Interactive plot (#140)

[JaniceShu]

Use plotly to display hover info.
Remove legend and caption (related issue #150).
Add x-axis label.

Hover info the numbers are not formatted correctly. I tried to format it, but couldn't get config to pass into the code properly. @arnfinn, if there is an easy way, please enlighten me!

I have taken the legend away completely. I don't think it is necessary as the range of the bins are easily obtained from the plot.

[arnfinn]

This will fail, from tests/testthat/test_module_tabs.R:

    histogram <- output$plot_histogram
    # Check y values (name of health trusts)
    expect_equal_to_reference(histogram[["coordmap"]][["panels"]][[1]][["domain"]][["discrete_limits"]][["y"]],
                              "data/plot_histogram_y.rds")
    expect_equal(histogram[["coordmap"]][["panels"]][[1]][["mapping"]][["x"]], "get(\"value\")")
    expect_equal(histogram[["coordmap"]][["panels"]][[1]][["mapping"]][["y"]], "get(\"area_name\")")
    expect_equal(histogram[["coordmap"]][["panels"]][[1]][["mapping"]][["fill"]], "get(\"brks\")")

It seems like output$plot_histogram is now a json-object. Maybe test it with something like this (as been done for output$plot_map and output$make_table)?

histogram <- as.character(output$plot_histogram)
expect_true(grepl("<SOME STRING>", histogram))

tests/testthat/test_plot_variation.R is of course also failing. Seems like the data in data <- readRDS("data/start_plot.rds") is not good enough for the new histogram code. Maybe use something like this instead?

  data <- testdata %>%
    dplyr::filter(level1 == 1) %>%
    dplyr::filter(level2 == 1) %>%
    dplyr::filter(level3 == 1)

[arnfinn]

I am considering deactivating these tests... It will just make our life harder with all these tests during development. The plots will change a lot...

[arnfinn]

I forgot about your config question...

Not sure what you have tried. You can send the config directly in your function by adding config = config

    # BARCHART
    output$plot_histogram <- plotly::renderPlotly({
      plot <- helseatlas::plot_variation(
        input_data = data(),
        xlab = config$plot$xlab[[language()]],
        ylab = as.character(data()$type),
        config = config # New line
        )

and add it to the plot_variation

plot_variation <- function(input_data = NULL, xlab = "Area", ylab = "Rate", type = "histogram", num_groups = 5, config = NULL) {

Or maybe better: only send in what is needed. Thus, something like this:

    # BARCHART
    output$plot_histogram <- plotly::renderPlotly({
      plot <- helseatlas::plot_variation(
        input_data = data(),
        xlab = config$plot$xlab[[language()]],
        ylab = as.character(data()$type),
        big_mark = config$num$big[[language()]],
        decimal_mark = config$num$decimal[[language()]]
        )

We can do it in another PR:)