Vfrdtyky

I'm plotting a heatmap using R plotly:

set.seed(1)

df <- reshape2::melt(matrix(rnorm(100*20),100,20,dimnames = list(paste0("G",1:100),paste0("S",1:20))))



library(plotly)

library(dplyr)

plot_ly(z=c(df$value),x=df$Var2,y=df$Var1,colors=grDevices::colorRamp(c("darkblue","gray","darkred")),type="heatmap",colorbar=list(title="Scaled Value",len=0.4)) %>%

  layout(yaxis=list(title=NULL),xaxis=list(tickangle=90,tickvals=10,ticktext="X-Label"))

As you can see, plotly is not showing all y-axis ticks. My question is whether it is possible, and if so how, to retrieve the y-axis tick labels plotly selected to show?

I saved the plot object:

plotly.obj <- plot_ly(z=c(df$value),x=df$Var2,y=df$Var1,colors=grDevices::colorRamp(c("darkblue","gray","darkred")),type="heatmap",colorbar=list(title="Scaled Value",len=0.4)) %>%

  layout(yaxis=list(title=NULL),xaxis=list(tickangle=90,tickvals=10,ticktext="X-Label"))

And looked around and it seems that perhaps plotly.obj$x$layoutAttrs should store this information but it doesn't:

> plotly.obj$x$layoutAttrs

$`102ce55fd393e`

$`102ce55fd393e`$yaxis

$`102ce55fd393e`$yaxis$title

NULL





$`102ce55fd393e`$xaxis

$`102ce55fd393e`$xaxis$tickangle

[1] 90



$`102ce55fd393e`$xaxis$tickvals

[1] 10



$`102ce55fd393e`$xaxis$ticktext

[1] "X-Label"

Any idea?

I don't think you can get the ticks, that are finally rendered. But you can get all the levels of the y-axis, that ploty can choose from.

levels(plotly.obj$x$attrs$`2c4c148651ae`$y)

The ticks that are finally rendered are dynamically chosen and will adapt, depending on your plot size etc.

You can also check out the attributes with plotly_json():

plot_ly(z=c(df$value),x=df$Var2,y=df$Var1,colors=grDevices::colorRamp(c("darkblue","gray","darkred")),type="heatmap",colorbar=list(title="Scaled Value",len=0.4)) %>%

      layout(yaxis=list(title=NULL),xaxis=list(tickangle=90,tickvals=10,ticktext="X-Label")) %>% 

      plotly_json()

I got the answer from a github issue I posted on ropensci/plotly:

set.seed(1)

df <- reshape2::melt(matrix(rnorm(100*20),100,20,dimnames = list(paste0("G",1:100),paste0("S",1:20))))



library(plotly)

library(dplyr)

plot_ly(z=c(df$value),x=df$Var2,y=df$Var1,colors=grDevices::colorRamp(c("darkblue","gray","darkred")),type="heatmap",colorbar=list(title="Scaled Value",len=0.4)) %>%

  layout(yaxis=list(title=NULL),xaxis=list(tickangle=90,tickvals=10,ticktext="X-Label")) %>% 

  htmlwidgets::onRender(

    "function(el, x) {

       alert(el._fullLayout.yaxis._vals.map(function(i) { return i.text; }));

    }"

  )

Will pop up a browser window with the tick labels.

The question now is if this can be saved/piped to an R variable or written to a file so it can be done automatically rather than interactively. That's going to be another post.