Ensuring same page rendering for Jupyter with multiple Plotly plots, each time the page is opened?

Question

I am currently using Jupyter on a server:

$ python3 --version
Python 3.8.10
$ pip3 list | grep 'jupyter \|nbextensions\|plotly'
jupyter                           1.0.0
jupyter-contrib-nbextensions      0.5.1
jupyter-nbextensions-configurator 0.4.1
plotly                            5.2.2

My problem is this: I have an .ipynb notebook, with multiple Plotly plots inside, generated by multiple Python code cells.

Then, I have both:

• enabled "Initialization cells" on all Python code cells, as per https://stackoverflow.com/a/39666438/6197439 , and saved the notebook
• Made a first %%html cell, with JavaScript to re-run all cells on page load, as per https://stackoverflow.com/a/38856870/6197439 , and saved the notebook

Now, the results are like this:

• If the kernel corresponding to the .ipynb notebook, has been shut down, and then I open the .ipynb notebook in a new browser tab, then indeed all cells run, and after a some seconds wait, I get the graphs/plots rendered in the page, as I want it
• But if now I close that browser tab, and the kernel corresponding to this .ipynb notebook still runs, and I click the .ipynb notebook again so it opens in a new tab, this time the graphs/plots are NOT visible in the page (instead, the corresponding Python code cells are expanded, even if the original saved state is collapsed - where I use the Codefolding nbextension, see collapse cell in jupyter notebook)
  • Although, at this time, I can call Cell/Run All, and the Plotly plots are generated - but the corresponding code cells remain expanded

So, my question is: is there anything I can do, so that whenever the .ipynb page is opened for the first time, all the Plotly plots are generated - regardless if this is the first page load since the kernel started, or later page loads?