#!/usr/bin/env Rscript

# Creates a t-SNE plot of a normalised SingleCellExperiment object.

# Load optparse we need to check inputs

library(optparse)
library(workflowscriptscommon)
library(LoomExperiment)
library(scater)
library(Rtsne)

# parse options

option_list <- list(
  make_option(
    c("-i", "--input-loom"),
    action = "store",
    default = NA,
    type = "character",
    help = "A SingleCellExperiment object file in Loom format."
  ),
  make_option(
    c("-c", "--colour-by"),
    action = "store",
    default = NULL,
    type = "character",
    help = "Feature (from annotation file) to colour t-SNE plot points by. The values represented in this options should be categorical"
  ),
  make_option(
    c("-s", "--size-by"),
    action = "store",
    default = NULL,
    type = "character",
    help = "Feature (from annotation file) to size t-SNE plot points by. The values represented in this options should be numerical and not categorical"
  ),
  make_option(
    c("-p", "--shape-by"),
    action = "store",
    default = NULL,
    type = "character",
    help = "Feature (from annotation file) to shape t-SNE plot points by. The values represented in this options should be categorical"
  ),
  make_option(
    c("-o", "--output-plot-file"),
    action = "store",
    default = NA,
    type = "character",
    help = "Path of the PDF output file to save plot to."
  )
)

opt <- wsc_parse_args(option_list, mandatory = c("input_loom", "output_plot_file"))

# Check parameter values

if (! file.exists(opt$input_loom)) {
  stop((paste("File", opt$input_loom, "does not exist")))
}

# Filter out unexpressed features

sce <- import(opt$input_loom, format = "loom", type = "SingleCellLoomExperiment")
sce <- logNormCounts(sce)
sce <- runTSNE(sce, perplexity = 10)
plot <- plotTSNE(sce, colour_by = opt$colour_by, size_by = opt$size_by, shape_by = opt$shape_by)

ggsave(opt$output_plot_file, plot, device = "pdf")