Dendrogram mode fails edge case with many clones

[whelena]

The dendrogram mode was tested with a simulated sample with 1000 subclones. This seems to break the node collision prevention algorithm. The resulting plot seems to not have any optimization applied. Specifiying specific angle or spread param in the dataframe does not change the result.

test tsv: /hot/software/package/public-R-CancerEvolutionVisualization/development/hwinata-manuscript/data/multi-sample/phylogeny-1000.tsv

code:

phy1000.dt <- fread(test.tsv, header = TRUE, sep = '\t');
phy1000.dt[, `:=`(
    mode = 'dendrogram',
    length.1 = 10,
    connector.type = 'solid',
    connector.col = 'black',
    connector.width = 3
    )];
phy1000.dt[parent == which(label == '1'), spread := 5];

plt <- SRCGrob(
    phy1000.dt,
    horizontal.padding = 1.5,
    yaxis1.label = 'Distance (# of SNV)',
    add.normal = TRUE,
    use.radians = FALSE,
    draw.nodes = FALSE
    );

Resulting plot: