Plotly.Py cannot load JSON created in ggplotly in Plotly.R

[rl-utility-man]

Hello,

I want to give my R-coding colleagues access to a custom Plotly.py layout script by saving JSON in R and then loading it into Python. However, when I try this I plotly.io.read_json reports that the JSON produced in Plotly.R contains invalid tags and an invalid "transparent" color. If I make knock out two keys with Null values and replace the "transparent" color with white, then the JSON loads.

The attached R_to_JSON_to_Python.zip show a simple test case where I create a graph from the ggPlotly documentation, save it to JSON, fail to load in Plotly.IO, then make three changes to the JSON and can load it.

I can replicate this on the latest R (x86-64 4.1.0, 5/18/21) and the latest Plotly on CRAN; and using Plotly 5.1.0 and Python 3.8.10 on Windows.

Is an appropriate way to bridge R and Python Plotly to save JSON in R and load it in Python?

Is the existing work around shown here going to be robust to adding more complexity to the graphs without endangering data?

Thanks!

-Rob

[rl-utility-man]

Here is the R code for the test case:

#R/Python/Plotly Proof of concept

install.packages(c("plotly"))
library(plotly)

#make a simple bar chart using GGPlot2 notation.
#code is from here:
#   https://plotly.com/ggplot2/geom_bar/
dat <- data.frame(
    time = factor(c("Lunch","Dinner"), levels=c("Lunch","Dinner")),
    total_bill = c(14.89, 17.23)
)

p <- ggplot(data=dat, aes(x=time, y=total_bill)) +
    geom_bar(stat="identity")
#then convert it to plotly
fig <- ggplotly(p)

#convert it to JSON and save it to disk
#following example here:    https://stackoverflow.com/questions/48445671/plotly-generate-json-in-r-save-and-reuse-in-html-javascript
json<-plotly_json(fig, FALSE)
write(json,'c:/data/test.json')

Here is the Python code for the test case:

#R/Python/Plotly Proof of concept
import plotly.io as pio
import json



try:
    fig2 = pio.read_json("c:/data/test.json")
except Exception as e:
    print(repr(e))



##the JSON coming out of R is not quite readable in Python; but if I replace the bad keyword "transparent" with the hex code for white and get rid of 
#three blank entries that Plotly Python does not support, then I can load it.
with open("c:/data/test.json") as f:
    json_text = f.read()
    json_text=json_text.replace("transparent", "#ffffff").replace(""",
      "frame": null""","").replace('"family": "",','').replace(""",
          "linetype": []""","")
    assert "transparent" not in json_text
    #graph_json = json.loads(json_text)
    with open("C:/data/test_replaced.json", "w") as out:
        out.write(json_text)



fig2 = pio.read_json("C:/data/test_replaced.json")

fig2.show()

[dHannasch]

I have no idea what could be causing this to fail, but as a workaround, in Jupyter (which I assume is what's ultimately executing your Python code?) you can pass data between R and Python directly and possibly make the plots that way. See e.g. https://document-cookiecutters.gitlab.io/fork-this-jupyter-pages-ci/notebook.html .

[T-Flet]

plotly/plotly.R#2009 (comment)

[gvwilson]

Hi - we are trying to tidy up the stale issues and PRs in Plotly's public repositories so that we can focus on things that are still important to our community. Since this one has been sitting for several years, I'm going to close it; if it is still a concern, please add a comment letting us know what recent version of our software you've checked it with so that I can reopen it and add it to our backlog. Thanks for your help - @gvwilson