# Load all required libraries
library(CEMiTool, quietly = TRUE, warn.conflicts = FALSE)
library(ggplot2, quietly = TRUE, warn.conflicts = FALSE)
library(getopt, quietly = TRUE, warn.conflicts = FALSE)
# setup R error handling to go to stderr
options(
    show.error.messages = FALSE,
    error = function() {
        cat(geterrmessage(), file = stderr())
        q("no", 1, FALSE)
    }
)

# we need that to not crash galaxy with an UTF8 error on German LC settings.
loc <- Sys.setlocale("LC_MESSAGES", "en_US.UTF-8")

################################################################################
### Input Processing
################################################################################

# Collect arguments from command line
args <- commandArgs(trailingOnly = TRUE)

# Get options, using the spec as defined by the enclosed list.
# Read the options from the default: commandArgs(TRUE).
spec <- matrix(
    c(
        "expressionMatrix", "M", 1, "character",
        "sampleAnnotation", "A", 2, "character",
        "pathwayList", "P", 2, "character",
        "interactions", "I", 2, "character",
        "set_beta", "B", 1, "integer",
        "filter", "f", 1, "logical",
        "filter_pval", "i", 1, "numeric",
        "apply_vst", "a", 1, "logical",
        "n_genes", "n", 1, "integer",
        "eps", "e", 1, "numeric",
        "cor_method", "c", 1, "character",
        "cor_function", "y", 1, "character",
        "network_type", "x", 1, "character",
        "tom_type", "t", 1, "character",
        "merge_similar", "m", 1, "logical",
        "rank_method", "r", 1, "character",
        "min_ngen", "g", 1, "integer",
        "diss_thresh", "d", 1, "numeric",
        "center_func", "h", 1, "character",
        "ora_pval", "o", 1, "numeric",
        "gsea_scale", "l", 1, "logical",
        "gsea_min_size", "w", 1, "integer",
        "gsea_max_size", "z", 1, "integer",
        "sample_column_name", "v", 1, "character"
    ),
    byrow = TRUE, ncol = 4
)

opt <- getopt(spec)
counts <- read.table(
    opt$expressionMatrix,
    header = TRUE,
    sep = "\t",
    strip.white = TRUE,
    stringsAsFactors = FALSE,
    check.names = FALSE
)


# Run CEMiTool

if (is.null(opt$sampleAnnotation)) {
    cem <- cemitool(
        counts,
        filter = opt$filter,
        filter_pval = opt$filter_pval,
        apply_vst = opt$apply_vst,
        n_genes = opt$n_genes,
        eps = opt$eps,
        cor_method = opt$cor_method,
        cor_function = opt$cor_function,
        network_type = opt$network_type,
        tom_type = opt$tom_type,
        merge_similar = opt$merge_similar,
        min_ngen = opt$min_ngen,
        diss_thresh = opt$diss_thresh,
        center_func = opt$center_func,
        verbose = TRUE,
        ora_pval = opt$ora_pval,
        set_beta = opt$set_beta
    )
} else {
    annotation <- read.table(
        opt$sampleAnnotation,
        header = TRUE,
        sep = "\t",
        strip.white = TRUE,
        stringsAsFactors = FALSE,
        check.names = FALSE
    )
    cem <- cemitool(
        counts,
        annotation,
        filter = opt$filter,
        filter_pval = opt$filter_pval,
        apply_vst = opt$apply_vst,
        n_genes = opt$n_genes,
        eps = opt$eps,
        cor_method = opt$cor_method,
        cor_function = opt$cor_function,
        network_type = opt$network_type,
        tom_type = opt$tom_type,
        merge_similar = opt$merge_similar,
        min_ngen = opt$min_ngen,
        diss_thresh = opt$diss_thresh,
        center_func = opt$center_func,
        verbose = TRUE,
        ora_pval = opt$ora_pval,
        sample_name_column = opt$sample_column_name,
        class_column = "Class",
        order_by_class = TRUE
    )
    cem <- mod_gsea(
        cem,
        gsea_scale = opt$gsea_scale,
        gsea_min_size = opt$gsea_min_size,
        gsea_max_size = opt$gsea_max_size,
        rank_method = opt$rank_method
    )
    cem <- plot_gsea(cem)
}

if (!is.null(opt$pathwayList)) {
    gmt_in <- read_gmt(opt$pathwayList)
    cem <- mod_ora(cem, gmt_in)
    cem <- plot_ora(cem)
}

if (!is.null(opt$interactions)) {
    interactions <- read.table(
        opt$interactions,
        header = TRUE,
        sep = "\t",
        strip.white = TRUE,
        stringsAsFactors = FALSE,
        check.names = FALSE
    )
    interactions_data(cem) <- interactions # add interactions
    cem <- plot_interactions(cem)
}

## Write analysis results into files
write_files(cem,
    directory = "./Tables",
    force = TRUE
)

generate_report(cem)

save_plots(cem,
    value = "all",
    directory = "./Plots",
    force = TRUE
)