#' @examples

+#' data(sceCeldaCGGridSearch)

+#' ## Returns same result as running celdaGridSearch with "bestOnly = TRUE"

+#' sce <- selectBestModel(sceCeldaCGGridSearch)

+#' @importFrom data.table as.data.table

+#' @export

+setMethod("selectBestModel", signature(x = "SingleCellExperiment"),

+    function(x, asList = FALSE, useAssay = "counts") {

+        logLikelihood <- NULL

+        group <- setdiff(colnames(runParams(x)),

+            c("index", "chain", "logLikelihood", "mean_perplexity", "seed"))

+        dt <- data.table::as.data.table(runParams(celdaList))

+        newRunParams <- as.data.frame(dt[, .SD[which.max(logLikelihood)],

+            by = group])

+        newRunParams <- newRunParams[, colnames(runParams(celdaList))]

+

+        ix <- match(newRunParams$index, runParams(celdaList)$index)

+        if (nrow(newRunParams) == 1 & !asList) {

+            x <- celdatosce(resList(x)[[ix]],

+                SummarizedExperiment::assay(x, i = useAssay))

+        } else {

+            x@metadata$celda_grid_search@runParams <-

+                as.data.frame(newRunParams)

+            x@metadata$celda_grid_search@resList <- resList(x)[ix]

+        }

+        return(x)

+    }

+)

+

+

 #' @examples

 #' data(celdaCGGridSearchRes)

 #' ## Returns same result as running celdaGridSearch with "bestOnly = TRUE"

 #' cgsBest <- selectBestModel(celdaCGGridSearchRes)

 #' @importFrom data.table as.data.table

 #' @export

-selectBestModel <- function(celdaList, asList = FALSE) {

-  if (!methods::is(celdaList, "celdaList")) {

-    stop("celdaList parameter was not of class celdaList.")

-  }

+setMethod("selectBestModel", signature(x = "celdaList"),

+    function(x, asList = FALSE) {

+        logLikelihood <- NULL

+        group <- setdiff(colnames(runParams(x)),

+            c("index", "chain", "logLikelihood", "mean_perplexity", "seed"))

+        dt <- data.table::as.data.table(runParams(x))

+        newRunParams <- as.data.frame(dt[, .SD[which.max(logLikelihood)],

+            by = group])

+        newRunParams <- newRunParams[, colnames(runParams(x))]

-  logLikelihood <- NULL

-  group <- setdiff(

-    colnames(runParams(celdaList)),

-    c("index", "chain", "logLikelihood", "mean_perplexity", "seed")

-  )

-  dt <- data.table::as.data.table(runParams(celdaList))

-  newRunParams <- as.data.frame(dt[, .SD[which.max(logLikelihood)],

-    by = group

-  ])

-  newRunParams <- newRunParams[, colnames(runParams(celdaList))]

-

-  ix <- match(newRunParams$index, runParams(celdaList)$index)

-  if (nrow(newRunParams) == 1 & !asList) {

-    return(resList(celdaList)[[ix]])

-  } else {

-    celdaList@runParams <- as.data.frame(newRunParams)

-    celdaList@resList <- resList(celdaList)[ix]

-    return(celdaList)

-  }

-}

+        ix <- match(newRunParams$index, runParams(x)$index)

+        if (nrow(newRunParams) == 1 & !asList) {

+            return(resList(x)[[ix]])

+        } else {

+            x@runParams <- as.data.frame(newRunParams)

+            x@resList <- resList(x)[ix]

+            return(x)

+        }

+    }

+)

 .createSCEceldaGridSearch <- function(celdaList,

@@ -2,35 +2,55 @@

 % Please edit documentation in R/celdaGridSearch.R

 \name{selectBestModel}

 \alias{selectBestModel}

+\alias{selectBestModel,SingleCellExperiment-method}

 \title{Select best chain within each combination of parameters}

 \usage{

-selectBestModel(celdaList, asList = FALSE)

+selectBestModel(x, ...)

+

+\S4method{selectBestModel}{SingleCellExperiment}(x, asList = FALSE, useAssay = "counts")

 }

 \arguments{

-\item{celdaList}{Object of class `celdaList`. An object containing celda

-models returned from `celdaGridSearch()`.}

+\item{x}{Object of class \linkS4class{SingleCellExperiment} or

+\code{celdaList}. An object containing celda

+models returned from \link{celdaGridSearch}.}

+

+\item{asList}{\code{TRUE} or \code{FALSE}. Whether to return the

+best model as a

+\code{celdaList} object or not. If \code{FALSE}, return the best model as a

+corresponding celda model object.}

-\item{asList}{`TRUE` or `FALSE`. Whether to return the best model as a

-`celdaList` object or not. If `FALSE`, return the best model as a

-corresponding `celda_C`, `celda_G` or `celda_CG` object.}

+\item{useAssay}{A string specifying which \code{assay}

+slot to use if \code{x} is a

+\linkS4class{SingleCellExperiment} object. Default "counts".}

 }

 \value{

-A new `celdaList` object containing one model with the best log

- likelihood for each set of parameters. If only one set of parameters is in

- the `celdaList`, the best model will be returned directly instead of a

- `celdaList` object.

+One of

+\itemize{

+ \item A new \linkS4class{SingleCellExperiment} object containing

+ one model with the best log-likelihood for each set of parameters in

+ \code{metadata(x)}. If there is only one set of parameters,

+ a new \linkS4class{SingleCellExperiment} object

+ with the matching model stored in the

+ \link[S4Vectors]{metadata}

+ \code{"celda_parameters"} slot will be returned. Otherwise, a new

+ \linkS4class{SingleCellExperiment} object with the subset models stored

+ in the \link[S4Vectors]{metadata}

+ \code{"celda_grid_search"} slot will be returned.

+ \item A new \code{celdaList} object containing one model with the best

+ log-likelihood for each set of parameters. If only one set of parameters

+ is in the \code{celdaList}, the best model will be returned directly

+ instead of a \code{celdaList} object.}

 }

 \description{

 Select the chain with the best log likelihood for each

- combination of tested parameters from a `celdaList` object gererated by

- `celdaGridSearch()`.

+ combination of tested parameters from a \code{SCE} object gererated by

+ \link{celdaGridSearch} or from a \code{celdaList} object.

 }

 \examples{

-data(celdaCGGridSearchRes)

+data(sceCeldaCGGridSearch)

 ## Returns same result as running celdaGridSearch with "bestOnly = TRUE"

-cgsBest <- selectBestModel(celdaCGGridSearchRes)

+sce <- selectBestModel(sceCeldaCGGridSearch)

 }

 \seealso{

-`celdaGridSearch()` can run Celda with multiple parameters and

- chains in parallel. `subsetCeldaList()` can subset the `celdaList` object.

+\link{celdaGridSearch} \link{subsetCeldaList}

 }