Plotly graph objects returns error when trying to export the plot despite the graph being plotted without any errors

Question

I tried to make a plot and save it as svg. When I run the code that makes the plot (I'm using Jupyter), it makes the plot, and everything works fine. However, if I try to save it as svg the program returns ValueError:

ValueError: Transform failed with error code 525: Failed to execute 'insertBefore' on 'Node': The new child element contains the parent.

Here is the part of the code that runs without any errors:

x_values = [2.71152, 1.80768, 0.90384, 0.90384, 0.67788]
y_values = [430.28256, 503.08416, -235.5592, 626.42848, 752.61792]

x_new = [2.71152, 1.80768, 0.90384, 0.67788]
y_new = [430.28256, 503.08416, 626.42848, 752.61792]

a, b, r_val, p_val, slope_err = stats.linregress(x_new, y_new)
yfit = [a * xi + b for xi in x_new]

fig = go.Figure()
fig.add_traces([go.Scatter(name = "Data", 
                           x = x_values, 
                           y = y_values, 
                           mode = "markers+text", 
                           text = ["", r"$y = -144.5 x + 798.5$", "", ""],
                           textposition = "bottom left"), 
              go.Scatter(name = "Fit-Line", 
                           x = x_new, 
                           y = yfit, 
                           mode = "lines")])
fig.update_layout(font = dict(family = "CMU Serif", size = 16),
                  template = "simple_white", 
                  width = 900,
                  xaxis_title = r"$m \cdot (T_f - T_i), \text{ kg K}$",
                  yaxis_title = r"$Q_c, \text{ J}$")
fig.show()

And here is the line that throws the error:

fig.write_image("images/part-2-2-plot.svg")

It also says it "Cannot infer image type from output path '{file}'." However, I checked everything, the "images" directory exists in the folder, but the problem persists. Same problem if I try exporting to .png instead of .svg.

I tried removing parts of the code to see which line messes up everything. Found out that when I remove attributes xaxis_title and yaxis_title from the fig.update_layout() method, it does not throw any errors and exports the plot. Could someone please explain why it behaves like that? Did not seem to have problems with similar plots I made before.

I searched the internet but sadly did not find anything useful.

I don't get the exact same error, but I'm pretty sure either way the problem is the math in the axes titles. There's this bug with kaleido and mathjax. Kaleido is responsible for saving the images and depends on mathjax for making the latex math syntax work. One solution would be to use a different engine than kaleido, orca. The plotly documentation recommends kaleido and calls orca legacy, but you might not care. Another approach is saving the figure as .html but you might not want that. — bas
– bas, Commented Apr 13, 2023 at 21:25
If either of the above approaches works for you I can show how you can make it work in an answer. — bas
– bas, Commented Apr 13, 2023 at 21:28
@5eb yes, the first approach sounds right. In that project, I decided to remove axes titles, export the SVG, create separate SVGs with the titles using LaTeX and manually attach them to the exported SVG. However, a shorter and more direct way of solving the problem would be really useful in future. — Azamat
– Azamat, Commented Jun 1, 2023 at 19:38

OmG · Accepted Answer · 2023-10-11 11:42:31Z

0

It seems that there is a conflict between math style axes titles properties (i.e., xaxis_title and yaxis_title) and simple_white template (i.e., error in the source). For example, by changing the template to plotly_dark, you will get the following:

Another piece of evidence is changing the title from math style to a simple text also resolves the error, i.e., by keeping the simple_white template.

edited Oct 11, 2023 at 11:42

answered Oct 11, 2023 at 11:22

OmG

19k13 gold badges69 silver badges96 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Plotly graph objects returns error when trying to export the plot despite the graph being plotted without any errors

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related