R/method-dr-tibble.R

In YuLab-SMU/MicrobiotaProcess: A comprehensive R package for managing and analyzing microbiome and other ecological data within the tidy framework

#' @title convert the results of multiple dimensionality reduction to tbl_df
#' @rdname tidydr
#' @param x the R object, options 'prcomp', 'princomp', 'pcoa'
#' @param display character  
#' @param digits integer indicating the number of decimal places,
#' default is 2.
#' @param ... additional parameters
#' @return tbl_df object
#' @noRd
tidydr <- function(x, ...){
    UseMethod("tidydr")
}

#' @method tidydr prcomp
#' @importFrom stats setNames
#' @rdname tidydr
#' @noRd
tidydr.prcomp <- function(x, display="sites", digits=2, ...){
    vars <- x$sdev^2 
    vars <- (100 * vars/sum(vars)) %>%
             round(digits=digits) %>%
             paste0("(", ., "%)")
    
    if (inherits(x, "prcomp")){
        sites <- "x"
        features <- "rotation"
    }else if (inherits(x, "princomp")){
        sites <- "scores"
        features <- "loadings"
    }

    da <- x[[sites]] %>% 
          as.data.frame() %>%
          setNames(
                   x[[sites]] %>% 
                   colnames() %>% 
                   paste(vars)
          ) %>%
          tibble::as_tibble(rownames="sites")

    if ("features" %in% display){
        da %<>% 
            add_attr(
                     attribute=x[[features]] %>% 
                     tibble::as_tibble(rownames="features"),
                     name="features_tb"
                )
    }

    return(da)
}

#' @method tidydr princomp
#' @rdname tidydr
#' @noRd
tidydr.princomp <- tidydr.prcomp

#' @method tidydr pcoa
#' @rdname tidydr
#' @noRd
tidydr.pcoa <- function(x, digits=2, ...){
    if ("Rel_corr_eig" %in% names(x$values)){
       vars <- 100 * x$values$Rel_corr_eig[seq_len(ncol(x$vectors))]
    }else{
       vars <- 100 * x$values$Relative_eig[seq_len(ncol(x$vectors))]
    }
    vars %<>% round(digits=digits) %>% 
             paste0("(", ., "%)")

    da <- x$vectors %>%
          as.data.frame() %>%
          setNames(
              paste0("PCo", 
                     x$vectors %>% 
                         ncol %>% 
                         seq_len
                     ) %>% 
              paste(vars)
          ) %>%
          tibble::as_tibble(rownames="sites")

    return(da)
}

#' @method tidydr decorana
#' @rdname tidydr
#' @noRd
tidydr.decorana <- function(x, display="sites", digits=2, ...){
    da <- vegan::scores(x, display="sites", ...) %>%
          as.data.frame() %>%
          tibble::as_tibble(rownames="sites")

    if ("features" %in% display){
        dat <- vegan::scores(x, display="species", ...) %>%
               tibble::as_tibble(rownames="features")
        da %<>% 
            add_attr(dat, name="features_tb")
    }
    return(da)
}

#' @method tidydr metaMDS
#' @rdname tidydr
#' @noRd
tidydr.metaMDS <- function(x, display="sites", ...){
    da <- x %>% 
          vegan::scores(display="sites",...) %>%
          as.data.frame() %>%
          tibble::as_tibble(rownames="sites") 

    if ("features" %in% display){
        dat <- tryCatch(
                   vegan::scores(x, display="species", ...),
                   error=function(i){NULL}
               ) %>%
               tibble::as_tibble(rownames="features")
        da %<>% 
            add_attr(dat, name="features_tb")
    }
    return(da)
}

#' @method tidydr cca
#' @rdname tidydr
#' @noRd
tidydr.cca <- function(x, display=c("sp", "wa", "lc", "bp", "cn"), digits=2, ...){
    vars <- x %>% 
            vegan::eigenvals() %>% 
            summary %>% 
            data.frame %>% 
            dplyr::slice(2) %>% 
            as.numeric() * 100
    vars %<>% 
        round(digits=digits) %>% 
        paste0("(", ., "%)")

    if (length(display)<=1){
        display <- "sp"
    }
    dalist <- x %>% 
              vegan::scores(display, choices=seq_len(length(vars)), ...)

    da <- dalist$sites %>%
          as.data.frame() %>%
          setNames(
                   dalist$sites %>%
                   colnames() %>%
                   paste(vars)
                  ) %>%
          tibble::as_tibble(rownames="sites")

    dalist <- dalist[-match("sites", names(dalist))]
    dalist <- dalist[vapply(dalist, function(i)!all(is.na(i)), logical(1))]
    if (length(dalist)>0){
        leftnm <- names(dalist)
        for (i in leftnm){
            da %<>%
                add_attr(dalist[[i]] %>% 
                         tibble::as_tibble(rownames=i),
                         name=i)
        }
    }
    return(da)
}

#' @method tidydr rda
#' @rdname tidydr
#' @noRd
tidydr.rda <- tidydr.cca

#' Extracting the internal tbl_df attribute of tibble.
#' @rdname dr_extract
#' @param name character the name of internal tbl_df attribute.
#' @param .f a function (if any, default is NULL) that pre-operate the
#' data
#' @export
#' @return tbl_df object
#' @author Shuangbin Xu
#' @examples
#' \dontrun{
#' library(vegan)
#' data(varespec, varechem)
#' mpse <- MPSE(assays=list(Abundance=t(varespec)), colData=varechem)
#' tbl <- 
#' mpse %>%
#'   mp_cal_nmds(.abundance=Abundance, action="add") %>%
#'   mp_envfit(.ord=NMDS, .env=colnames(varechem), action="only") 
#' tbl 
#' tbl %>% attributes %>% names
#' # This function is useful to extract the data to display with ggplot2
#' # you can also refer to the examples of mp_envfit.
#' dr_extract(name=NMDS_ENVFIT_tb)(tbl)
#' # add .f function 
#' dr_extract(name=NMDS_ENVFIT_tb, 
#'            .f=td_filter(pvals<=0.05 & label!="Humdepth"))(tbl)
#' }
dr_extract <- function(name, .f=NULL){
    function(.data){
        name <- rlang::enquo(name)
        .data %<>% attr(rlang::as_name(name))
        if (!is.null(.f)){
            .data <- .f(.data)
        }
        return (.data)
    }
}