#!/usr/bin/R
# -----------------------------------
# Load updated metadata / umap basis:
# -----------------------------------
library(cbrbase)
set_proj('DEVTRAJ', 'mosaicism')

source(paste0(sbindir, 'CKP25/load_metadata.R'))

# Load in the updated UMAPs w/ reclustering as well:
# --------------------------------------------------
meta.file = 'CKP25/CKp25_QCed_cells_metadata.tsv'
if (!file.exists(meta.file)){
    # Coordinates for Welch paper:
    new.umap.file = 'Annotation/CKP25_UMAP_clusters_120321_dataframe.tsv'
    updf = read.delim(new.umap.file, header=T)
    updf2 = updf[,c('cell','Xumap1','Xumap2')]
    names(updf2) = c('cell','U1','U2')
    meta = merge(meta, updf2, all.x=TRUE)

    # Coordinates for Dileep paper:
    dileep.umap.file = 'Annotation/CKP25_UMAP_clusters_03132023_dataframe.tsv'
    vddf = read.delim(dileep.umap.file, header=T)
    vddf = vddf[,c('cell','Xumap1','Xumap2')]
    names(vddf) = c('cell','V1','V2')
    meta = merge(meta, vddf, all.x=TRUE)
    write.table(meta, file=meta.file, quote=F, row.names=F, col.names=T, sep="\t")
} else {
    meta = read.delim(meta.file, header=T, sep="\t")
}


# Select UMAP basis to plot:
# --------------------------
reclustered = TRUE
if (reclustered){
    imgpref = paste0(imgpref, 'reclustered_'); 
    smeta = meta[!is.na(meta$U1),]
    u1 = 'U1'
    u2 = 'U2'
} else {
    no.deid = TRUE
    smeta = meta[meta$label != 'Batch',]
    if (no.deid){ 
        imgpref = paste0(imgpref, 'nodeid_'); 
        smeta = smeta[smeta$sublabel2 != 'Deid',]
    }
    u1 = 'Xumap1'
    u2 = 'Xumap2'
}

# Shuffle points for unbiased plotting:
ind = sample(1:nrow(smeta), nrow(smeta), replace=FALSE)
smeta = smeta[ind,]